Answers are grounded in this paper and cite their sources.
Current detection methods often struggle with serious false alarm residue and missed detections when extracting infrared small targets in intricate backgrounds. To solve this problem, a detection method based on a multiscale trilayer local statistical difference measure (TLSD) is proposed. First, the local statistical characteristics (including average, median, and standard deviation) between small targets and background clutter are analyzed from the perspective of a three-layer filter template. On this basis, a local difference measure (LDM) and a target likelihood measure (TLM) are constructed to reasonably represent the target signal. Subsequently, the TLSD is defined by integrating LDM and TLM, which can greatly reduce false alarms and enhance the target saliency. Finally, the multiscale calculation is used to extract infrared targets with different sizes in the image. Results prove that the algorithm can efficiently detect weak targets in complicated background, and it can achieve more robust and effective performance compared with existing algorithms according to visual and quantitative comparisons.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: