Dim small targets detection based on statistical block low-rank background modeling

How to effectively detect weak targets from complex background is always a challenging problem and is a meaningful research subject with practical significance. In this paper, the complex video frame images are considered as a spatial random process, and the stationarity and low-rank characteristics of different components of the image are related to theirs statistical characteristics. According to this view, a statistical block low-rank background modeling algorithm (for short: SBLR) is proposed. This paper first analyzes the regional statistical characteristics of the image, and then uses k-mean statistical clustering algorithm to divide the image into statistical blocks to obtain the statistical block images. Then, the characteristics of each component of the statistical block image are analyzed to establish a model composed of statistical block low rank background and sparse components. Next, according to the characteristics of each component of the model, the solution scheme of principal component analysis is adopted, and the specific solution algorithm is given. Finally, the background reconstruction experiment according to SBLR algorithm and target detection experiment are carried out. Experiments show that the algorithm proposed in this paper achieves good accuracy in the background reconstruction of complex scenes, the background is significantly suppressed, the target is significantly enhanced, and the target detection rate is high.