No-reference defogging image quality evaluation method based on HSI color space

Abstract

The invention discloses a no-reference defogging image quality evaluation method based on an HSI color space, and belongs to the technical field of image processing. For two types of distortion problems, namely color distortion and insufficient definition, which are easy to occur in a defogged image, the method converts the image into an HSI color space for processing, and sub-images corresponding to three channels are respectively obtained; based on different distribution laws, the sub-images of an H channel and an S channel are utilized to complete color evaluation, and the sub-image of an I channel is utilized to complete the definition evaluation, so as to achieve the high consistency of an objective evaluation result and the subjective feeling; the problem that quality evaluation of defogged images is difficult to complete due to lack of reference images is solved. The method has few processing steps, does not need to use fogless reference images and a large amount of data sets for model training, and is high in final evaluation result accuracy and simple in application.

Description

technical field [0001] The invention belongs to the technical field of image processing, and more specifically relates to a reference-free dehazing image quality evaluation method based on HSI color space. Background technique [0002] With the continuous development of industry and the improvement of urbanization level, the degree of deterioration of air quality is also deepening, corresponding to the frequent occurrence of smog weather. Haze is a common phenomenon in nature. It is due to the existence of suspended particles and small water droplets in the atmosphere. These particles affect the scattering and absorption of sunlight by the atmosphere, resulting in reduced visibility and affecting human visual perception. Due to the existence of smog, image imaging is affected, which greatly interferes with the image captured by the equipment, and causes the quality of the captured image to be damaged, which is not conducive to the acquisition of image information. [0003] ...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to solve the problem that the quality of defogged images is difficult to achieve blind evaluation and improve the reliability of the objective evaluation results of such images, and provides a method for evaluating the quality of defogged images without reference based on HSI color space; the present invention mainly aims at Dehazing images are often prone to problems such as color distortion and insufficient clarity. By converting it to the HSI color space, observing the sub-images corresponding to the three channels, and extracting features from the sub-images of the H and S channels based on natural statistical laws, Obtain the color evaluation result; at the same time, use the sub-image of the I channel to perform cumulative blur probability detection to obtain the sharpness evaluation result, and then combine the evaluation results of the two to complete the final quality evaluation, thus solving the problem of image clarity without reference comparison. degree assessment questions

Images