Research On Recognition Of Infrared Point Target

Infrared early warning system is through the passive infrared radiation to accept target identification of the target. Compared with conventional radar warning system, it is mainly good for hiding, mobile, free from electromagnetic interference and can provide intuitive image information. It can make a timely warning in itself not to be found and destroyed under the premise of conventional and anti-implicit goal of the near-Earth class.In the infrared early warning system, the halfway point of the complex background target recognition is the key technology. Attacks target and system distance is generally far, aimed at present in the infrared image for point like, accounted for only a few or ten few pixels, without apparent form and texture features such as information can use, and its strength is generally weak, easy to be complicated background submerged. This article first track detection method based on dynamic programming method for target identification, and multi-stage hypothesis testing into one to improve.Throughout the identification process, the principle of infrared imaging and infrared imaging model is constructed, then the original infrared image preprocessing background suppression, and finally the use of dynamic programming detection method to identify the target. Improved method in the original method of hypothesis testing two segmentation process, the first split is to remove the image below the image mean split second threshold grayscale cumulative value the introduction of the dynamic programming method, have to consider the accumulation process have to consider the current frame in the process of accumulation. Experiment using two methods, shows that the improved algorithm the average computation time of4.56seconds,5.24seconds less than the original method, and the requirements meet the detection probability of the signal to noise ratio SNR≥1.8, lower than the original method SNR≥3.0.