Detecting slow-moving underwater scale targets using multiple higher order lacunarity

This paper discusses the impact of high level clutter backgrounds caused by platform low-speed motion and shallow sea reverberation on target detection. Based on the traditional high-order time lacunarity (HOT-Lac), a multi-high-order lacunarity (MHO-Lac) algorithm is proposed, incorporating local and global computations. This method effectively alleviates background fluctuation issues caused by the platform’s low-speed movement and characterizes the dynamic properties of targets relative to clutter in continuous range and bearing sonar echographs. In addition, a blind deconvolution-based target bearing point spread function solution method is introduced based on the classical theory of sparse representation. By identifying the potential location of the low-speed moving scale target, it effectively mitigates the misclassification of foreground pixels caused by low-speed motion and suppresses clutter. Sea trials show that under the interference of clutter in shallow sea, the area under the curve (AUC) value of the method proposed in this paper has increased by more than 0.06 compared to the robust high-order flux tensor algorithm, demonstrating superior performance.