|FENG, A. - University Of Missouri|
|ZHOU, J. - University Of Missouri|
|Sudduth, Kenneth - Ken|
|ZHANG, M. - Jiangsu Academy Agricultural Sciences|
Submitted to: Biosystems Engineering
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/20/2020
Publication Date: 3/8/2020
Citation: Feng, A., Zhou, J., Vories, E.D., Sudduth, K.A., Zhang, M. 2020. Yield estimation in cotton using UAV-based multi-sensor imagery. Biosystems Engineering. 193:101-114. https://doi.org/10.1016/j.biosystemseng.2020.02.014.
Interpretive Summary: Remote sensing can provide an efficient way to obtain information about spatial variation across fields and landscapes. Recent availability of unmanned aerial vehicle (UAV) technology has created interest in developing new remote sensing applications for use in precision agriculture. Within-season estimation of crop yield would allow farmers to make management and marketing decisions before harvest, potentially improving their profitability. In this project, data from multiple image sensors mounted on a UAV were investigated as an estimator of cotton yield in a research field in Southeast Missouri. While good results were obtained with just a single feature calculated from the images, better results were obtained using a combination of three features. The tradeoff between increased accuracy and the higher cost for obtaining additional features would need to be evaluated for specific applications. This study has demonstrated the potential of using UAV-based imaging to estimate within-field yield variation in precision agriculture. This approach may be useful to researchers and to farmers who are interested in obtaining within-season yield estimations at relatively low cost and high resolution.
Technical Abstract: Timely monitoring of crop development and accurate estimation of yield is important to improve field management and crop production. This study aimed to evaluate the performance of an unmanned aerial vehicle (UAV)-based remote sensing system in cotton yield estimation. The UAV system equipped with a RGB camera, a multispectral camera and an infrared thermal camera was used to acquire images of a cotton field in two growth stages. The sequential images from the three cameras were processed to generate georeferenced orthomosaic images and a digital elevation model (DEM) of the cotton field, which were registered to the georeferenced yield data acquired by a yield monitor mounted on a harvester. Eight image features were extracted from the orthomosaic images and the DEM, including normalized difference vegetation index (NDVI), green normalized difference vegetation index (GNDVI), triangular greenness index (TGI), a* channel in CIE-LAB color space (a*), canopy cover, plant height, canopy temperature and cotton fiber index (CFI). A correlation analysis was conducted between yield and image features. Models were developed to evaluate the accuracy of each image feature for yield estimation. Results show that plant height and CFI were the best single features for cotton yield estimation, both with R2 = 0.90. The combination of plant height and CFI, plant height and a*, or plant height and temperature were the best two-feature models with R2 from 0.92 to 0.94. The best three-feature models were among the combinations of plant height, CFI, temperature and a*, which indicated that those four features were the most important ones for the estimation of cotton yield. This study found that image features from UAV-based multiple sensors were able to estimate cotton yield accurately.