Development of mobile applications for crop scouting with small unmanned aircraft systems

Abstract

Small unmanned aircraft systems (sUAS) have been in commercial use since the1980’s and over 8-12% of its current uses are in the agricultural sector, but only involving limited uses like surveying, mapping and imaging, which is expected to increase to 47% according to AUVSI with the association of Artificial Intelligence over the next decade. Our research is one such effort to help farmers utilize advanced sUAS technology coupled with Artificial Intelligence and give them meaningful results in a widely used and user friendly interface, like a mobile application. The vision for this application is to provide a completely automated experience to the farmer for a repetitive and periodic analysis of his/her crops where all the instruction needed from the farmer is a push of a button on a one time configured application and ultimately providing results in seconds. This would help the farmer scout their crops, assess yield potential, and determine if additional inputs are needed for increasing grain yield and profit per acre. For making this application we focused on user-friendliness by abstracting crop algorithms, minimized necessary user inputs, and automate the construction of flight paths. Due to internet connection not always being available at farm fields, processing was kept to on-board compute systems and the mobile device to give live results to farmers without reliance on cloud-based analytics. The application is configured to work with DJI Aircraft using OpenCv for video processing and mobile vision, GIS and GPS data for accurate mapping, locating device, sUAS on the mobile application, and FFMPEG for encoding and decoding compressed video data. An algorithm developed by Precision-Ag Lab at the K-State Agronomy Department was implemented into the sUAS application for providing real time yield estimations and nitrogen recommendation algorithm for winter wheat.