Target saliency detection method for improving anti-interference performance based on gray scale stability

Abstract

The invention provides a target saliency detection method for improving anti-interference performance based on gray scale stability, belongs to the technical field of visible light image target detection, and solves the problem of inaccurate target detection caused by complex background and relatively small target in a target detection process. In order to overcome near fine clutter interference such as ground gravel, ruts and plaques in a field scene in background interference of a target, the method comprises the following steps of: performing equal-interval threshold segmentation on a grayscale image, and setting different judgment thresholds to obtain a plurality of MSER regions; screening the number of pixels and the aspect ratio of each MSER region, and superposing the screened MSER regions to form an MSER saliency map; according to the method, the target and the background are subjected to saliency distinguishing by using the gray stability of the target area at different judgment thresholds, a relatively complete target area contour is obtained on the premise of suppressing background interference, and a relatively accurate saliency detection effect is achieved.

Description

technical field [0001] The invention belongs to the technical field of visible light image target detection, and relates to a target saliency detection method. Background technique [0002] In the terminal guidance technology, the target detection and identification technology is the key technology for the initialization of the non-human guided target at the end of the loop, which directly affects the effect of the terminal guidance. [0003] For general targets, many very comprehensive datasets (such as Pascal VOC, COCO) can be provided for neural network training. [0004] For the TV seeker to detect the position of military vehicles in the field, it is difficult to obtain image data sets of military vehicle targets in the field, especially the data sets of military vehicle targets seen by the seeker in the terminal guidance and control scene are even more difficult to obtain. The reason is: due to the sensitivity of military vehicles, it is impossible to directly collect...

AI Technical Summary

Problems solved by technology

[0009] The technical problem solved by the present invention is to solve the problem of inaccurate target detection caused by complex background, similar target size and interference size (or relatively small target) existing in the detection process of the target in the wild scene by the TV camera, and provides A target saliency detection method, which can eliminate the background interference of the target, overcomes the problem that the existing target saliency detection method is difficult to detect small targets (due to the long detection distance), and obtains a saliency image. saliency target area

[0010] The present invention proposes a target saliency detection method based on grayscale stability to improve anti-interference performance, which is a new target saliency detection method, with Solve the problem of saliency detection of multiple small targets in a complex background, and overcome the background interference of the target, including ground gravel, ruts, plaques and other near-fine clutter in field scenes

Images