Hierarchical system of lists for detecting and tracking moving weakly emitting point objects using image sequences

Subject of study. This study was focused on the detection and tracking of moving weakly emitting point objects. Aim of study. The aim was to develop an image sequence processing method for the stable detection and tracking of objects with low signal-to-noise ratios. Method. The proposed method for image sequence processing was implemented using a hierarchical list system consisting of a list of targets, an initial list, and a buffer list. All the detections that were found during single-frame processing and were not associated with the objects in these lists were entered into the buffer list, which occupied the lowest level of the hierarchy. If a detection in the current frame could be successfully associated with an object in the buffer list, it was transferred to the initial list, which included the candidates for track initiation. Upon meeting specific confirmation conditions, these candidates were transferred to the list of targets. The method employed various rules for the data association of frame detections with objects across these lists, rules for recalculating track and energy characteristics, and conditions for maintaining an object within a list. Main results. The proposed method enables the detection and tracking of objects with a low probability of detection on a single frame. Statistical experiments confirm the efficiency of this approach for target tracking and track maintenance. Practical significance. The proposed method enhances the detection range and tracking stability for poorly visible objects in real-time detection systems.