A defect detection method of self-compared template matching based on ESPRIT is presented

which is mainly for the wafer that has same pattern. Firstly

the ESPRIT arithmetic is used to calculate precisely the period of two one-dimensional images and get the size of the building block. We can reconstruct the building block by shifting a window of proper size throughout the image and adding the corresponding pixel values. By averaging over all of the blocks in the image

the amount of noise and the effect of defects are reduced considerably. Standard models are formed by extending standard building block. We compare the original image with the standard models

and the point may be identified as a possible defect if the difference is larger than a threshold. The experiment verifies that this method has high-speed and high-precision

to meet the need of real-time detection. The complexity of this algorithm is O(N to the 3/2th power) and the precision is 0.04 pixel.