Thermal infrared human target detection technology is traditionally mainly used in the military field, but with the technological progress and price reduction of infrared imaging devices and equipment in recent years, its application has gradually expanded to transportation, safety protection, medical treatment, industrial production, fire rescue, etc. In the civil field, it has shown very broad development and application prospects, and has become an important supporting technology for the construction of an information-based and intelligent society in the era of artificial intelligence.

 

In terms of image characteristics, the inherent imaging mechanism of infrared imaging causes thermal infrared images to have the problems of poor imaging quality, less available information, and difficult processing. On the detection object, the complex appearance and behavior patterns of the human body make human features lack good aggregation. These two factors make it quite difficult to correctly detect, identify and analyze human objects from thermal infrared images. Therefore, based on the image processing and pattern recognition algorithm level, based on the basic characteristics of thermal infrared images and human targets, Shenzhen Sytong Technology conducts a more systematic analysis and discussion on the detection methods of human targets based on thermal infrared images. The content involves thermal infrared images. Image preprocessing, image segmentation, statistical recognition of human objects, and single human action recognition.