Object Detection and Danger Warning of Transmission Channel Based on Improved YOLO Network
Wenbo Ning, Xiaochen Mu, Chong Zhang, Taotao Dai, Sheng Qian, Xiaotong Sun · 2020 IEEE 4th Information Technology, Networking, Electronic and Automation Control Conference (ITNEC) · 2020
In view of the wide monitoring area of the transmission channel, the proportion of object in the captured pictures is small, especially the problem of early warning of targets in inspection images under complex backgrounds such as mutual occlusion, forests, mountains and rivers and multiple interference factors. A YOLOv3-MobileNet algorithm for power transmission channel danger warning is proposed. According to the characteristics of the transmission channel image, the number of network layers is increased while maintaining a small amount of calculation, and the feature mapping module is enriched to provide more accurate semantic information for the prediction layer. An IoU-Kmeans algorithm is proposed, which improves the predicted position of the target. Experimental results show that the optimized algorithm's detection accuracy (mAP) is 90%. Compared with SSD and YOLOv3 object detection algorithms, the detection accuracy of the algorithm is improved by 75% and 3.5%.