Small Target Detection in Noisy Image Sequences

JC Li,ZK Shen,WP Yang
DOI: https://doi.org/10.1109/naecon.1997.622742
1997-01-01
Abstract:This paper presents a new approach for detecting small, moving targets in noisy image sequences. This approach takes the following steps: 1) clutter suppressing, 2) adaptive thresholding, and 3) recursive maximum similarity filter. Among these processes, clutter suppressing algorithm and recursive maximum similarity (RMS) filter are a main contribution of this paper. Both of them can be implemented by fully parallel algorithm, and real-time is possible for image sequences at high frame rate. Test results show that the approach has ability to detect small low-contrast moving targets whose positions and trajectories are unknown.
What problem does this paper attempt to address?