Cashew Trees Detection And Yield Analysis Using UAV-Based Map

摘要

In this study we developed a novel method to detect cashew trees in an orthophoto map derived from images collected by an unmanned aerial vehicle (UAV). We also suggest a way in which these detections can be used to analyze the yield of the cashew farm. The proposed method uses image analysis to find the tops of trees, to merge different tops located on the same tree, and to segment individual tree. The segmented trees are used in a deep learning framework to know the exact location of cashew trees. The preliminary cashew detection from UAV-based map is promising. This study can be interesting for developing countries where UAV system are nowadays gaining popularity in agriculture. Our method does not require any additional sensor other than the RGB camera onboard the UAV. This low-cost solution is suitable for small and medium cashew farmers. The developed method can also be extended to other types of trees, other than the cashew.