Infrared Dim-small Object Detection Algorithm Based on Saliency Map Combined with Target Motion Feature

In response to the problem of high false alarm rates in infrared dim-small target detection with complex ground backgrounds, this paper proposes an infrared dim-small detection algorithm based on saliency map, combined with target motion features. This algorithm firstly transforms the small target detection problem of IR images into a problem of recovering low rank matrices and sparse matrices from original matrices, and then removes the false targets with an improved morphological filtering algorithm. By observing the target motion characteristics between multiple frames, a method combining with Kalman filter is proposed to correlate the target in order to further reduce the false alarm rate. The simulation results show that the proposed algorithm can effectively reduce the false alarm rate of infrared vehicle target detection in complex background, ensuring a high detection rate and real-time performance, while also solving the problem of short-term target obscuration and disappearance.