A Dynamic Programming Track-Before-Detect Method Based on Multiple Hypothesis Testing

Abstract

The invention discloses a multi-hypothesis test-based dynamic programming detection-before-tracking method, which belongs to the field of target detection, and in particular relates to the technical field of weak target detection and tracking. Firstly, the value function is calculated by using all the echo data in the sliding window, and the value function is detected, and the detection result and target track estimation are output; then the information of the current first frame data is removed from the value function in turn, and the processed value function Detection, output detection results and target track estimation. This method considers the detection of different sub-sliding windows, but does not need to reprocess the data in the sub-sliding windows to calculate the value function. Therefore, while keeping the amount of calculation almost unchanged, it effectively solves the detection delay problem of the traditional dynamic programming tracking before detection method, improves the detection probability of targets whose appearance time is later than the sliding window start time, and greatly improves the performance of the algorithm.

Description

technical field [0001] The invention belongs to the field of target detection, in particular to the technical field of weak target detection and tracking. Background technique [0002] Dynamic programming tracking before detection algorithm is an effective method to detect and track weak targets. In recent years, it has received extensive attention and research from scholars at home and abroad, and has been applied to many fields such as radar tracking and image processing. This method is a batch processing method, which jointly processes the latest several frames of echo data at each moment, which is called a sliding window. [0003] In practical applications, due to the limited visual range of the sensor, target splitting, and mutual occlusion of targets, new targets may appear at any time within the detection range of the sensor. For example, in radar detection, the target may enter the radar's monitoring area after the radar starts scanning for a period of time; or in ...

AI Technical Summary

Problems solved by technology

In these cases, the first few frames of data in the sliding window do not contain the information of the newborn target, but the traditional dynamic programming tracking before detection algorithm still jointly processes all the echo data in the sliding window, resulting in the value function of the target state being Noise and clutter interference cannot be detected in this sliding window, and the detection probability of the algorithm is damaged

Although as the sliding window moves, the data that does not contain new target information will be gradually moved out of the sliding window, and the value function can be accumulated correctly, but this will cause the detection delay of the target and bring loss to the system performance

Images