红外传统与智能融合目标检测算法研究OA

The continuous development of infrared decoy technology has increased the scene complexity,putting higher demands on the design of anti-interference algorithms.Traditional image processing algorithms rely on manually defined rules,lacking the ability to describe complex scenes.To compensate for this disadvantage,the intelligent image recognition algorithm is introduced,and the establishment,training,and testing of the artificial intelligence algorithm network based on improved YOLOv5 are completed.By introducing the concept of confidence coefficient,and fusing the algorithms at the pixel level,feature level and decision level,a traditional and intelligent fusion algorithm is proposed,which combines the clear scene division and strong interpretability of the traditional algorithm with the powerful feature extraction and learning capabilities of the intelligent algorithm.The simulation verification with 3 072 image sequences shows that the fusion algorithm improves the comprehensive target detection success probability by 12.99%in comparison with the traditional algorithm,verifying its effectiveness,highlighting its significance in enhancing the target detection capabilities of infrared air-to-air missiles.