基于YOLO智能网络的红外弱小多目标检测技术

复杂背景下的红外弱小多目标检测是红外目标检测的难点,现有算法的稳健性难以满足实际弹载应用需求。针对弹载环境下的红外场景图像开展了基于典型深度学习网络模型的目标检测应用研究,提出一种基于YOLO(you only look once)网络的智能目标检测方法,通过高维特征学习表征和推理实现红外弱小多目标检测。采用传统的模板匹配算法和YOLO深度学习算法进行识别性能对比分析,验证了YOLO网络在红外弱小多目标检测方面的良好性能。实验结果表明:YOLO算法的检测概率可达92.2%,平均检测精度为0.844,与传统的模板匹配方法相比,YOLO具有明显的优势。

Infrared Dim-small Multi-target Detection Technology Based on YOLO Intelligent Network

Infrared dim-small multi-target detection under complex background is a difficulty, and the robustness of existing algorithms is difficult to meet the actual application requirements of missile load. In this paper, the target detection application research based on a typical deep learning network model is carried out for infrared scene images under the missile environment. An intelligent target detection method based on YOLO (you only look once)network is proposed, which is realized by learning and representing of high-dimensional features and reasoning. The traditional template matching algorithm and the YOLO deep learning algorithm are used for the comparative analysis of recognition performance, which verifies the good performance of the YOLO network in the infrared dim-small multi-target detection. The experimental results show that the detection probability of the YOLO algorithm can reach 92.2% and the average detection accuracy is 0.844. Compared with the traditional template matching method, YOLO has obvious advantages.