Infrared Target Detection And Tracking

This paper addresses the issue of detecting and tracking small moving targets in infrared (IR) image sequences, which is a hard work due to the high variability of target types and low signal-to-noise ratio. The issue has applications in air defense, air traffic control and long rang surveillance. It is always desirable to achieve very high sensitivity in the detection of moving objects with the lowest possible false alarm rates. To achieve the goal, the paper presents some algorithms for detecting and tracking moving targets in IR image sequences.The detection algorithms are classified into two categories according to target size. For segmenting larger targets, a modified PCNN algorithm is proposed. Otsu algorithm and genetic algorithm are compared and analyzed with the proposed one. Experimental results show that the proposed one can get more complete region and edge information with much lower complexity and higher speed.For small targets, the paper compares the performance of five different high-pass filters for background suppressing. Then two feasible segmenting algorithms are introduced according to the different characteristics of filtered images. For further improving the performance of segmenting, a morphological filter is employed to extract targets of interest. Experimental results demonstrate that the algorithm is effective, time-saving and easy for hardware implementation.Combining kalman filter with fuzzy algorithm, the paper also develops a method for tracking moving targets at a long rang. Kalman filter is exploited for predicting locations of targets and fuzzy algorithm is developed to simulate human memory for determining the probability of targets. Experimental results show that the algorithm is effective even though the targets are in heavy clutter and may become occluded.