F. Chen, J. Li, and Z. Jing (PRC), and H. Leung (Canada)
target detection, nearest neighbor, data association,infrared, image sequence
In most actual applications, the real infrared images have very low SNR. Small targets cannot be detected reliably using one image frame. In this paper, a local entropy method is employed to preprocess the real image data. With image sequences, a small, moving target detection and tracking algorithm is proposed by employing nearest neighbor association in the image plane, which integrates the target energy along its track and decreases the false alarms considerably. Simulation results with the real image sequences show the proposed algorithm can successfully detect and track the small moving target.
Important Links:
Go Back
