Infrared Small Target Detection and Tracking
|School||Dalian University of Technology|
|Course||Computer Software and Theory|
|Keywords||Small Target Pre-Processing Morohology Wavelet Transform KalmanFilter|
With the development of infrared technology, infrared small target detection and tracking has been widely used in the medical field and military field, such as medical imaging, precision-guided, automatic control. When the target is relatively far from the infrared detector, the occupied area of targets is small in the imaging plane, and the signal-to-noise ratio of the target is low, which increases the difficulty of the small target detection and tracking.This paper first introduces the current research status of infrared small target detection and tracking, and proposes the difficulties of detection and tracking, then analyzes the infrared radiation characteristics, the background clutter and noise characteristics, the geometric characteristics and of small target and the motion characteristics of small target.Infrared image pre-processing, infrared small target detection and infrared small target tracking, which were deeply researched in this paper. The pre-processing of infrared image can enhance the signal-to-noise ratio of the small target, and suppress the background clutter. This paper briefly introduces several common pre-processing method of infrared image, such as histogram equalization, high-pass filtering, median filtering and difference method etc. Presently, the method of infrared small target detection is mainly including DBT(detection-before-tracks) and TBD(tracks-before-detection), In this paper, infrared small target detection method belongs to the category of DBT. In the spatial domain, morphological top-hat operator which is used for infrared image processing and small target detection. In the frequency domain, wavelet transform is used for small target detection. A fusion method which combines spatial with frequency domain, increasing the rate of detection. Based on small target detection, this article uses kalman filter to track small targets, which can predict the position of the small target at the next moment and reduce the amount of calculation, thereby it ensures the real-time tracking.