Multi-target segmentation method for illumination non-uniform image

Abstract

The invention relates to the field of image processing, in particular to a multi-target segmentation method for an image with non-uniform illumination, which comprises the following specific steps of: (1) inputting the image and converting the image into a grayscale image; (2) after the step (1), setting an image pixel neighborhood, and starting to traverse image pixels; (3) after the step (2), calculating a gray average value and an image entropy value, and dividing image pixel neighborhoods into four local neighborhoods, namely brightness neighborhoods and characteristic strength neighborhoods; and (4) after the step (3), processing different pixel points by adopting different neighborhood sizes and threshold calculation methods until all the pixel points of the image are traversed. According to the method, the gray level distribution aggregation degree and spatial characteristics of pixel neighborhoods can be fully considered, even if a halo effect exists in an illumination non-uniform image, appropriate neighborhoods and threshold values can be automatically selected, whether the current pixel is in a target area or not can be better judged, and it is guaranteed that the outline of a target or a defect can be accurately divided.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a method for multi-target segmentation of unevenly illuminated images. Background technique [0002] In image processing, the collected images often cannot make the illumination of multiple targets uniform, and even in some cases where the illumination is smooth or the field of view is too large, the image containing the target can only be collected under the condition of uneven illumination. Uneven illumination leads to different processing speeds and accuracy for targets at different positions in the image. Therefore, it is necessary to segment images with uneven illumination to separate the contours of each target, which is convenient for subsequent processing such as positioning, detection, and measurement. [0003] The segmentation methods of unevenly illuminated images are mainly based on histogram equalization combined with local threshold segmentation and homomorphic fil...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome at least one of the above-mentioned defects in the prior art, the present invention provides a multi-object segmentation method for unevenly illuminated images, which divides image pixel neighborhoods into bright and dark and feature intensity by using the gray mean value and image entropy Four kinds of local neighborhoods, for different pixels, adopt different neighborhood sizes and threshold calculation methods, which can fully consider the gray distribution and spatial characteristics of pixel neighborhoods. In unevenly illuminated images, even if there is light The halo effect can also automatically select the appropriate neighborhood and threshold to better judge whether the current pixel is in the target area, and ensure that the outline of the target or defect can be accurately divided

Images