Thermal infrared images have many non-ideal characteristics, and the target regions obtained by many methods are incomplete. To facilitate subsequent applications, post-processing is often required to improve its integrity. In this section, a repair algorithm based on Schrödinger transform and LSAC model is described for the existence of broken human targets in the binarized image segmentation results. This method does not rely on prior knowledge of human body shape, and automatically determines the attribution of human body fragments according to the gray distribution of connected areas, eliminates interfering structures, and fills the defect areas between human body fragments at the same time.