Detecting animals, especially medium to large-sized animals, in agriculture fields before running harvester machines is crucial to protecting wildlife from unnecessary deaths. This is not only an issue of protecting wildlife but also of avoiding contaminating the crop harvest. Typically, animals like deer behave problematically when they see or hear an agricultural vehicle approaching. Instead of fleeing, they freeze and hide, ultimately resulting in getting run over by the combine harvester vehicles. Currently, the only solution to this problem is to scan the field on foot and scare away the animals.

Our solution to this problem is an intelligent sonar based on a wide-band sensing system installed on a drone. The sonar is based on a prototype we developed, "Agry-Sonar" for yield estimation in cherry tomatoes. We have shown successfully that this sonar can 'see through' corn stalks and pepper plants, identifying the next row or a hidden object up to 5m away. "Agry-Sonar" uses a wide-band sonar chirp signal and a machine-learning algorithm that can perform classification from the echo.

The sonar sensing system will be integrated into a purpose-built drone to efficiently scan the field before operating harvester machines. IMS-ZHAW will incorporate the novel sonar sensor package from Smart-Pick into a drone platform designed in-house suitable for conducting wildlife surveys in agricultural fields. The drone platform will be developed to integrate all sensing hardware and flight components into a compact and robust package, allowing simple operation in agricultural fields.

A primary goal of this project is to conduct the necessary tests to validate the proof-of-concept system and the feasibility of this novel idea. We will use the results of this preliminary study as a basis to plan a subsequent innovation project.