Local threshold segmentation method and defect detection method

Abstract

The invention discloses a local threshold segmentation method and a defect detection method. The local threshold segmentation method comprises the following steps: selecting an initial central mark point at an initial position of an image; setting a local selection area A and a selection area B by taking the central mark point as a center, calculating a threshold value at the central mark point, and performing binarization processing on the point according to the threshold value, wherein the selection area B is a sub-area of the selection area A; in the direction of the image coordinate system, marking the next pixel point of the mark point or the pixel point with the distance t from the center mark point as a new center mark point, and repeatedly binarizing the new center mark point untilthe threshold segmentation of the whole image is completed; the method has obvious advantages in image detection of low contrast and fine defects, is good in real-time performance, and is suitable for monitoring product quality in an industrial field.

Description

technical field [0001] The invention relates to the field of visual inspection, in particular to a local threshold segmentation method and a defect detection method. Background technique [0002] Image segmentation is a key step from image processing to image analysis. The existing image segmentation methods are mainly divided into the following categories: threshold-based segmentation methods, region-based segmentation methods, edge-based segmentation methods, and segmentation methods based on specific theories, etc. , ideally for image segmentation, the exposure of the image is controlled within a reasonable range, the contrast is strong, the brightness is moderate, and the image segmentation effect is obvious. Influencing factors such as uneven distribution of gray scale and image noise, for example, for automotive stamping parts, due to the dark color of the stamping parts themselves and the complex detection environment of the automobile production workshop, the traditi...

AI Technical Summary

Problems solved by technology

[0002] Image segmentation is a key step from image processing to image analysis. The existing image segmentation methods are mainly divided into the following categories: threshold-based segmentation methods, region-based segmentation methods, edge-based segmentation methods, and segmentation methods based on specific theories, etc. Ideally, for image segmentation, the image exposure is controlled within a reasonable range, the contrast is strong, the brightness is moderate, and the image segmentation effect is obvious. However, due to the influence of light and the material of the measured object, the actual collected image is prone to overexposure and image blur Influencing factors such as uneven distribution of gray levels, image noise, etc. For example, for automotive stamping parts, due to the dark color of the stamping parts themselves and the complex detection environment of the automobile production workshop, the traditional threshold segmentation method is difficult to perform in low-contrast image segmentation. It is impossible to achieve better results on the surface, and the tiny defects on the surface (such as scratches, embossing defects, zinc dust defects) are easy to miss

Images