Infrared Night Camouflage Target Detection Algorithm Based on Improved YOLO v3

Based on the problem of camouflage multi-target detection in no-light scenes at night, this paper studies an infrared nighttime camouflage target detection algorithm based on improved YOLO v3. By improving its feature extraction network, the size of the output feature map of the original feature extraction network is adjusted. The feature information of infrared camouflage targets is strengthened, and the structure of the feature fusion network is improved to improve the detection performance and accuracy of the algorithm for nighttime camouflage targets. The comparative experiments on the infrared dataset based on night camouflaged people show that compared with the YOLO v3 algorithm, the recall rate of this algorithm is increased from 91.9% to 95.9%, and the mAP value is increased from 90.2 % to 95.6%. This proves the effectiveness of the proposed algorithm for the multi-target detection problem of no-light camouflage at night.