Weak target detection method by use of time-space domain filtering adaptive to pipe diameter

Abstract

The invention discloses a weak target detection method by use of time-space domain filtering adaptive to a pipe diameter. The weak target detection method comprises steps of: carrying out background detection on a to-be-processed image by use of anisotropic differential algorithm so as to improving subsequent target detection ability; segmenting a difference graph by use of a local maximum value so as to acquire a binary image; initializing a time domain parameter (an accumulation frame length) and a space domain parameter (the size of the pipe diameter), and successively inputting a series of binary images with the accumulation frame length of N; and finally, by use of the time-space domain filtering adaptive to the pipe diameter, detecting multiple frames of images so as to acquire a real target point, and superposing the detection results so as to output a target movement track. Compared with a traditional pipeline filtering target detection method with a fixed pipe diameter, the weak target detection method is advantageous in that based on the correlation between a target and the time-space domain multi-frame movement, the size of the pipe diameter is adaptively modified according to the change of a target scale, so a detection problem caused by the fact that the target becomes small/big due to unchanged pipe diameter is effectively solved and target detection precision is greatly improved.

Description

technical field [0001] The invention relates to a weak and small target detection method, which is characterized in that the weak and small target detection method adopts pipe diameter self-adaptive time-space domain filtering, and is used for image processing, computer vision and target detection and positioning. The invention belongs to the technical field of target detection in a photoelectric capturing and tracking system. Background technique [0002] The detection of weak and small targets is one of the core technologies in the photoelectric capture and tracking system. It uses image processing algorithms to detect targets in low signal-to-noise ratio and strong noise environments. The performance of the algorithm is very important for the operating distance and early warning of the photoelectric capture and tracking system. The essential. Due to the small imaging area of ​​the target at a long distance, the energy of the target signal is relatively weak, especially u...

AI Technical Summary

Problems solved by technology

The pipeline filtering algorithm is a classic multi-frame detection algorithm, including the mobile weighted pipeline filtering algorithm and the multi-constraint pipeline filtering algorithm proposed later, which can suppress the interference of noise to a certain extent, but the diameter of these methods is fixed, and the real The size of the moving small target is constantly changing. The traditional pipeline filtering algorithm can only detect the target whose size is smaller than the pipe diameter, and cannot adapt to the change of the target size, which often leads to detection failure in practical applications.

Images