An Intelligent Image Fusion Method Driven by Target Features

Abstract

The invention relates to an intelligent image fusion method driven by target features. Firstly, the salient target area in the infrared image is obtained based on the saliency detection and segmentation; then the image feature set and the fusion algorithm set are constructed, and the principle and method of fuzzy mathematics are used Analyze the fuzzy mapping relationship between the two; finally, under the NSCT decomposition framework, use the characteristics of the target area to drive the fusion of the target area, and select the low-frequency visible light that retains more visible light information for the background area, and take a large high-frequency absolute value strategy. The inverse NSCT transform achieves the final image fusion. This method can adaptively select the optimal fusion method of the target area, maintain the target characteristics to a great extent, retain more background details of the image, and solve many problems such as information loss, incomplete targets, and background blur. Defects, the fusion image presents the characteristics of high contrast, high detail information, and target highlighting, and has a good visual effect.

Description

technical field [0001] The invention belongs to infrared and visible light image fusion methods, and relates to an intelligent image fusion method driven by target features, which applies the fuzzy control decision theory in fuzzy mathematics to the field of heterogeneous image fusion, and the invention results can be applied to various military or civilian applications Heterogeneous image processing system. Background technique [0002] Image fusion is the process of combining several images from multiple sources, multiple phases, and multiple resolutions in the same scene into one image. The fused image contains all the information that people are interested in in the input image. Through the fusion of multiple images, the limitations of a single sensor image in terms of geometry, spectrum, or spatial resolution can be overcome to make the final image information more accurate and comprehensive. Therefore, image fusion technology has important practical significance in th...

AI Technical Summary

Problems solved by technology

However, there are still defects in this method: the fusion strategy based on artificial prior selection is often effective for a single scene, and cannot adapt to the dynamic changes of image features.

[0006] The traditional image fusion method cannot be fused adaptively according to the difference of target features, resulting in the disadvantages of unclear target and blurred background in the fused image

Images