An Image Dehazing Method Based on Derivative Graph Fusion Strategy

Abstract

The invention discloses an image defogging method based on a derivative graph fusion strategy, which includes the following steps: Step A: Construct a sample set based on a large number of original foggy images and their corresponding fog-free images; Extract the derivative image of the original foggy image from two angles to enhance the detail recovery of the image's distant and near scenes by the dehazing method, eliminate color cast and enhance contrast; step C: build a U-shaped convolutional neural network; step D: cascade step A to obtain Derivative map of the original foggy image and the original foggy image are used as input, and the haze-free image is used as the output to train the network built by step C; Step E: use the network trained by step D, and concatenate the original foggy image and the derivative map corresponding to the foggy image as input Predict haze-free images after dehazing. The invention has good defogging effect.

Description

technical field [0001] The invention belongs to the field of image information processing, and in particular relates to an image defogging method based on a derivative graph fusion strategy. Background technique [0002] Image dehazing aims to recover the details of a scene and estimate an unknown haze-free image from a given hazy image. Dehazing algorithms are required in many situations, such as everyday photography, automatic surveillance systems, satellite remote sensing, outdoor object recognition, and visual navigation in low-visibility environments, etc. However, the image quality captured under severe weather conditions such as fog and haze is easily disturbed by the fog, which reduces the contrast of the captured image and seriously degrades the color. Such foggy images often lack visual vividness and clarity. Therefore, image dehazing techniques are urgently needed not only in everyday photography but also in many computer vision applications. [0003] At presen...

AI Technical Summary

Problems solved by technology

[0007] The problem to be solved by the present invention is to provide an image defogging method based on a derivative map fusion strategy in view of the existing defogging method relying on the transmission map and the atmospheric value prediction mechanism, which has problems such as incomplete defogging and local color shift. It can effectively extract the derivative image of the foggy image from multiple angles through the fusion of the defogging network, thereby realizing image defogging, and has a good defogging effect on images of various fog concentrations in various scenes

Images