Infrared noise suppression and detail enhancement method based on multi-scale characteristic

Abstract

The invention discloses an infrared noise suppression and detail enhancement method based on a multi-scale characteristic. Aiming at a condition that a processed infrared image has problems that a background noise is large, a contrast ratio is low and details are fuzzy, in order to reach purposes of noise suppression and detail enhancement, a Bayes optimum estimation model is adopted so as to provide a prior model suitable for depicting an infrared image characteristic with a low signal to noise ratio. Through further optimization of a multi-scale structure characteristic, the noise suppression and the detail enhancement are finally performed based on the multi-scale structure characteristic. In the invention, when the infrared image with the low signal to noise ratio is processed, a background noise can be eliminated, naturalness of an original image is maintained and an edge contour of a target is sharpened; and when the image with a high signal to noise ratio is processed, target details and textures can be adaptively increased and simultaneously an over-enhancement state of a sharp edge can be avoided.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to infrared noise suppression and detail enhancement based on multi-scale features. Background technique [0002] The infrared image is formed by detecting the thermal radiation difference between the surface of the target object and the background, and its imaging is not affected by the bright light of the external environment, and has a strong ability to detect and identify camouflage; in addition, the infrared thermal imager does not need to radiate energy outward, It only needs to passively accept external radiation, it is not easy to be detected and monitored by enemy detectors, and it has strong anti-jamming and camouflage capabilities. Therefore, infrared thermal imaging cameras are widely used in site monitoring [1] , medical imaging [2] and military remote sensing imaging [3] etc. However, affected by the application environment and acquisition equipment at the...

AI Technical Summary

Problems solved by technology

TV, however, in a noisy environment, the gradient value is not enough to be used as a feature to distinguish noise and region boundaries. After denoising, the image has flat areas that are not smooth enough, and there are still obvious noise and block effects.

The background noise suppression effect of the existing technology 2 on the infrared image is as follows: image 3 shown

Images