An effective image segmentation method for noisy low-contrast unbalanced background in Mura defects using balanced discrete-cosine-transfer (BDCT)

This article presents a new image segmentation method to extract the detecting component from its background image. 2D automatic optical inspection (AOI) technology for defect detection and classification has played a vital role for on-line manufacturing industrial sectors nowadays. Image segmentation is a crucial step to extract the component information from its neighboring background. Due to potential complexity in such an image processing operation, considerable challenges are still encountered from establishing a robust approach. In general, three major factors play a significant influence on the result of the segmented image objects: (1) brightness distribution of the background image; (2) degree of unbalanced brightness of the background image; and (3) noise level near the object feature to be detected. This research addresses these important factors and develops a balanced discrete-cosine-transfer (BDCT) method. Exclusive advantage of the BDCT method is to overcome the current limitations of discrete cosine transform (DCT) methods in reconstructing the background image in horizontal or vertical directions only. The segmentation performance of the developed method is up to 25% better than the DCT method, in terms of accuracy of segmentation. From the test results on some real industrial image cases, it is verified that the method is capable of detecting the expected component accurately.