Dim Moving Target Detection Based on Imaging Uncertainty Analysis and Hybrid Entropy

Abstract

The detection of weak moving targets is the fundamental work for target recognition and plays an important role in precise guidance systems, missile warning and defense systems, space target surveillance and detection tasks, satellite remote sensing systems, and other areas. In low signal-to-noise ratio conditions, the target signal is often submerged by strong background and noise signals, making it difficult to detect. Existing detection methods mainly focus on infrared bright target detection, which cannot solve the problem of darker targets compared to the background. Additionally, these methods do not sufficiently consider the temporal signal characteristics of moving targets in video sequences. To address these difficulties, this paper introduces an uncertainty analysis approach, establishes a target confidence model for weak moving target detection based on the principle of uncertainty in the time domain, and proposes a mixed entropy enhancement detection method based on single-pixel temporal signals. This method effectively suppresses background noise and enhances target signals in gaze scenes, enabling the detection of weak moving dark and bright targets. The proposed method demonstrates good performance in real-life scene experiments.