Research On High-Reliability Detection Of The Space Infrared Dim Small Target

With the rapid development of opto-mechanical structure and large focal plane infrared detection technology, much more attention are paid to space-based infrared target detection system because it has a lot of advantage such as long distance detection, being capable of working around the clock, great anti-interference ability and being capable of monitoring large field of view, etc. This kind of system is widely used in space Infrared target reconnaissance and infrared astronomical observation. In the field of infrared target detection and tracking, highly reliable detection technology for dim infrared targets detection under complex background is the core technology that determines the detection ability of the systems. The highly reliable detection and tracking methods for dim infrared targets in space are deeply studied in this paper combining with the real infrared detection systems.To analyze the impact of the various parts of the infrared imaging system on target detection, from the time and space domains, the model of the optical system, infrared detectors and information acquisition circuit are built. By analyzing the detector’s electronic noise, spatial sampling effect, the non-uniformity response, and this paper calculated their influence on target SCR, energy concentration and other parameters, and finally get the infrared imaging system’s quantized effect on dim weak targets.To ensure the highly reliable operation of space target detection system, a top-to-bottom program control flow platform plan is designed in this paper with taking use of TMR(Triple Module Redundancy), EDAC(Error Detect and Correct), FPGA dynamic partial reconfiguration and current measurement and control technology. Fault injection test results show that this program can achieve a protection of target detection system against both SEU(Single Event Upset) and SEL(Single Event Lock).To analyze the relationship among the target SNR, the detection probability and the false probability, this paper has theoretically proved that target SNR is the key parameter to both the detection probability and the false probability. To improve the target SCR, a non-uniformity correction algorithm based the equivalent integral capacitance of detector unit is come up with. Compared with the results of the two points correction algorithm, this algorithm can depress the infrared image non-uniformity down to 1%.In order to achieve rapid detection of mass infrared image data, a target fast search algorithm based on multi-frame target collection fusion is designed and 5 times faster detection is achieved in this paper. To improve low detection probability problems for the low SCR target in the single-frame target detection algorithm, this paper designs a target detection algorithm based on multi-feature probability fusion and achieve highly reliable target detection of single-frame and low SNR. Meanwhile, an improved TBD target detection algorithm based on DP-TBD is presented after studying DP-TBD algorithm intensively and the highly detection probability 0.86 for targets with low SCR is achieved in the condition of reducing both the complexity and computation amount of the algorithm.