A pulse crop dataset of agronomic traits and multispectral images from multiple environments

Authors: UMANI, KINGSLEY - Washington State University; ZHANG, CHONGYUAN - Washington State University; SANJAN, WORASIT - Washington State University; McGee, Rebecca; Vandemark, George; SANKARAN, SINDUJA - Washington State University

Submitted to: Data in Brief
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/21/2023
Publication Date: 12/26/2023
Citation: Umani, K., Zhang, C., Sanjan, W., Mcgee, R.J., Vandemark, G.J., Sankaran, S. 2023. A pulse crop dataset of agronomic traits and multispectral images from multiple environments. Data in Brief. 53:110013. https://doi.org/10.1016/j.dib.2023.110013.
DOI: https://doi.org/10.1016/j.dib.2023.110013

Interpretive Summary: Unpiloted aerial vehicles (UAV) take pictures of research field plots that can be used by plant breeders to rapidly and accurately evaluate large numbers of plants for specific traits such as yield, disease resistance, and early maturity. Typically, applying these technologies initially involves a cooperation between plant breeders and experts in remote sensing. Plant breeders measure traits of interest in the field to determine the "ground truth" values for each trait, while remote sensing experts take photos of the plots. The breeder and remote sensing experts work together to determine how accurately photos taken by UAV can predict plant traits. This paper provides more than 275 photos taken of USDA pea and chickpea breeding plots over 2017-2019 at several locations in Washington and Idaho. Ground truth data for several important field traits including yield, seed size, and days to mature,m are also provided for all plots. These resources can be used to teach scientists how to process and interpret field photos taken by UAV, and how to accurately predict the performance of breeding lines and varieties for selected traits based on photographic images. This is important for increasing capacity across the United States in the use of remote sensing for plant breeding.

Technical Abstract: Crop yield potential in breeding trials can be captured using unmanned aerial vehicle (UAV) based multispectral imagery. Several digital traits or phenotypes such as vegetation indices can represent canopy crop vigor and overall plant health, which can be used to evaluate differences in performance across varieties in crop breeding programs. This dataset contains agronomic data for named cultivars and breeding lines of dry pea and chickpea, with over 275 multispectral images from advanced and preliminary breeding trials. The breeding trials were located at three locations in the “Palouse” region of Eastern Washington and Northern Idaho of the United States across 2017, 2018 and 2019 cropping seasons. The multispectral images were captured at multiple time points using a UAV flight integrated with 5-band camera from early growth through pod development growth stages during each cropping season. This dataset details seed yield information from trials of dry pea and chickpea that were obtained from each location, as well as additional agronomic and phenological data recorded at one location (mostly Pullman, WA) for each cropping season. The dataset also includes 20-78 megabytes (MB) Tagged Image Format (TIF) uncalibrated stitched orthomosaic images generated from the photogrammetric software. The images can be processed using any convenient image processing algorithm to obtain vegetation indices and other useful information. Zhang et al. [1] and Marzougui et al. [2] demonstrated the usage of this dataset.