Research On Photoelectric Countermeasure Jamming Detection Technology Based On Infrared Characteristics

In the development of modern weapons and equipment and the construction of national defense,the position and role of photoelectric countermeasure in modern warfare are increasingly prominent.In view of the complex military environment,how to quickly detect and accurately detect enemy targets has always been a hot research issue,especially with the development of artificial intelligence technology,the research on target detection algorithms based on deep learning is also ongoing.Compared with visible light detection,infrared detection can meet the requirements of all-weather work,and has high concealment and strong penetrability.However,compared with visible light image detection,infrared small target detection faces more problems to be solved.Clutter interference exists in the complex background,and infrared targets have problems such as weak signal strength and small spatial scale.How to realize timely detection of incoming targets and fast and accurate target detection and recognition is a challenging problem.In this paper,the design of infrared detection optical system and detection methods are studied.The design of multi band infrared detection system and the detection method of infrared warning small target are studied:(1)In order to solve the problem of weak infrared target signal caused by clutter interference in complex background and long detection distance,an infrared multi band detection optical imaging system is proposed.By analyzing the infrared imaging characteristics of different military targets in each band,different optical imaging methods are designed for different infrared bands.In order to obtain more complete information about small infrared targets,infrared detection systems are deployed in different directions,So as to improve the detection accuracy and provide a basis for the subsequent infrared small target detection.(2)In order to improve the detection effect and reduce the false alarm rate of the infrared warning system,GAM-YOLOv5 and GAM-Mob-YOLOv5 detection algorithms are proposed based on the improved YOLOv5 infrared multi azimuth small target detection method.Firstly,according to the requirement of infrared small target detection,a multi view dataset is made and labeled,and the small target dataset is expanded by image enhancement algorithm.In order to fuse the infrared features among various azimuth views,GAM-YOLOv5 algorithm framework is constructed by adding GAM Attention attention mechanism to associate the information between multiple azimuth views to achieve the aggregation of multiple azimuth view features.In order to further optimize the deep learning architecture and improve the network detection efficiency,Mobile One lightweight backbone network is introduced on the basis of GAM-YOLOv5 to establish GAM-Mob-YOLOv5 detection algorithm.Finally,the improved GAM-YOLOv5 and GAM-Mob-YOLOv5 algorithms are compared with YOLO5 algorithm through experiments.The experimental results show that the detection accuracy and speed of multi azimuth small target detection based on GAM-Mob-YOLOv5 have been improved.Compared with the best detection result of single view of YOLOv5,the detection accuracy of GAM-Mob-YOLOv5 in multi azimuth view and enhanced multi azimuth view has been improved by 10.7%and 11.9%,and the detection speed has been improved by 98 fps,which verifies the effectiveness of the algorithm.