CN-121998882-A - Image enhancement method, image enhancement device and computer storage medium

CN121998882ACN 121998882 ACN121998882 ACN 121998882ACN-121998882-A

Abstract

The application provides an image enhancement method, an image enhancement device and a computer storage medium. The image enhancement method comprises the steps of carrying out multi-level segmentation on a target image to obtain a plurality of first images of a first segmentation level and a plurality of second images of a second segmentation level segmented by the first images, determining at least one partitioned image set, carrying out partition mean histogram equalization on the images in the partitioned image set, determining leaf images, carrying out pixel value histogram equalization on the leaf images, carrying out brightness correction on the result of the partition mean histogram equalization, carrying out brightness correction on the result of the pixel value histogram equalization by utilizing the result of the partition mean histogram equalization after the brightness correction, and forming an enhanced image according to all the leaf images after the brightness correction. By the image enhancement method, the continuity and the balance of the brightness distribution of the image are improved.

Inventors

YU KEQIANG
LI XINGLIANG
YUAN WENJUN
HAO DEJUN

Assignees

浙江大华技术股份有限公司

Dates

Publication Date: 20260508
Application Date: 20251211

Claims (10)

1. An image enhancement method, the image enhancement method comprising: Performing multistage segmentation on a target image to obtain a plurality of first images of a first segmentation level and a plurality of second images of a second segmentation level segmented by the first images; Determining at least one partitioned image set, wherein the partitioned image set comprises a plurality of first images or second images obtained by dividing the same image; carrying out partition mean histogram equalization on the images in the partition image set; determining a leaf image, wherein the leaf image is a first image or a second image which is not segmented; performing pixel value histogram equalization on the leaf image; performing brightness correction on the result of the partition mean histogram equalization; Carrying out brightness correction on the pixel value histogram equalization result by utilizing the brightness corrected partition mean value histogram equalization result; and constructing an enhanced image according to all the leaf images after brightness correction.
2. The method of image enhancement according to claim 1, wherein, The multi-stage segmentation of the target image comprises: performing image recognition on the target image to divide a foreground area and a background area; Determining a foreground region of the target image as a first image, or segmenting a plurality of first images; and determining the background area of the target image as a first image or segmenting a plurality of first images.
3. The method of image enhancement according to claim 2, wherein, The image enhancement method further comprises the following steps: when a plurality of foreground objects exist in the first image, continuously dividing a plurality of second images from the first image according to the image area of each foreground object; and continuously dividing the first image into a plurality of second images according to the preset threshold when the number of the image pixels of the first image is larger than the preset threshold.
4. The method of image enhancement according to claim 1, wherein, The performing the partition mean histogram equalization on the images in the partition image set includes: Determining a mean value of pixel values of each image in the partitioned image set; And carrying out histogram equalization on the pixel value average value of all the images in the partitioned image set.
5. The method for image enhancement according to claim 1 or 4, wherein, The brightness correction of the result of the partition mean histogram equalization comprises the following steps: determining a first father image of each second image and a first adjacent image of the same segmentation level of the first father image, wherein the first father image is the first image for segmenting the second image; determining a first correction parameter according to the first father image and the left first adjacent image; determining a second correction parameter according to the first father image and the right first adjacent image; And carrying out brightness correction on the histogram equalization mean value of the second image by utilizing the first correction parameter and the second correction parameter.
6. The image enhancement method according to claim 5, wherein, The brightness correction is carried out on the pixel value histogram equalization result by utilizing the brightness corrected partition mean value histogram equalization result, and the brightness correction comprises the following steps: Determining a second adjacent image for each leaf image; Determining a third correction parameter and a fourth correction parameter according to the average value equalization result of the leaf image, the average value equalization result of the left second adjacent image and the average value equalization result of the right second adjacent image; And carrying out brightness correction on the histogram equalization pixel values of the leaf image by using the third correction parameter and the fourth correction parameter.
7. The method of image enhancement according to claim 6, wherein, The determining a second adjacent image for each leaf image includes: Determining the left adjacent image as the left second adjacent image and the right adjacent image as the right second adjacent image in response to the leaf image having the left adjacent image and the right adjacent image segmented by the same image at the segmentation level; In response to the leaf image having a right adjacent image segmented from the same image at the segmentation level and not having a left adjacent image segmented from the same image, determining a left adjacent image of a second parent image of the leaf image as the left second adjacent image and determining the right adjacent image as the right second adjacent image; And in response to the leaf image having a left adjacent image segmented by the same image at the segmentation level and not having a right adjacent image segmented by the same image, determining a right adjacent image of a second parent image of the leaf image as the right second adjacent image and determining the left adjacent image as the left second adjacent image.
8. The method of image enhancement according to claim 7, wherein, And determining a third correction parameter and a fourth correction parameter according to the average value equalization result of the leaf image, the average value equalization result of the left second adjacent image and the average value equalization result of the right second adjacent image, wherein the method comprises the following steps: Determining a third correction parameter according to a mean value equalization result of a second father image of the leaf image and a mean value equalization result of the left second adjacent image when the right adjacent image obtained by dividing the same image exists in the division level of the leaf image and the left adjacent image obtained by dividing the same image does not exist; And determining a fourth correction parameter according to the average value equalization result of the leaf image and the average value equalization result of the right second adjacent image.
9. An image enhancement device, comprising a memory and a processor coupled to the memory; Wherein the memory is for storing program data and the processor is for executing the program data to implement the image enhancement method according to any one of claims 1 to 8.
10. A computer storage medium for storing program data which, when executed by a computer, is adapted to carry out the image enhancement method according to any one of claims 1 to 8.

Description

Image enhancement method, image enhancement device and computer storage medium Technical Field The present application relates to the field of image processing technologies, and in particular, to an image enhancement method, an image enhancement apparatus, and a computer storage medium. Background The image sensor acquires the image with various problems such as unobtrusive details, image get confused and the like, which is unfavorable for subsequent scientific research and also affects the visual experience of people. Therefore, image enhancement is an important branch of image processing and is continuously paid attention to by a plurality of scientists at home and abroad. When the field Jing Guangzhao is poor and the brightness distribution is extremely uneven, a large area of the pixel value of the photographed image appears. Rendering existing histogram equalization-based image enhancement ineffective. Many times, existing histogram equalization techniques encounter the above-described scene, and the enhanced image is instead worse than before. ‌ A Disclosure of Invention In order to solve the technical problems, the application provides an image enhancement method, an image enhancement device and a computer storage medium. In order to solve the technical problems, the present application provides an image enhancement method, which includes: Performing multistage segmentation on a target image to obtain a plurality of first images of a first segmentation level and a plurality of second images of a second segmentation level segmented by the first images; Determining at least one partitioned image set, wherein the partitioned image set comprises a plurality of first images or second images obtained by dividing the same image; carrying out partition mean histogram equalization on the images in the partition image set; determining a leaf image, wherein the leaf image is a first image or a second image which is not segmented; performing pixel value histogram equalization on the leaf image; performing brightness correction on the result of the partition mean histogram equalization; Carrying out brightness correction on the pixel value histogram equalization result by utilizing the brightness corrected partition mean value histogram equalization result; and constructing an enhanced image according to all the leaf images after brightness correction. The multi-stage segmentation of the target image comprises the following steps: performing image recognition on the target image to divide a foreground area and a background area; Determining a foreground region of the target image as a first image, or segmenting a plurality of first images; and determining the background area of the target image as a first image or segmenting a plurality of first images. The image enhancement method further comprises the following steps: when a plurality of foreground objects exist in the first image, continuously dividing a plurality of second images from the first image according to the image area of each foreground object; and continuously dividing the first image into a plurality of second images according to the preset threshold when the number of the image pixels of the first image is larger than the preset threshold. The step of performing the partition mean histogram equalization on the images in the partition image set includes: Determining a mean value of pixel values of each image in the partitioned image set; And carrying out histogram equalization on the pixel value average value of all the images in the partitioned image set. Wherein the performing brightness correction on the result of the partition mean histogram equalization includes: determining a first father image of each second image and a first adjacent image of the same segmentation level of the first father image, wherein the first father image is the first image for segmenting the second image; determining a first correction parameter according to the first father image and the left first adjacent image; determining a second correction parameter according to the first father image and the right first adjacent image; And carrying out brightness correction on the histogram equalization mean value of the second image by utilizing the first correction parameter and the second correction parameter. The brightness correction of the pixel value histogram equalization result by using the brightness corrected partition mean value histogram equalization result comprises the following steps: Determining a second adjacent image for each leaf image; Determining a third correction parameter and a fourth correction parameter according to the average value equalization result of the leaf image, the average value equalization result of the left second adjacent image and the average value equalization result of the right second adjacent image; And carrying out brightness correction on the histogram equalization pixel values of the leaf image by using the third correction pa