Radishes, which are common root vegetables, are rich in vitamins and minerals, and contain low calories. This vegetable is known for its rapid growth. Nevertheless, the variety of leaf diseases where leaves get affected by various bacterial and fungal diseases can hinder the healthy growth of radish. Furthermore, there is a high risk of inaccurate identification of diseases if the farmers try to use traditional methods in recognizing these diseases. With the purpose of precise identification of radish leaf diseases for the finest growth of this vegetable, total of 2801 images of the radish leaves are collected from vegetable field in Bangladesh. The collected dataset includes comprehensive images of healthy leaves as well as four types of leaf affected by various diseases such as Black Leaf Spot, Downey Mildew, Flea Beetle and Mosaic. Utilizing this robust dataset, deep learning models can be trained to identify the leaf diseases which helps to detect the diseases in order to reduce the harm of the cultivation of radish. By identifying the diseases on radish leaves accurat-ely and maintaining healthy production of radish, this dataset contributes to the broader sustainability in the agricultural sector.
Reproduction assets foundThe paper is a Data in Brief article describing a public Mendeley Data repository of 2801 smartphone images of radish leaves (healthy plus four disease classes) collected in Bangladesh, which is the paper's own phenotyping image dataset and is directly accessible.Dataset · publicortant role for classifying the radish plant healthy and unhealthy leaves.
Data source location
1. Vegetable field of Kathalkandi, Nasirnagar, Brahmanbaria, Bangladesh (latitude: 24.1915°, longitude: 91.1826°)
Data accessibility
Repository name: Mendeley Data
Data identification number: 10.17632/s973cz2jcd.1
Direct URL to data: https://data.mendeley.com/datasets/s973cz2jcd/1
1
Value of the Data
•
The dataset containing several classes of radish leaves where each class clearly representing the unhealthy leaf as well as healthy leaf. All the images are captured with high resolution that ensuing the high-quality of leaves images, helps to recognize the pattens of diseases.
•
The dataset presentOpen asset ↗Mendeley Data · 10.17632/s973cz2jcd.1lines:1-50Code / dataset availability confirmedCrossref · OpenAlex · checked 15 Sept 2026
Cover crops (CC) immobilize mineral soil N in their biomass, preventing N losses during crop rotation intervals. As the CC biomass is incorporated into the soil and decomposes, N is released for the following main crop. The efficiency of CC N uptake and release depends on CC quantity and quality, which can be enhanced in mixtures. Traditional N uptake measurements are labour-intensive and limited in capturing spatial variability. We calibrated relationships between traditional measurements and multispectral data from an Unmanned Aerial Vehicle (UAV) to quantify CC traits with minimal disturbance and high spatial resolution in both monocultures and mixtures. This innovative approach combined vegetation indices, textural features, and a photogrammetry-derived canopy surface model to predict CC traits. Linear models were trained for biomass, N uptake, and C:N predictions, while a K-Nearest-Neighbour model was trained for N concentration. When evaluated on the test set, the calibrated remote sensing models accurately predicted CC aboveground biomass (R 2 : 0.71, RMSE: 287.1 kg/ha, NRMSE: 11.74 %), N concentration (R 2 : 0.80, RMSE: 1.77 gN /kg, NRMSE: 6.96 %), N uptake (R 2 : 0.56, RMSE: 9.38 kgN /ha, NRMSE: 15.08 %), and C:N ratio (R 2 : 0.62, RMSE: 1.86, NRMSE: 10.98 %). The field experiment included monocultures, bi-, and tri-species mixtures of common vetch ( Vicia sativa ), black oat ( Avena strigosa ), and fodder radish ( Raphanus sativus ). N uptake was similar between treatments, yet the CC species differed in strategies, producing high biomass with low N concentration or vice versa. This study provides a basis for spatially predicting key CC traits using UAV optical data.
Reproduction assets foundThe paper's Data availability statement explicitly states the authors' R code for image processing, model training, and figure production is publicly available on the authors' WUR GitLab repository (uav4covercroptraits). No phenotype dataset or image deposit is stated separately.Code · publictal for the
UAV data acquisition.
Supplementary materials
Supplementary material associated with this article can be found, in
the online version, at doi:10.1016/j.atech.2024.100608.
Data availability
The R code generated during this study to process the images, train
the models and produce the figures, is publicly available at https://git.wur.nl/dall002/uav4covercroptraits.References
[1] C. Aita, S.J. Giacomini, Crop residue decomposition and nitrogen release in singleOpen asset ↗git.wur.nl/dall002/uav4covercroptraitspdf-raw-page:10 lines:1-89Code / dataset availability confirmedEurope PMC · checked 7 Sept 2026
RadishRiceX-ray / CTRoot2D/3D reconstructionGrowth / time-series analysisRoot system architecture
Crop genetic engineering for better root systems can offer practical solutions for food security and carbon sequestration; however, soil layers prevent the direct visualization of plant roots, thus posing a challenge to effective phenotyping. Here, we demonstrate an original device with a distributed fiber-optic sensor for fully automated, real-time monitoring of underground root development. We show that spatially encoding an optical fiber with a flexible and durable polymer film in a spiral pattern can significantly enhance sensor detection. After signal processing, the resulting device can detect the penetration of a submillimeter-diameter object in the soil, indicating more than a magnitude higher spatiotemporal resolution than previously reported with underground monitoring techniques. Additionally, we also developed computational models to visualize the roots of tuber crops and monocotyledons and then applied them to radish and rice to compare the results with those of X-ray computed tomography. The device's groundbreaking sensitivity and spatiotemporal resolution enable seamless and laborless phenotyping of root systems that are otherwise invisible underground.
Reproduction assets foundThe authors publicly provide MATLAB code for virtual root reconstruction and the sample datasets used in the study via their GitHub repository Fiber-RADGET, with explicit availability statements in the Methods and Data availability sections.Code · publicThe custom code for the virtual root reconstruction in MATLAB (MathWorks, Massachusetts, USA) is available at https://github.com/mtei1/Fiber-RADGET.git .Open asset ↗mtei1/Fiber-RADGETlines:126-223Code / dataset availability confirmedEurope PMC · bioRxiv · checked 7 Sept 2026
RadishRiceX-ray / CTRootObject detection2D/3D reconstructionGrowth / time-series analysisVisualization / data managementRoot system architecture
Crop genetic engineering for better root systems can offer practical solutions for food security and carbon sequestration; however, soil layers prevent direct visualization. Here, we demonstrate an original device with a distributed fiber-optic sensor for fully automated, real-time monitoring of underground root development. We demonstrate that spatially encoding an optical fiber with a flexible and durable polymer film in a spiral pattern can significantly enhance sensor detection. After signal processing, the resulting device can detect the penetration of a submillimeter-diameter object in the soil, indicating more than a magnitude higher spatiotemporal resolution than previously reported with underground monitoring techniques. We also developed computational models to visualize the roots of root crops and monocotyledons, and then applied them to radish and rice to compare the results with those of X-ray computed tomography. The device’s groundbreaking sensitivity and spatiotemporal resolution enable seamless and laborless phenotyping of root systems that are otherwise invisible underground.
Reproduction assets foundThe paper's custom MATLAB code for virtual root reconstruction from fiber-optic strain data is explicitly stated to be publicly available on the authors' GitHub repository (Fiber-RADGET). No separate public phenotype dataset deposit is mentioned; the supplementary movie is not a qualifying dataset URL.Code · publicThe
custom code for the virtual root reconstruction in MATLAB
(MathWorks, Massachusetts, USA) is available at
https://github.com/mtei1/Fiber-RADGET.git.Open asset ↗mtei1/Fiber-RADGETpdf-page:12 lines:1-24