Infrared small target detection technology is an important subject in computer vision. In infrared detection system, it is required to quickly read and accurately segment moving small targets from sequence images, and the clutter suppression effect of image preprocessing algorithm directly affects the detection probability of the system. However, due to the traditional detection using fixed threshold segmentation, the poor segmentation effect leads to a large amount of data for small targets to be detected, which often takes a long time to process in embedded platform, and does not meet the real-time requirements. In addition, the long-distance infrared small targets have low signal-to-noise ratio, few pixels and lack of geometric features, which often lead to false alarms in detection, thus reducing the reliability of the detection system. Based on the above analysis, this paper designs an infrared small target adaptive detection system based on DSP+FPGA by using adaptive threshold segmentation and small target feature information, and transplants the improved algorithm to the embedded platform of DSP+FPGA. Moreover, the actual images are collected for verification. The experimental results show that the system designed in this paper can suppress the background clutter well, implement the adaptive detection of small infrared targets, and meet the real-time and reliability of detection.
|