An Improved TLLCM Infrared Small Target Detection Method Based on Difference of Gaussians

Small target detection has great theoretical significance in the field of infrared (IR). Local contrast measure (LCM) has shown commendable performance in infrared small target detection. However, existing LCM methodologies primarily emphasize the background suppression rate, which leads to a decline in detection accuracy. In this paper, an enhanced algorithm based on the framework of Tri-layer LCM(TLLCM) is introduced, which incorporates the Difference of Gaussians (DoG) algorithm and morphological filtering operations. The proposed algorithm demonstrates an improvement in both Precision and Recall metrics, while preserving the low false alarm rate advantage inherent to TLLCM. On dataset IRSTD-1k, proposed method achieves BSF of 1934.61, SCRG of 27.74, Precision rate of 0.52 and Recall rate of 0.64.