High-throughput Phenotyping as an Auxiliary Tool for Watermelon Germplasm Accession Selection

Abstract

Abstract Watermelon cultivation plays an important socioeconomic role in Brazil. Despite all efforts, in genetic improvement programs, selection often involves a limited number of accessions due to complex cultivation requirements for this species, including large evaluation areas. The requirement of large areas results in increased costs and time of evaluations using conventional methodologies. In this context, we assessed the use of imaging to optimize resources and field evaluation time. The study aimed to use image phenotyping with various vegetation indices for watermelon germplasm. In this study, we evaluated 118 watermelon genotypes. In-field manual measurements included the SPAD index of leaves and average Brix and weight of fruits. To determine the potential use of imaging in phenotyping, four vegetation indices —NGRDI, GLI, SAVI, and NDVI — were studied. The germplasm exhibited genetic dissimilarity and the agronomic performance was monitored using images. The NGRDI, NDVI, SAVI and GLI indices remarkably correlated with the BRIX variable and SPAD index, even in germplasms with high dissimilarity. In addition, remote sensing was used to monitor field parameters that were not directly related to the leaf. The NGRDI, NGRDI and SAVI indices were sensitive in capturing this indirect relationship with the BRIX variable.

Code and data availability

The article describes UAV image phenotyping of 118 watermelon genotypes (vegetation indices vs. SPAD, Brix, fruit weight) and multivariate analyses, but contains no data availability statement, no public dataset or image deposit, and no author code/workflow URL. Software mentioned (ENVI Classic, ArcMap, MiniTab, Genes)

No evidence-backed public reproduction asset is currently recorded.