Research on target detection method based on CNN

Abstract

Abstract Traditional target detection methods have poor tracking effect for targets with excessive displacement or occlusion between adjacent frames, and it is difficult to deal with high-speed moving targets. In view of the low detection rate of the current system in detecting long-distance weak and small targets, false targets and occlusion in the air, the working mode of convolutional neural network (CNN) and its advantages in target detection are analyzed, and a target detection method based on CNN is proposed, which can ensure the target detection accuracy and meet the real-time requirements of combat scenes, Finally, compared with other algorithms, the proposed method achieves the best results on four types of data. While the center error is low, it has high tracking success rate and low time complexity, which meets the requirements of high precision and real-time of the system.