TFT LCD mura defect detection method based on hybrid self-adaptive level set model and multi-channel fusion

Abstract

The invention discloses a TFT LCD mura defect detection method based on a hybrid self-adaptive level set model and multi-channel fusion, and belongs to the technical field of LCD mura defect machine vision detection technologies. The invention discloses a mura detection method based on a hybrid self-adaptive model with global information and local information fusion. The hybrid self-adaptive model can increase the curve evolution speed and effectively overcome the interference caused by non-uniformity of the background gray scale; furthermore, the hybrid self-adaptive model can be self-adaptively decreased when getting close to a target region so as to prevent over convergence and realize accurate partitioning of a weak edge. Meanwhile, the invention discloses a detection scheme based on multi-channel fusion of gray-scale maps and s channel images so as to be compatible with detection for different types of mura. According to the TFT LCD mura defect detection method, an ROI region can be accurately extracted, and a texture background of the ROI region is suppressed; the self-adaptive model is used for overcoming the interference caused by the non-uniformity of the background gray scale and solving the difficulty of extremely low contrast ratio of the weak edge, so that accurate partitioning for the edge of a mura defect can be realized.

Description

technical field [0001] The invention relates to the technical field of machine vision detection of TFT LCD mura defects, and more specifically relates to a method for detecting TFT LCD mura defects based on a hybrid adaptive level set model combined with multi-channels. Background technique [0002] With the development of liquid crystal displays (Liquid crystal display, LCD) in the direction of large screen, thinner, and higher resolution, the defect detection work of the screen is becoming more and more important. The mura defect is used to describe the brightness imbalance perceived by the human eye when observing the display. It is a low-contrast object with no fixed shape and blurred edges. The main reasons are defects in the circuit or structure and uneven material properties. It is the research focus and difficulty of LCD defect detection. At present, most of the methods at home and abroad use manual inspection to complete the back-end inspection of the TFT LCD produ...

AI Technical Summary

Problems solved by technology

This application solves the problem that the mura defect cannot be accurately segmented due to the uneven brightness of the image background, but the application also cannot take into account different types of mura detection, and cannot solve the problem that the weak edge position of the mura is difficult to accurately locate

Images