Infrared Maritime Target Detection Based on Temporal Weight and Total Variation Regularization Under Strong Wave Interferences

Infrared maritime remote target detection (IMRTD) faces great challenges when suffering strong wave interference, because the saliency of the strong wave interference may exceed that of the target in successive frames. This leads to a significant decline in the performance of existing methods accompanied by a considerable number of false alarms and even missing the real target. In this manuscript, a novel spatio-temporal tensor model named temporal weight and total variation regularization (TWTVR) is proposed. First, the temporal coherent weight (TCW) of the target component is designed by the temporal coherency of the target and the temporal variability of the wave interference to strengthen the target. Second, to suppress strong wave interferences, we design a spatio-temporal multicomponent total variation (STMCTV) regularization. Finally, the TWTVR model is designed to be solved by the alternating direction multiplier method (ADMM). An efficient adaptive penalty growth rate is applied to reduce the number of iterations under the premise of ensuring detection accuracy. Qualitative and quantitative analysis of sequences from 12 different scenes demonstrates the superiority of TWTVR, which verifies the superior performance of TWTVR on IMRTD with strong wave interference.