Functional-structural plant models simulate plant responses to environmental conditions, but their development and evaluation are often limited by the lack of datasets combining detailed architectural and physiological measurements. Here, we present a comprehensive dataset acquired from four oil palm plants ( Elaeis guinnensis) grown under controlled and contrasting climate scenarios. The dataset includes (i) three-dimensional reconstructions of plant architecture derived from terrestrial lidar point clouds, (ii) leaf-level gas exchange measurements used to parameterize photosynthesis and stomatal conductance models, and (iii) continuous plant-scale measurements of CO 2 and H 2 O fluxes obtained in a microcosm under precisely monitored and manipulated environmental conditions (light, temperature, humidity, and CO 2 concentration) across height climate scenarios. By combining detailed structural data with physiological measurements at both leaf and whole-plant scales, this database has been designed to build and evaluate digital twins (or shadows) of plants functioning under controlled conditions. It provides a valuable resource for calibrating biophysical models (light interception and photosynthesis), benchmarking model predictions across scales, and investigating the consistency between leaf-level parameterization and plant-level fluxes. All data and processing workflows are openly available, facilitating reuse for model development, evaluation, and intercomparison in plant and crop modelling communities.
Penyakit Basal Stem Rot (BSR) merupakan penyakit penting pada tanaman kelapa sawit yang dapat menurunkan produktivitas. Analisis berbasis citra dapat mendukung deteksi penyakit, tetapi karakteristik distribusi warna perlu dipahami sebelum tahap klasifikasi. Penelitian ini bertujuan menganalisis distribusi warna citra daun kelapa sawit kelas Healthy dan BSR menggunakan Histogram pada ruang warna RGB dan HSV serta mengevaluasi kemampuan diskriminatif fitur warna. Dataset terdiri atas 2.438 citra dari repositori terbuka Roboflow. Setiap citra diproses melalui resize 224×224 piksel, Gaussian Blur, konversi RGB ke HSV, serta ekstraksi 12 parameter statistik berupa mean dan standar deviasi. Perbedaan distribusi antar kelas diuji menggunakan Mann–Whitney U dan besarnya perbedaan dihitung menggunakan Cohen's d. Kemampuan diskriminatif G_Mean, S_Mean, dan V_Mean dievaluasi menggunakan Receiver Operating Characteristic (ROC) dan Area Under the Curve (AUC). Hasil menunjukkan ketiga fitur memiliki perbedaan signifikan (p-value = 0,0000) dengan ukuran efek besar. Nilai AUC G_Mean, S_Mean, dan V_Mean masing-masing sebesar 0,9794, 0,7362, dan 0,9790. Hasil menunjukkan bahwa analisis distribusi warna dapat mengidentifikasi fitur diskriminatif sebagai dasar pemilihan fitur dan perancangan pra-pemrosesan citra BSR.
Oil palm is an essential commodity for the economy; however, basal stem rot caused by Ganoderma boninense poses a significant threat to plantation productivity and long-term vitality. It highlights the importance of early detection of stem disease to facilitate timely intervention and minimize potential economic losses. This study presents an image-based approach to diagnosing oil palm stem maladies, leveraging handcrafted color and texture features within a supervised machine learning framework. The dataset contained 525 images of oil palm stems, of which 205 depicted healthy specimens, and 320 depicted diseased ones. These were captured within their natural environment. Color features were derived by analyzing color moments within the HSV color space, while texture features were extracted from the Grey-Level Co-occurrence Matrix (GLCM). The extracted features were classified employing an Artificial Neural Network (ANN) and were subsequently contrasted with classifiers including Decision Tree, K-Nearest Neighbors, Naive Bayes, and Support Vector Machine. Model performance was evaluated using k-fold cross-validation with k = 5 and k = 10 to ensure the consistency and reliability of the assessment. The experimental results demonstrated that the highest accuracy of 97.52% was achieved when the ANN model was used to classify the integrated color and texture features. The innovative aspect of this research resides in demonstrating that handcrafted features integrated with artificial neural networks can attain high detection accuracy in scenarios with limited data, providing a viable alternative to data-intensive deep learning techniques. This method facilitates a dependable, computer vision-driven early detection system for oil palm stem diseases, thereby promoting sustainable plantation management.
The presence of soil hydrocarbon parameters (SHPs), including total petroleum hydrocarbons (TPHs), total organic carbon (TOC; %), and soil toxicity (EC50; mg L−1), can affect vegetation in several ways. This study assessed the impact of SHPs on vegetation in the Niger Delta using field-measured, leaf-scale hyperspectral data acquired across the region. Red-edge position (REP) and four hyperspectral vegetation indices (HVIs)—mND705, photochemical reflectance index (PRI), Normalised Difference Vegetation Vigour Index (NDVVI844,447; a vegetation vigour index), and modified DATT (MDATT; a chlorophyll-sensitive red-edge index)—were used to quantify chlorophyll content in the vegetation types of Awolowo grass, elephant grass, mango trees, oil palm trees, and mangrove vegetation and to explore their variation with SHPs. The results show that mangrove vegetation was the most impacted by TPHs (R = −0.683), while mango vegetation was the most impacted by TOC (R = −0.725), based on Pearson correlation coefficients derived from the mND705 index. Similarly, mango and mangrove vegetation showed the strongest responses to soil toxicity (EC50; mg L−1), based on Spearman correlation coefficients (rs = 0.657 and rs = 0.870, respectively) using the MDATT index. These findings highlight species-specific physiological responses to soil hydrocarbon contamination and demonstrate the applicability of red-edge-based hyperspectral techniques for assessing vegetation stress in complex coastal ecosystems such as the Niger Delta.
ABSTRACT Oil palm production has increased rapidly since 2012, particularly in Guatemala, Malaysia, and Indonesia. Accurate Fresh Fruit Bunch classification is vital for oil yield and quality, but manual grading is inefficient and existing deep learning methods are computationally intensive. To overcome these challenges, this study proposes an optimized Oil Palm Fresh Fruit Bunch Ripeness classification framework based on a Mixed‐Order Relation‐Aware Recurrent Neural Network (OP‐FFBR‐MORA‐RNN). The proposed methodology integrates Confidence Partitioning Sampling Filtering (CPSF) for effective image localization, cropping, and resizing, followed by Revised Tunable Q‐Factor Wavelet Transform (RTQFWT) to extract discriminative features. These features are subsequently classified using a Mixed‐Order Relation‐Aware Recurrent Neural Network (MORA‐RNN), with its parameters optimized via the Fractional Pelican African Vulture Optimization (FPAVO) algorithm to enhance classification accuracy. The model categorizes FFBs into five ripeness classes: overripe, ripe, abnormal, empty fruit, and under‐ripe. Experimental evaluation on an oil palm dataset demonstrates that OP‐FFBR‐MORA‐RNN achieves 97.8% accuracy, 97.6% precision, and a low error rate of 2.4%, outperforming existing methods such as Oil Palm Fresh Fruit Bunch Ripeness categorization on mobile devices using Convolutional Neural Network (OP‐FFBR‐MD‐CNN), Machine Vision for maturity classification using Artificial Neural Network (MVM‐OP‐FFBR‐ANN), and Object Detection for Oil Palm Fruit Bunches using You‐Only‐Look‐Once (OB‐OPFBR‐YOLOv7). These results confirm the framework's effectiveness for reliable, scalable, and efficient ripeness classification, supporting improved agricultural monitoring and optimized crop management.
Ganoderma boninense infection affects nearly half of Indonesia's palm oil plantations, which leads to severe economic losses. Conventional detection methods like PCR are limited by cost and complexity, while electrochemical detection offers a promising alternative by targeting plant-derived metabolites. This study focuses on developing a low-cost, custom fabricated screen-printed electrode (SPE) modified with MWCNTs and Au/Pt nanoparticles to enable practical, scalable early detection of G. boninense in oil palm plantations using various electrochemical techniques. The SPE was fabricated using conductive carbon ink on sticker paper and modified with MWCNT-Au/Pt nanocomposites to enhance performance. Pt nanoparticles (2.5, 5, and 10 mM) were synthesized hydrothermally, while Au (5 mM) was prepared via citrate reduction. The nanocomposites were formed by sonication and applied via drop-casting. Characterization was performed using UV-vis, FTIR, XPS, SEM, contact angle, four-point probe, CV, and EIS. Electrochemical detection of phytol, quinoline, and stigmasterol was conducted using DPV and chronoamperometry in phosphate buffer saline. The results indicate that surface modification enhances conductivity and hydrophilicity, improving the charge transfer process in electrochemical detection. The best combination for SPE modification was the 3 : 1 ratio and 5 mM Pt concentration. The best detection results using DPV were achieved for phytol, quinoline, and stigmasterol, with respective R 2 and LOD of 0.98 and 2.85 mM, 0.95 and 2.36 µM, and 0.98 and 1.36 µM. Repeated testing on the SPE demonstrated good stability, despite being designed as a disposable test kit. Further, the SPEs demonstrate low detection limits and high sensitivity, highlighting their potential for early diagnosis and rapid screening of Ganoderma -infected oil palm trees as an accessible alternative for disease prevention.
In this work, we present a Physics-Informed Neural Network (PINN) framework for the classification of oil palm fresh fruit bunch (FFB) ripeness using RGB images. Unlike conventional Convolutional Neural Networks (CNNs) that learn solely from visual patterns, the proposed PINN integrates a physics-based index—derived from the red-to-green pixel intensity ratio—directly into the network architecture and loss function. This hybrid design embeds wavelength-dependent physical knowledge related to chlorophyll degradation during ripening, enabling the model to learn more robust and generalizable features even with limited and imbalanced training data. The PINN model achieves a peak accuracy of 0.73, outperforming the purely data-driven CNN baseline (0.68) by a margin of 5%. Overall, the PINN demonstrates superior performance in minority-class detection and maintains stable convergence under three different lighting conditions (different light spectra). These results highlight the effectiveness of integrating domain-specific physical insights into deep learning models, offering a promising pathway toward reliable, non-destructive, and automated ripeness assessment for agricultural applications.
Accurate identification of oil palm fruit varieties is essential for supporting breeding programs and optimizing seed quality in plantation operations. Manual approaches often lead to inconsistencies due to the high visual similarity among fruit types, particularly between dura and tenera. This study proposes an automatic classification model for oil palm fruit cross-sections using HSV-based color feature extraction combined with a Gaussian Naïve Bayes classifier. A dataset of 186 cross-sectional fruit images was used, consisting of 90 training samples and 96 testing samples representing the dura, pisifera, and tenera varieties. The methodology includes preprocessing, segmentation, HSV feature extraction, model training, and performance evaluation through a confusion matrix. Experimental results show that the proposed model achieves an accuracy of 85%, with misclassifications primarily occurring in the tenera class due to its close resemblance to the dura variety. Compared to Linear Discriminant Analysis (LDA), the proposed approach demonstrates faster computation time and competitive accuracy. These findings indicate that Gaussian Naïve Bayes, supported by HSV feature descriptors, provides an efficient solution for lightweight and cost-effective digital classification of oil palm fruit varieties
Introduction High-resolution plant imagery is vital for phenotyping, disease monitoring, and precision agriculture. However, image acquisition in real-world conditions is frequently limited by sensor resolution, cost, and environmental noise, resulting in low-quality images. While deep learning-based super-resolution (SR) approaches show promise, they often fail to recover fine structural details that are essential for plant science applications. Methods We propose a plug-and-play high-frequency feature enhancement (HF-FE) module that can be seamlessly integrated into existing SR architectures. The module selectively amplifies high-frequency information, thereby improving the reconstruction of subtle details such as leaf venation, lesion boundaries, and texture patterns, while maintaining computational efficiency. Performance was evaluated on three diverse plant datasets: an oil palm dataset for large-scale plantation imagery, the UAV-based AqUAVPlant dataset for aquatic plants, and the Plant Pathology 2020 dataset for crop disease imagery. Results Across all datasets, models incorporating the HF-FE module achieved consistent improvements over state-of-the-art (SOTA) baselines, with notable gains in PSNR and SSIM. Visual assessments further confirmed enhanced clarity of fine structural features, particularly in challenging plant imaging scenarios. Discussion The proposed HF-FE module provides a flexible and effective enhancement strategy for plant image SR. By improving the fidelity of reconstructed plant imagery, it supports more accurate visualization and analysis, offering a methodological advancement that contributes to intelligent plant sensing, supports digital agriculture, and facilitates sustainable crop management.
Oil palm (Elaeis guineensis) productivity was frequently constrained by foliar diseases, which were often difficult to detect at an early stage using conventional visual inspection methods. To address this challenge, the present study proposed a hybrid deep learning framework for automated oil palm leaf disease detection. A dataset comprising 1,200 oil palm leaf images, equally distributed across three disease classes (400 images per class), was utilized. The dataset was partitioned into training (70%), validation (15%), and testing (15%) subsets, with training and validation data obtained from public repositories, while testing data were collected directly to ensure model generalizability. The proposed hybrid architecture combined U-Net for precise leaf lesion segmentation, ResNet-50 as a deep feature extractor to capture high-level discriminative representations. U-Net segmentation enabled isolation of infected regions, while ResNet-50 provided robust feature embeddings that enhanced separability between visually similar disease classes. Experimental evaluation demonstrated that the baseline U-Net + SVM approach achieved an accuracy of 84.2%, precision of 82.5%, recall of 83.1%, and F1-score of 82.8%. In contrast, the hybrid U-Net + ResNet-50 + SVM method yielded superior results with 91.6% accuracy, 90.8% precision, 91.2% recall, and 91.0% F1-score, reflecting an improvement of approximately 7.4%.
Oil palm is a globally significant economic crop, especially in Malaysia and Indonesia, where it contributes substantially to national economies. Accurate monitoring of plantations, particularly tree count and health, is essential for effective management and yield estimation. However, traditional field-based methods are costly and labor-intensive, and while UAVs and very high-resolution satellite imagery offer precision, they are often limited by high cost, limited coverage, and technical constraints such as altitude variation and image mosaicking. This study proposes a scalable and cost-effective pipeline that utilizes UAV imagery implementing two tree-counting approaches Template Matching, serving as a classical baseline, and YOLOv8, a deep learning object detection model chosen for its high accuracy and inference speed. The oil palm tree density was classified into High, Medium, and Low category for easy result observation. Preliminary results demonstrate promising True Positive Rate (TPR), False Negative Rate (FNR), and Estimation Error (EE) values, with YOLOv8 achieving its best performance in medium-density regions (TPR: 109.54%, FNR: 9.50%, EE: 12.91 %) and Template Matching showing the weakest performance in low-density areas (TPR: 305.45%, FNR: 205.45 %, EE: 205.45 %). These results indicate that the proposed approach offers a practical, real-time, and resourceefficient solution for large-scale oil palm monitoring, although improvements are still required in low-density plantations.
Vannesa Montoya‐Sánchez · Anna K. Schweiger · Michael Schlund · Gustavo Brant Paterno · Stefan Erasmi · Holger Kreft · Dirk Hölscher · Fabian Brambach · Bambang Irawan · Leti Sundawati · Delphine Clara Zemp
Oil palmAerial / UAVField / plotLiDAR / point cloudMultispectral / hyperspectralRaman / spectroscopyWhole plant / canopy / plot / field
Abstract Assessing plant diversity using remote sensing, including airborne imaging spectroscopy, shows promise for large‐scale biodiversity monitoring in landscape restoration and conservation. Enriching plantations with native trees is a key restoration strategy to enhance biodiversity and ecosystem functions in agricultural lands. In this study, we tested how well imaging spectroscopy characterizes plant diversity in 37 experimental plots of varying sizes and planted diversity levels in a biodiversity‐enriched oil palm plantation in Sumatra, Indonesia. Six years after establishing the plots, we acquired airborne imaging spectroscopy data comprising 160 spectral bands (400–1000 nm, at ~3.7 nm bandwidth) at 0.3 m spatial resolution. We calculated spectral diversity as the variance among image pixels and partitioned spectral diversity into alpha and beta diversity components. After controlling for differences in sampling area through rarefaction, we found no significant relationship between spectral and plant alpha diversity. Further, the relationships between the local contribution of spectral beta diversity and plant beta diversity revealed no significant trends. Spectral variability within plots was substantially higher than among plots (spectral alpha diversity ~82%–87%, spectral beta diversity ~11%–18%). These discrepancies are likely due to the structural dominance of oil palm crowns, which absorbed most of the light, while most of the plant diversity occurring below the oil palm canopy was not detectable by airborne spectroscopy. Our study highlights that remote sensing of plant diversity in ecosystems with strong vertical stratification and high understory diversity, such as agroforests, would benefit from combining data from passive with data from active sensors, such as LiDAR, to capture structural diversity.
The vast size of oil palm (Elaeis guineensis) plantations has led to lightweight unmanned aerial vehicles (UAVs) being identified as cost effective tools to generate inventories for improved plantation management, with proximal aerial data capable of resolving single palm canopies at potentially, centimetric resolution. If acquired with sufficient overlap, aerial data from UAVs can be processed within structure-from-motion (SfM) photogrammetry workflows to yield volumetric point cloud representations of the scene. Point cloud-derived structural information on individual palms can benefit not only plantation management but is also of great environmental research interest, given the potential to deliver spatially contiguous quantifications of aboveground biomass, from which carbon can be accounted. Using lightweight UAVs we captured data over plantation plots of varying ages (2, 7 and 10 years) at peat soil sites in Sarawak, Malaysia, and we explored the impact of changing spatial resolution and image overlap on spatially variable uncertainties in SfM derived point clouds for the ten year old plot. Point cloud precisions were found to be in the decimetre range (mean of 26.7 31 cm) for a 10 year old plantation plot surveyed at 100 m flight altitude and >75% image overlap. Derived canopy height models were used and evaluated for automated palm identification using local height maxima. Metrics such as maximum canopy height and stem height, derived from segmented single palm point clouds were tested relative to ground validation data. Local maximum identification performed best for palms which were taller than surrounding undergrowth but whose fronds did not overlap significantly (98.2% mapping accuracy for 7 year old plot of 776 palms). Stem heights could be predicted from point cloud derived metrics with root-mean-square errors (RMSEs) of 0.27 m (R2= 0.63) for 7 year old and 0.45 m (R2=0.69) for 10 year old palms. It was also found that an acquisition designed to yield the minimal required overlap between images (60%) performed almost as well as higher overlap acquisitions (>75%) for palm identification and basic height metrics which is promising for operational implementations seeking to maximise spatial coverage and minimise processing costs. We conclude that UAV-based SfM can provide reliable data not only for oil palm inventory generation but allows the retrieval of basic structural parameters which may enable per-palm above-ground biomass estimations.
Lethal Wilt is a limiting disease for oil palm cultivation in the eastern and central zones of Colombia. In the eastern zone, it caused the eradication of approximately 8700 ha of oil palm between 2010 and 2022, with economic losses of more than 185 million dollars. Studies conducted by Cenipalma reported that the pathogen causing this disease is Candidatus Liberibacter, which is possibly transmitted by Haplaxius crudus (Van Duzee). The adults feed on the foliage of the palms and move between them, spreading the pathogen in the plantation. A strategy to contribute to the management of LW is establishing cultivars resistant to the insect vector; however, no resistant cultivars or sources of resistance have been identified in the country's commercial cultivars or germplasm collections. Therefore, this work aimed to design and validate a methodology to characterize the oil palm genotypes Elaeis guineensis and Elaeis oleifera and interspecific OxG hybrids against adults of H. crudus , evaluating resistance through antixenosis and antibiosis to identify genotypes with possible sources of resistance. An arena with leaflets of the different genotypes in free-choice tests was used to assess antixenosis. For antibiosis, entomological sleeves were installed on the palm leaves, which were infested with adults of H. crudus from a breeding unit. The results of antixenosis and antibiosis in both the first phase (design) and the second phase (validation) indicated greater preference and survival for the genotypes of E. guineensis and lower preference and survival for the interspecific hybrids and E. oleifera . In the genotype E. guineensis , the average mortality was reached after 30 days, while in E. oleifera and the hybrids, it occurred between the third and fourth days. The results of this research provide a reproducible methodology for the evaluation of oil palm germplasms against H. crudus and sucking insects for the selection of sources of resistance for incorporation into breeding programs.
Abstract Background Automating stomatal trait measurement has gained popularity because of their inherent importance for field phenotyping application as stomata are critical for both carbon capture and water use efficiency in plants. Such tool has been reported for rice, wheat, tomato, barley and oil palm. However, none exist yet for canola, which is an important economic and agronomic crop globally. Results We developed a new toolkit called Stomatal Comprehensive Automated Neural Network or SCAN by combining the use of high-resolution portable digital microscopy with machine learning based on You Only Look Once algorithm (YOLOv8). Digital micrographs of leaf surfaces enter the SCAN pipeline, which includes stomata detection, stomata segmentation and stomatal pore segmentation models, to measure stomatal density, stomatal size and stomatal pore area, respectively. In addition to SCAN’s ability to measure leaf stomatal traits in canola at 89 to 94% accuracy, we also showed that SCAN can be used to predict stomatal density even in species not included in the training set such as Arabidopsis, tobacco, rice, wheat, maize and proso millet. SCAN was designed for the biological science community with the premise that users are not required to possess advanced programming capabilities to manage dependency prerequisites, execute the models, and integrate the analysis. This was achieved by packaging the models into a desktop application system that can be accessed offline. Conclusion Overall, SCAN provides a non-destructive, real-time, portable, and high-throughput measurement of leaf stomatal traits in canola. The minimised hardware requirement and user-friendly desktop application system make SCAN suitable for field phenotyping application.
Reproduction assets foundThe paper's authors publicly deposit the SCAN pipeline's model weights, hyperparameters, training scripts, and datasets in the FD_detection GitHub repository, and provide the SCAN application itself (with download and demonstration) in a second GitHub repository. Both are paper-specific, public, and actionable.Code · publicin Table S1.
123
124
The training tasks were carried out on an Ubuntu 20.04 Linux server at the Research School of Biology in
125
Australian National University, using two Nvidia A30 (24G) Graphic Processing Units (GPUs). The full
126
details of models’ weights, hyperparameters, training scripts and datasets can be found at
127
https://github.com/William-Yao0993/FD_detection.128
129
Model evaluation
130
131
Mean Average Precision (mAP, Fig. 3) and F1 score were used to assess model ability (Fig. 4). mAP is
132
calculated as the mean value of each class area under the precision-recall curve over thresholds, and the
133
F1 score is the harmonic mean of precision and recall. The formulas are deOpen asset ↗William-Yao0993/FD_detectionpdf-raw-page:4 lines:1-81Plant phenotyping relevance matchCrossref · checked 15 Sept 2026
Crop growth status detection is significant in agriculture and is vital in planting planning, crop yield, and reducing the consumption of fertilizers and workforce. However, little attention has been paid to detecting the growth status of each crop. Accuracy remains a challenging problem due to the small size of individual targets in the image. This paper proposes an object detection model, HR-YOLOv8, where HR means High-Resolution, based on a self-attention mechanism to alleviate the above problem. First, we add a new dual self-attention mechanism to the backbone network of YOLOv8 to improve the model’s attention to small targets. Second, we use InnerShape(IS)-IoU as the bounding box regression loss, computed by focusing on the shape and size of the bounding box itself. Finally, we modify the feature fusion part by connecting the convolution streams from high resolution to low resolution in parallel instead of in series. As a result, our method can maintain a high resolution in the feature fusion part rather than recovering high resolution from low resolution, and the learned representation is more spatially accurate. Repeated multiresolution fusion improves the high-resolution representation with the help of the low-resolution representation. Our proposed HR-YOLOv8 model improves the detection performance on crop growth states. The experimental results show that on the oilpalmuav dataset and strawberry ripeness dataset, our model has fewer parameters compared to the baseline model, and the average detection accuracy is 5.2% and 0.6% higher than the baseline model, respectively. Our model’s overall performance is much better than other mainstream models. The proposed method effectively improves the ability to detect small objects.
Oil and water content are an important quality criteria of crude palm oil (CPO) resulted from palm oil fruit processing. Those contents are usually determined using chemical method in the laboratory. This method is time consuming, long procedure, and destructive. Some efforts had been carried out to determine oil and water content of palm oil fruit non-destructively using some methods including Near-Infrared Spectroscopy (NIRS), but the results had not been satisfied. This research aims to assess Artificial Neural Network (ANN) and NIRS method to predict oil and water content of palm oil fruit’s non-destructively. The samples were palm oil fruits with ten maturity levels harvested from plantation in Bogor. Sample’s reflectance was measured with spectrometer NIR-Flex 500 at wavelength of 1000-2500 nm. After that, oil and water content were determined using chemical method. Some pre-treatments of NIR spectra namely normalization, savitzky-golay first derivative, their combinations, and standard normal variate were applied. Multivariate analysis such as PLS were carried out and the results of Factor Component (FC) were input for ANN model. The result showed the best method to predict oil content was combination savitzky-golay first derivative and normalization pre-treatment using PLS-ANN with 20 FC (R2=0.99; SEC=0,58%, RPD = 29.89; CV = 2.47%). For water content, the best prediction was standard normal variate pre-treatment using PLS-ANN with 20 FC (R2=0.99; SEC=1,07%, RPD=20.68; CV=1,73%). The result shows that developed ANN and NIRS can predict oil and water content of palm oil fruit non-destructively.
The efficient production of oil from oil palm trees is heavily dependent on their health status, reflected in the oil extraction rate (OER). The 17th frond of the oil palm trees contains a significant amount of organic compounds that directly influence the overall health of the tree. Achieving an optimal balance of essential nutrients such as nitrogen (N), phosphorus (P), and potassium (K) is crucial for classifying a tree as healthy, as it results in an increased oil to bunch and fruit to bunch ratio. To accurately assess the health level of oil palm trees, this study explores the application of Raman spectroscopy, a non-invasive technique in determining the molecular fingerprint of an organic sample. In this research, Raman spectroscopy is employed to determine the health level of oil palm trees, and a machine learning-based health level classification algorithm is developed. The algorithm analyzes the organic compounds found in oil palm leaves, which were collected from 20 different trees. The extracted spectral features from these leaves are used to classify them into two health levels: healthy and not healthy. For this purpose, 31 machine learning models are tested to identify the most accurate classifier. The findings reveal that the Tree and fine K-Nearest Neighbors (KNN) classifier demonstrates the highest overall accuracy of 95% using three significant features, namely the Raman intensity, Full Width at Half Maximum (FWHM), and area under the curve. This result signifies the potential of Raman spectroscopy as a reliable and promising method for non-invasively phenotyping oil palm leaves, enabling precise prediction of the health status of oil palm trees.
Oil palm is the most efficient oil-producing crop but its extension leads to increased deforestation in Southeast Asia. Oil palm height enables the quantitative estimation for carbon stock or palm oil yield. Nevertheless, there are still no accurate characterization of oil palm height providing information for the tradeoff between forest damage and carbon stock of oil palm in Southeast Asia. The new generation of spaceborne LiDAR provides large-extent canopy height samples, offering an opportunity for mapping the oil palm height at the regional scale yet with challenge to extrapolate the footprint heights to spatially coherent maps. Here, we proposed a new method by combining Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2) footprint data with stand age that is closely related to tree height. We first developed a semi-automatic filtering algorithm to filter the low-quality ICESat-2 data, and used a change detection algorithm to optimize the planting year map of oil palm. Then, an empirical age-height model was derived by linking ICESat-2 footprint canopy height with oil palm age that was used to estimate spatially and temporally oil palm height for the whole Peninsular Malaysia. A validation with independent ICESat-2 footprint data suggests a high agreement for the height estimates in 2020 from the age-height model (R² = 0.63; RMSE = 1.64 m) with a bias within ±3 m for >90% of the height estimates. Using the age-height model and planting year map of oil palm, we produced the first comprehensive wall-to-wall maps of long-term yearly oil palm height at a spatial resolution of 30 m in Peninsular Malaysia during 2001 through 2020. Our results suggest that the mean height of oil palm in all and regionally-disturbed areas have increased by 10.82 m and 9.29 m respectively during the last two decades in Peninsular Malaysia. Our results indicate that combining stand age and ICESat-2 footprint data has great potential in spatially-explicit mapping regional oil palm height that contributes to a better quantification for regional plantation carbon stock.
Oil palm is a high value crop with an estimated 5% yearly replanting rate. To dispatch high quality seeds, stringent culling upon the germinated seeds is necessary. The shape of germinated part of an oil palm seed is a key manual criterion for distinguishing good seeds from bad ones. Accurate segmentation of the germinated part would serve as an important preprocessing step for automatic phenotypic analysis and quality classification of germinated oil palm seeds. In this paper, we pioneer the study of semantic segmentation of germinated oil palm seeds by convolutional neural networks (CNNs). Leveraging the state-of-the-art ‘SE-ResNext + U-Net’ architecture for image segmentation, we propose two modifications to address the difficulty of accurately segmenting the germinated part of a seed since it is much smaller compared to the seed body. Firstly, we design local spatial channel attention (LS-SE) to replace the Squeez-Excitation (SE) module to retain local information of a feature channel. Then we pass the features generated by the encoder to the same level of the decoder part twice along the decoding direction (DC-UNet) to retain the original features. This helped address the problem where the edge segmentation details of oil palm seeds require higher-resolution detail information to improve the segmentation accuracy. In addition, the number of parameters of the proposed DC-UNet is much smaller than that of other state-of-the art U-Net variants such as Unet++, significantly reducing the training time. Our proposed DC-UNet with LS-SE obtained an MIOU that is 1.2% higher than U-Net, with a clearly better visual segmentation around the boundaries of the germinated parts.
Oil palm is a key agricultural resource in Malaysia. However, palm disease, most prominently basal stem rot caused at least RM 255 million of annual economic loss. Basal stem rot is caused by a fungus known as Ganoderma boninense . An infected tree shows few symptoms during early stage of infection, while potentially suffers an 80% lifetime yield loss and the tree may be dead within 2 years. Early detection of basal stem rot is crucial since disease control efforts can be done. Laboratory BSR detection methods are effective, but the methods have accuracy, biosafety, and cost concerns. This review article consists of scientific articles related to the oil palm tree disease, basal stem rot, Ganoderma Boninense , remote sensors and deep learning that are listed in the Web of Science since year 2012. About 110 scientific articles were found that is related to the index terms mentioned and 60 research articles were found to be related to the objective of this research thus included in this review article. From the review, it was found that the potential use of deep learning methods were rarely explored. Some research showed unsatisfactory results due to limitations on dataset. However, based on studies related to other plant diseases, deep learning in combination with data augmentation techniques showed great potentials, showing remarkable detection accuracy. Therefore, the feasibility of analyzing oil palm remote sensor data using deep learning models together with data augmentation techniques should be studied. On a commercial scale, deep learning used together with remote sensors and unmanned aerial vehicle technologies showed great potential in the detection of basal stem rot disease.
Background and aims Oil palms showing exceptional vigour and dubbed as 'giant palms' were identified in some progeny during breeding. A panel of phenotypical traits were studied to characterize these trees. The hypothesis that gigantism and other anomalies might be linked to polyploidy was investigated. Methods Twenty sib pairs of palms from different crosses, each comprising a giant and a normal oil palm, were studied by flow cytometry with rice 'Nipponbare' as standard reference. In parallel, palms were assessed in the field using 11 phenotypic traits. A principal component analysis (PCA) was conducted to define relationships between these phenotypical traits, and a linear discriminant analysis (LDA) to predict ploidy level and giant classification. Finally, a co-dominant molecular marker study was implemented to highlight the sexual process leading to the formation of 2n gametes. Key results The first group of oil palms presented an oil palm/rice peak ratio of around 4.8 corresponding to diploid oil palms, whereas the second group presented a ratio of around 7, classifying these plants as triploid. The PCA enabled the classification of the plants in three classes: 21 were normal diploid palms; ten were giant diploid palms; while 11 were giant triploid palms. The LDA revealed three predictors for ploidy classification: phyllotaxy, petiole size and circumference of the plant, but surprisingly not height. The molecular study revealed that triploid palms arose from 2n gametes resulting from the second division restitution of meiosis in parents. Conclusions This study confirms and details the process of sexual polyploidization in oil palm. It also identifies three phenotypical traits to assess the ploidy level of the giant oil palms in the field. In practical terms, our results provide a cheap scientific method to identify polyploid palms in the field.
Over the years, numerous studies have been conducted on the integration of computer vision and machine learning in plant disease detection. However, these conventional machine learning methods often require the contour segmentation of the infected region from the entire leaf region and the manual extraction of different discriminative features before the classification models can be developed. In this study, deep learning models, specifically, the AlexNet convolutional neural network (CNN) and the combination of AlexNet and support vector machine (AlexNet-SVM), which overcome the limitation of handcrafting of feature representation were implemented for oil palm leaf disease identification. The images of healthy and infected leaf samples were collected, resized, and renamed before the model training. These images were directly used to fit the classification models, without the need for segmentation and feature extraction as in models, without the need for segmentation and feature extraction as in the conventional machine learning methods. The optimal architecture of AlexNet CNN and AlexNet-SVM models were then determined and subsequently applied for the oil palm leaf disease identification.Comparative studies showed that the overall performance of the AlexNet CNN model outperformed AlexNet-SVM-based classifier.
Global biodiversity losses erode the functioning of our vital ecosystems. Functional diversity is increasingly recognized as a critical link between biodiversity and ecosystem functioning. Satellite earth observation was proposed to address the current absence of information on large-scale continuous patterns of plant functional diversity. This study demonstrates the inference and spatial mapping of functional diversity metrics through satellite remote sensing over a large key biodiversity region (Sabah, Malaysian Borneo, ~53,000 km2) and compares the derived estimates across a land-use gradient as an initial qualitative assessment to test the potential merits of the approach. Functional traits (leaf water content, chlorophyll-a and -b, and leaf area index) were estimated from Sentinel-2 spectral reflectance using a pre-trained neural network on radiative transfer modeling simulations. Multivariate functional diversity metrics were calculated, including functional richness, divergence, and evenness. Spatial patterns of functional diversity were related to land-use data distinguishing intact forest, logged forest, and oil palm plantations. Spatial patterns of satellite remotely sensed functional diversity are significantly related to differences in land use. Intact forests, as well as logged forests, featured consistently higher functional diversity compared to oil palm plantations. Differences were profound for functional divergence, whereas functional richness exhibited relatively large variances within land-use classes. By linking large-scale patterns of functional diversity as derived from satellite remote sensing to land-use information, this study indicated initial responsiveness to broad human disturbance gradients over large geographical and spatially contiguous extents. Despite uncertainties about the accuracy of the spatial patterns, this study provides a coherent early application of satellite-derived functional diversity toward further validation of its responsiveness across ecological gradients.
The implementations of deep learning combined with other methods such as transfer learning and data augmentation in oil palm fresh fruit bunch (FFB) ripeness classification have been researched throughout the years. However, most of the methods require devices with high computational resources which could not be implemented in mobile applications. To overcome this problem, this research would focus on creating a mobile application to classify the ripeness levels of oil palm FFB using lightweight Convolutional Neural Network (CNN). In this research, we implemented ImageNet transfer learning on 4 lightweight CNN models with a novel data augmentation method named “9-angle crop”, which would be further optimized using post-training quantization. Transfer learning with 3 unfrozen convolution blocks and 9 angle crop successfully increased the classification accuracy on MobileNetV1, and when it was compared to other lightweight models, EfficientNetB0 performed best with 0.898 test accuracy on Keras. Float16 quantization also proved to be the most suitable post-training quantization method for this model, halving the size of EfficientNetB0 with the least increase in image classification time and an accuracy drop of only 0.005 after the model was converted to TensorFlow Lite interpreter. In conclusion, the best model created in this research is EfficientNetB0, trained with the combination of transfer learning, 9 angle crop, and float16 quantization, which enabled it to achieve an overall test accuracy of 0.893 on TensorFlow Lite with 96 ms classification time per image, far surpassing the other 3 compared models with the second-best model being MobileNetV1 with 0.811 accuracy. The model itself was able to achieve similar results when it was implemented on an Android application to classify the ripeness levels of oil palm FFB images obtained through live camera input.
Drone is a UAV vehicle which is currently widely used in various activities, one of which is for aerial photography (photogrammetry). The concept of efficiency is the main goal of using this drone, which is to produce detailed and up-to-date aerial photo images with adjustable area coverage, relatively short time, affordable costs and minimal personnel required. The output is an orthophoto image that already has coordinates and can be used as primary data for convenience in the tree counting process so that tree populations in blocks based on design and area statements can be known in detail and accountably. The aerial mapping process using a copter unmanned vehicle with a height of 80-meters above ground level at an image resolution of 2.23 cm/pixel produces 3,795 photos with side overlap and front overlap photos of 70% and 80% with an area covering 2.72 km2 or 272 ha. The results of the calculation of oil palm trees as many as 3,147 pkk with a statement area of 29.08 ha in Block M06 and obtained an SpH of 108 pkk which is smaller than the ideal SpH ranging from 135-143 pkk/ha so that there is a need for compaction activities or plant fulfillment in the block.
Main conclusion Karyotyping using high-density genome-wide SNP markers identified various chromosomal aberrations in oil palm (Elaeis guineensis Jacq.) with supporting evidence from the 2C DNA content measurements (determined using FCM) and chromosome counts. Oil palm produces a quarter of the world's total vegetable oil. In line with its global importance, an initiative to sequence the oil palm genome was carried out successfully, producing huge amounts of sequence information, allowing SNP discovery. High-capacity SNP genotyping platforms have been widely used for marker-trait association studies in oil palm. Besides genotyping, a SNP array is also an attractive tool for understanding aberrations in chromosome inheritance. Exploiting this, the present study utilized chromosome-wide SNP allelic distributions to determine the ploidy composition of over 1,000 oil palms from a commercial F 1 family, including 197 derived from twin-embryo seeds. Our method consisted of an inspection of the allelic intensity ratio using SNP markers. For palms with a shifted or abnormal distribution ratio, the SNP allelic frequencies were plotted along the pseudo-chromosomes. This method proved to be efficient in identifying whole genome duplication (triploids) and aneuploidy. We also detected several loss of heterozygosity regions which may indicate small chromosomal deletions and/or inheritance of identical by descent regions from both parents. The SNP analysis was validated by flow cytometry and chromosome counts. The triploids were all derived from twin-embryo seeds. This is the first report on the efficiency and reliability of SNP array data for karyotyping oil palm chromosomes, as an alternative to the conventional cytogenetic technique. Information on the ploidy composition and chromosomal structural variation can help to better understand the genetic makeup of samples and lead to a more robust interpretation of the genomic data in marker-trait association analyses.
The information on biophysical parameters—such as height, crown area, and vegetation indices such as the normalized difference vegetation index (NDVI) and normalized difference red edge index (NDRE)—are useful to monitor health conditions and the growth of oil palm trees in precision agriculture practices. The use of multispectral sensors mounted on unmanned aerial vehicles (UAV) provides high spatio-temporal resolution data to study plant health. However, the influence of UAV altitude when extracting biophysical parameters of oil palm from a multispectral sensor has not yet been well explored. Therefore, this study utilized the MicaSense RedEdge sensor mounted on a DJI Phantom–4 UAV platform for aerial photogrammetry. Three different close-range multispectral aerial images were acquired at a flight altitude of 20 m, 60 m, and 80 m above ground level (AGL) over the young oil palm plantation area in Malaysia. The images were processed using the structure from motion (SfM) technique in Pix4DMapper software and produced multispectral orthomosaic aerial images, digital surface model (DSM), and point clouds. Meanwhile, canopy height models (CHM) were generated by subtracting DSM and digital elevation models (DEM). Oil palm tree heights and crown projected area (CPA) were extracted from CHM and the orthomosaic. NDVI and NDRE were calculated using the red, red-edge, and near-infrared spectral bands of orthomosaic data. The accuracy of the extracted height and CPA were evaluated by assessing accuracy from a different altitude of UAV data with ground measured CPA and height. Correlations, root mean square deviation (RMSD), and central tendency were used to compare UAV extracted biophysical parameters with ground data. Based on our results, flying at an altitude of 60 m is the best and optimal flight altitude for estimating biophysical parameters followed by 80 m altitude. The 20 m UAV altitude showed a tendency of overestimation in biophysical parameters of young oil palm and is less consistent when extracting parameters among the others. The methodology and results are a step toward precision agriculture in the oil palm plantation area.
Providing an accurate evaluation of palm tree plantation in a large region can bring meaningful impacts in both economic and ecological aspects. However, the enormous spatial scale and the variety of geological features across regions has made it a grand challenge with limited solutions based on manual human monitoring efforts. Although deep learning based algorithms have demonstrated potential in forming an automated approach in recent years, the labelling efforts needed for covering different features in different regions largely constrain its effectiveness in large-scale problems. In this paper, we propose a novel domain adaptive oil palm tree detection method, i.e., a Multi-level Attention Domain Adaptation Network (MADAN) to reap cross-regional oil palm tree counting and detection. MADAN consists of 4 procedures: First, we adopted a batch-instance normalization network (BIN) based feature extractor for improving the generalization ability of the model, integrating batch normalization and instance normalization. Second, we embedded a multi-level attention mechanism (MLA) into our architecture for enhancing the transferability, including a feature level attention and an entropy level attention. Then we designed a minimum entropy regularization (MER) to increase the confidence of the classifier predictions through assigning the entropy level attention value to the entropy penalty. Finally, we employed a sliding window-based prediction and an IOU based post-processing approach to attain the final detection results. We conducted comprehensive ablation experiments using three different satellite images of large-scale oil palm plantation area with six transfer tasks. MADAN improves the detection accuracy by 14.98% in terms of average F1-score compared with the Baseline method (without DA), and performs 3.55%-14.49% better than existing domain adaptation methods.
Reproduction assets foundThe authors explicitly state that their code and datasets (satellite images and annotations used for oil palm tree detection) are publicly available on GitHub.Code · publiche oil palm tree detection performance across different
remotely sensed images acquired from different sensors, regions and dates, without using labeled samples in the
target region. Our MADAN is proposed for enhancing both the generalization capacity and the transferability of
our model. Our codes and datasets are available on https://github.com/rs-dl/MADAN. The major contributions of
our work are as follows:
(1) We propose an adaptive object detector named MADAN for oil palm tree counting and detection across different
satellite images, which is the first work for large-scale domain adaptive tree crown detection using multi-source and
multi-temporal remote sensing images.
(2) WeOpen asset ↗rs-dl/MADANpdf-raw-page:6 lines:1-22Plant phenotyping relevance matchOpenAlex · Europe PMC · checked 9 Sept 2026
Ground-based LiDAR also known as Terrestrial Laser Scanning (TLS) technology is an active remote sensing imaging method said to be one of the latest advances and innovations for plant phenotyping. Basal Stem Rot (BSR) is the most destructive disease of oil palm in Malaysia that is caused by white-rot fungus Ganoderma boninense, the symptoms of which include flattening and hanging-down of the canopy, shorter leaves, wilting green fronds and smaller crown size. Therefore, until now there is no critical investigation on the characterisation of canopy architecture related to this disease using TLS method was carried out. This study proposed a novel technique of BSR classification at the oil palm canopy analysis using the point clouds data taken from the TLS. A total of 40 samples of oil palm trees at the age of nine-years-old were selected and 10 trees for each health level were randomly taken from the same plot. The trees were categorised into four health levels - T0, T1, T2 and T3, which represents the healthy, mildly infected, moderately infected and severely infected, respectively. The TLS scanner was mounted at a height of 1 m and each palm was scanned at four scan positions around the tree to get a full 3D image. Five parameters were analysed: S200 (canopy strata at 200 cm from the top), S850 (canopy strata at 850 cm from the top), crown pixel (number of pixels inside the crown), frond angle (degree of angle between fronds) and frond number. The results taken from statistical analysis revealed that frond number was the best single parameter to detect BSR disease as early as T1. In classification models, a linear model with a combination of parameters, ABD - A (frond number), B (frond angle) and D (S200), delivered the highest average accuracy for classification of healthy-unhealthy trees with an accuracy of 86.67 per cent. It also can classify the four severity levels of infection with an accuracy of 80 per cent. This model performed better when compared to the severity classification using frond number. The novelty of this research is therefore on the development of new approach to detect and classify BSR using point clouds data of TLS.
Basal stem rot (BSR) disease in oil palm is caused by Ganoderma boninense fungus. This plant disease is deemed highly destructive and would cause substantial economic loss. The use of spectroscopy technique with the capacity to deal with a large amount of spectral data has gained growing attention as a robust method, particularly to identify the symptoms of plant disease in its initial stage. The dimensionality reduction is pivotal in the use of spectroscopy technique due to its improved prediction performance and optimum processing. Considering that, this study assessed the feasibility of utilising dielectric spectral properties to classify the severity levels of BSR disease in oil palm across a frequency range of 100 kHz–30 MHz. The support vector machine-feature selection (SVM-FS) and principal component analysis (PCA) were applied as data reduction methods. After selecting the optimum number of significant frequencies, this study proceeded to assess the effectiveness of the support vector machine (SVM) and quadratic discriminant analysis (QDA) classifiers in identifying the four different levels of BSR disease. The performance of both classifiers with and without data reduction methods was subsequently compared in terms of the classification accuracy, while the whole spectrum data served as part of the control method. The resultant outcomes revealed that the use of QDA classifier with PCA recorded the highest classification accuracy (up to 96.36%). As for the case of without using data reduction methods, the SVM classifier recorded the highest classification accuracy at only 79.55%. Conclusively, this study proved the significance of dimensionality reduction of dielectric spectral data for the classification of BSR disease in oil palm.
Basal stem rot disease caused by the white‐rot fungus Ganoderma boninense is a major threat to the oil palm industry, and hence, the ability to detect infections at an early stage of development is desired. In this study, a headspace solid‐phase microextraction (HS‐SPME) method coupled with gas chromatography–mass spectrometry (GC‐MS) was employed to analyse the volatile organic compounds (VOCs) released from G. boninense cultures and infected oil palm wood. We examined VOCs released from three types of samples: G. boninense mycelium, oil palm wood and oil palm wood colonized by G. boninense. This preliminary study led to the tentative identification of 57 VOCs, including alcohols, alkanes, volatile acids, ketones, aldehydes, esters, sesquiterpenes and polycyclic aromatic hydrocarbon groups. Aliphatic compounds with eight‐carbon atoms, such as 1‐octen‐3‐ol, 3‐octanone, 1‐octanol and (E)‐2‐octenal, were the most abundant constituents of the Ganoderma samples, whereas furfural and hexanal were the major compounds detected in the oil palm wood samples. Chemometric analyses using cluster heat maps and principal component analyses were used to discriminate between the VOC profiles. The results indicated that the novel method described here could be used to detect Ganoderma disease and, more generally, for chemoecological studies of plant–pathogen interactions.
Abstract Transpiration at the stand level is often estimated from water use measurements on a limited number of plants and then scaled up by predicting the remaining plants of a stand by plant size‐related variables. Today, drone‐based methods offer new opportunities for plant size assessments. We tested crown variables derived from drone‐based photogrammetry for predicting and scaling plant water use. In an oil palm agroforest and an oil palm monoculture plantation in lowland Sumatra, Indonesia, tree and oil palm water use rates were measured by sap flux techniques. Simultaneously, aerial images were taken from an octocopter equipped with an Red Green Blue (RGB) camera. We used the structure from motion approach to compute several crown variables such as crown length, width, and volume. Crown volumes for both palms (69%) and trees (81%) explained much of the observed spatial variability in water use; however, the specific crown volume model differed between palms and trees and there was no single linear model fitting for both. Among the trees, crown volume explained more of the observed variability than stem diameter, and in consequence, uncertainties in stand level estimates resulting from scaling were largely reduced. For oil palms, an appropriate whole‐plant size‐related predictor variable was thus far not available. Stand level transpiration estimates in the studied oil palm agroforest were lower than those in the oil palm monoculture, which is probably due to the small‐statured trees. In conclusion, we consider drone‐derived crown metrics very useful for the scaling from single plant water use to stand‐level transpiration.
The vast size of oil palm (Elaeis guineensis) plantations has led to lightweight unmanned aerial vehicles (UAVs) being identified as cost effective tools to generate inventories for improved plantation management, with proximal aerial data capable of resolving single palm canopies at potentially, centimetric resolution. If acquired with sufficient overlap, aerial data from UAVs can be processed within structure-from-motion (SfM) photogrammetry workflows to yield volumetric point cloud representations of the scene. Point cloud-derived structural information on individual palms can benefit not only plantation management but is also of great environmental research interest, given the potential to deliver spatially contiguous quantifications of aboveground biomass, from which carbon can be accounted. Using lightweight UAVs we captured data over plantation plots of varying ages (2, 7 and 10 years) at peat soil sites in Sarawak, Malaysia, and we explored the impact of changing spatial resolution and image overlap on spatially variable uncertainties in SfM derived point clouds for the ten year old plot. Point cloud precisions were found to be in the decimetre range (mean of 26.7 cm) for a 10 year old plantation plot surveyed at 100 m flight altitude and >75% image overlap. Derived canopy height models were used and evaluated for automated palm identification using local height maxima. Metrics such as maximum canopy height and stem height, derived from segmented single palm point clouds were tested relative to ground validation data. Local maximum identification performed best for palms which were taller than surrounding undergrowth but whose fronds did not overlap significantly (98.2% mapping accuracy for 7 year old plot of 776 palms). Stem heights could be predicted from point cloud derived metrics with root-mean-square errors (RMSEs) of 0.27 m (R2 = 0.63) for 7 year old and 0.45 m (R2 = 0.69) for 10 year old palms. It was also found that an acquisition designed to yield the minimal required overlap between images (60%) performed almost as well as higher overlap acquisitions (>75%) for palm identification and basic height metrics which is promising for operational implementations seeking to maximise spatial coverage and minimise processing costs. We conclude that UAV-based SfM can provide reliable data not only for oil palm inventory generation but allows the retrieval of basic structural parameters which may enable per-palm above-ground biomass estimations.
The paradigm of functional-structural models (FSPM) assumes that studying the detailed organisation of plant structure allows a better understanding of functional processes; in particular the way plants capture light for performing photosynthesis. However, much attention must be paid toward the consistency between virtual plants and plants in the field in terms of size and geometry to accurately evaluate light interception. This paper thus aimed at i) assessing the capacity of a 3D architectural model based on oil palms (Elaeis guineensis) to accurately represent plants structural characteristics at both the scale of the individual plant and the cultivated plot and ii) employing the validated 3D mock-ups to investigate how light interception efficiency varies among progenies that exhibit different architectures. Innovative indicators related to plant geometry and topology were derived from terrestrial LiDAR scanners (TLS) and hemispherical photographs (HP) in order to assess a 3D plant model. Indicators such as plant height, width and volume, gap fractions and solid angle projections were established from field measurements and were compared to equivalent indicators that had been extracted from virtual TLS (VTLS) and virtual HP (VHP) simulated on 3D mock-ups. Indicators were then evaluated for their significance in terms of light interception. Progeny effect on light interception efficiency was finally evaluated for five progenies.The structural indicators estimated from VTLS and VHP were significantly correlated with equivalent indicators estimated from TLS and HP, respectively, and with simulated outputs related to light interception. Light interception efficiencies estimated from validated 3D mock-ups differed significantly among the five progenies under study, most notably along plant development.Our results highlight the relevance of combining TLS- and HP-derived indicators to evaluate the reliability of virtual 3D reconstruction of plants in relation to light capture, at both the plant and plot scales. The study paves the way for further investigations aiming at unravelling the relationships between oil palm architecture and the physiological processes driving its production.
Basal stem rot (BSR) is a prominent plant disease caused by Ganoderma boninense fungus, which infects oil palm plantations leading to large economic losses in palm oil production. There is need for novel disease detection techniques that can be used to reduce the oil palm losses due to BSR. Thus, this paper investigated the feasibility of utilizing electrical properties such as impedance, capacitance, dielectric constant, and dissipation factor in early detection of BSR disease in oil palm tree. Leaf samples from different oil palm trees (healthy, mild, moderate, and severely-infected) were collected and measured using a solid test fixture (16451B, Keysight Technologies, Japan) connected to an impedance analyzer (4294A, Agilent Technologies, Japan) at a frequency range of 100 kHz–30 MHz with 300 spectral interval. Genetic algorithm (GA), random forest (RF), and support vector machine-feature selection (SVM-FS) were used to analyze the electrical properties of the dataset and the most significant frequencies were selected. Following the selection of significant frequencies, the features were evaluated using two classifiers, support vector machine (SVM) and artificial neural networks (ANN) to determine the overall and individual class classification accuracies. The selection model comparative feature analysis demonstrated that the best statistical indicators with overall accuracy (88.64%), kappa (0.8480) and low mean absolute error (0.1652) were obtained using significant frequencies produced by SVM-FS model. The results indicated that the SVM classifier shows better performance as compared to ANN classifier. The results also showed that the classes, features selection models, and the electrical properties were found to be significantly different (p < .1). The impedance values were highly classified by Ganoderma disease at different levels of severity with overall accuracies of more than 80%. Impedance can be considered as the best electrical properties that can be used to estimate the severity of BSR disease in oil palm using spectroscopy technique. As such, this study demonstrates the potentials of utilizing electrical properties for detection of Ganoderma diseases in oil palm.
Three-dimensional (3D) reconstruction of plants is time-consuming and involves considerable levels of data acquisition. This is possibly one reason why the integration of genetic variability into 3D architectural models has so far been largely overlooked. In this study, an allometry-based approach was developed to account for architectural variability in 3D architectural models of oil palm (Elaeis guineensis Jacq.) as a case study. Allometric relationships were used to model architectural traits from individual leaflets to the entire crown while accounting for ontogenetic and morphogenetic gradients. Inter- and intra-progeny variabilities were evaluated for each trait and mixed-effect models were used to estimate the mean and variance parameters required for complete 3D virtual plants. Significant differences in leaf geometry (petiole length, density of leaflets, and rachis curvature) and leaflet morphology (gradients of leaflet length and width) were detected between and within progenies and were modelled in order to generate populations of plants that were consistent with the observed populations. The application of mixed-effect models on allometric relationships highlighted an interesting trade-off between model accuracy and ease of defining parameters for the 3D reconstruction of plants while at the same time integrating their observed variability. Future research will be dedicated to sensitivity analyses coupling the structural model presented here with a radiative balance model in order to identify the key architectural traits involved in light interception efficiency.
Oil palm plantations consist of diverse plant density level that influence the appearance of soil surface or commonly in remote sensing terms called as soil background.Choosing the right density coefficient of vegetation transformation can decrease the noise of soil background for estimating oil palm yield.This research aims 1) to examine the accuracy of SPOT-6 to identify the oil palm l plant growth level and to estimate their yield 2) to know the variation of oil palm yield based on SAVI index vegetation using different density coefficient; and 3) to determine the best density coefficient to estimate the yield of oil palm.This research was held in part of Air Molek, Indragiri Hulu Regency, Riau, one of the largest oil palm plantations in Indonesia.This research method utilises SAVI transformation with density coefficient L-0, L-0.5, and L-1, and regression statistics analysis.The land-cover primary data is derived from SPOT-6 imagery archived in 13 rd June 2013.The field survey was conducted in the same month of image's acquisition time and 120 sample areas were taken during that time.Two steps of regression analyses were applied to see the correlation between, first, vegetation index value and oil palm plant; and second, oil palm plant, vegetation index values, and oil palm yield from field observation.These steps produced a model to estimate the oil palm yield based on the index values of yield and vegetation, and the productivity estimation.The result shows that SPOT-6 imagery has 96% accuracy level which is considered high for identifying the oil palm variation.The R value for L-0 density coefficient is 0.8, for L-0.5 is 0.81 whereas for L-1 is 0.82.The best plant's density coefficient for estimating oil palm yield/yield is L-0 with yield estimation accuracy of 83.33%.
In this study, the oil palm fresh fruit bunch (FFB) was harvested and its images were recorded in a photographic studio. The bunch was recorded from various distances (2, 7, 10, 15m) using five lighting configuration, i.e. ultraviolet lamp (600 watts), visible lamp (600 and 1000 watt), as well as IR lamp (600 and 1000 watts). The FFB images were processed in order to obtain 15 colour components making up the image, consist of three primary colours (R, G, B) and their transformations (H, S, I, RI, GI, BI, RG, RB, GB, GR, BR, BG). The prediction model of FFB’s oil content was built to evaluate the amount of oil on FFB accurately based on its image. The model was built using deep neural networks, where the colour components served as inputs, and 10 hiden layers were introduced to describe the relationship between all these variables and oil content. Of various recording setup, only four were selected based on their coefficient of correlation, namely: 10m_UV (R2 = 1); 10m_Vis2 (R2 = 1); 10m_IR2 (R2 = 1); and 2m_IR2 (R2 = 0.981). The determinant colour of the FFB’s image which mostly influence the prediction models were the ratio of R to B (RB) for 10m_UV; the value of H and S on 10m_Vis2; the I and S of the 10m_IR2; and RB, H, and B for the 2m_IR2 treatment. Keywords: deep neural network, FFB, nondestructive, oil content, photogrammetry