For thermal infrared sequence images with relatively static background, it is difficult to obtain ideal results by directly using background subtraction to segment human targets. However, compared with the optical flow method and the frame difference method, background subtraction is easier to obtain the complete motion area, and the calculation cost is relatively small. Therefore, this section puts it under the framework of the LSAC model, and expounds the real-time background subtraction based on the real-time background subtraction based on LSAC (RT-BSLSAC) model [86] The model uses the evolution mechanism of the active contour curve to detect the foreground At the same time, the fast numerical algorithm is designed to ensure the standard and impose space constraints to suppress the noise on the operating efficiency of the pre-certification model.