Small and Dim Target Detection Based on Learning Mechanism in Complex Background

The detection of small and dim targets in complex scenes is currently an important and difficult problem in the field of target detection. This paper proposes an improved background modeling method based on Gaussian mixture model (GMM), which implements adaptive adjustment of modeling parameters and judgment of foreground from multiple dimensions. In addition, we combine the results of discriminative correlation filter tracker to aid the detection. Experiments are conducted in real scenes using the new method in comparison with mainstream background modeling methods, and our method can effectively improve the detection performance of small and dim targets in terms of qualitative and quantitative metrics.