A Multi-Target Tracking and Positioning Technology for UAV Based on Siamrpn Algorithm

Abstract

UAV s play a pivotal role in the field of security due to their flexibility, high efficiency, and low cost. This article uses yolov4 convolutional neural network technology to achieve the target detection process of power inspection photos. First, use labelimg to accurately label the power inspection training data set, and then use the fusion target detection network yolov4 and the detection-based multi-target tracking algorithm DeepSORT to address the problem of UAV positioning the target. Use the SiamRPN algorithm to achieve high-precision positioning to meet the needs of power inspection services and quickly identify targets in batches.