Crop scouting using UAV imagery: a case study for potatoes

Abstract

Crop scouting is essential to manage crops such as potatoes and to detect stresses. In fact, conventional approaches require a lot of time and staff. The rise of unmanned aerial vehicle imagery offers interesting perspectives in this area. However, the development of interpreted and generalizable cartographic products that can be used directly by producers still poses challenges in terms of processing complexity and production time. The purpose of this study was to develop a tool for the phytosanitary surveillance of potato crops using scouting maps in support of conventional methods. The approach is based on relatively simple and efficient preprocessing methods when no radiometric correction data are available. The first step establishes the best correlations between some biophysical parameters directly related to phytosanitary problems and vegetation indices in the visible and infrared domains. The second step allows the development of an approach to classify the following three stresses: pests, diseases, and development problems. Validated by scouting field sites, the developed approach makes it possible to quickly produce accurate scouting maps that producers can use directly thanks to their high potential for generalization to other areas and crop productions.