Infrared image significance detection method

Abstract

The invention discloses an infrared image significance detection method. The infrared image significance detection method comprises steps that multidimensional local sparse representation of an infrared image is executed to acquire an image salient map based on the multidimensional local sparse representation; the improved local contrast measurement processing of the infrared image is executed to acquire an image salient map based on the improved local contrast measurement; the above mentioned image salient maps are integrated together according to certain weight, and then the final salient map of the original infrared image is acquired. The infrared image significance detection method is advantageous in that the salient area of the infrared image is protruded completely, and at the same time, the background of the infrared image is fully inhibited, and by comparing with other methods, the detection rate of the signification is higher, and a false alarm rate is lower.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to an infrared image saliency detection method based on multi-scale local sparse representation and improved local contrast measurement. Background technique [0002] The saliency detection of infrared images aims to allow the computer to quickly identify the salient regions in infrared images that contain useful information and are most concerned by human vision. The result of the saliency detection is that the salient area is completely highlighted, and the background is well suppressed, so as to facilitate the subsequent further processing of the image. [0003] In current image processing research, a series of methods of saliency detection are successfully applied to visible light images. For example, the publication number CN102609911A is an edge-based image saliency detection method, which determines the salient area in the image through the detection of salient poin...

AI Technical Summary

Problems solved by technology

This method can highlight the target evenly, but it relies too much on the accuracy of the convex hull. While highlighting the target, a large part of the error area in the convex hull is also displayed, thus reducing the accuracy of saliency detection.

[0005] In short, although there are many target saliency detection methods, they all have some shortcomings. The specific performance is that the algorithm is more sensitive to noise due to the interference of the background in the image, and it is difficult to take into account the local features of the target and the scale features of the target. Low detection accuracy and high false alarm rate

In addition, for infrared images, if the existing image saliency detection algorithm for visible light is directly used to detect salient regions, since the special imaging mechanism of infrared images and the special properties of salient regions in the image are not considered, the performance of the algorithm will be reduced. Greatly reduced

Images