Detecting volunteer cotton plants in a corn field with deep learning on UAV remote-sensing imagery

Volunteer cotton (VC) plants growing in the fields of inter-seasonal and rotated crops, like corn, can serve as hosts to boll weevil pests once they reach pin-head square stage (5–6 leaf stage). The VC therefore need be detected, located, destroyed or sprayed. In this paper, we present a study on using deep learning (DL) detect corn field RGB images collected with an unmanned aerial vehicle (UAV). objectives were (i) determine whether YOLOv3 DL algorithm could used for detection based UAV-derived images, (ii) investigate behavior at three different pixel scales (320 × 320, S1; 416 416, S2; 512 512, S3). metrics evaluate results average precision (AP), mean (mAP) F1-score 95 % confidence level. It was found that able accuracy more than 80 %, 78.5 mAP 80.38 %. With respect size, no significant differences existed among scales, but difference AP between S1 S3 (p = 0.04) S2 0.02). A also overall goal minimize pest infestation by maximizing true positive which is represented values. lack these all indicated trained model irrespective input image sizes. capability demonstrates potential algorithms real-time mitigation computer vision spot-spray capable UAV.