| 摘要 |
PURPOSE:To surely recognize a fine flaw disposed on an object to be observed by excluding picture data at the periphery of a boundary from decision data on the picture of a master and the object to be observed. CONSTITUTION:The original picture PTO of the master is dislocated by + or -one picture element in X, Y directions to obtain dislocated pictures PTO1-PTO 4. The brightness information is added one by one picture element to respective picture parts, thereafter, divided with the 4 sheet number to obtain the average brightness distribution of an overlapped picture PT1. Then, the respective brightness information of the PT1 is subtracted from the picture PTO to obtain a difference picture PT2. Then, a threshold level TH is set based on the brightness information of the picture PT2, a boundary detecting and excluding picture PT3 formed by formed by allocating a level 1 to a picture part having the brightness of the picture PT2 higher than the value TH is formed, a zero level is allocated to a picture part having a lower brightness to exclude the detecting and excluding area PT31 of level and the picture to be observed is compared with the master picture only in the detecting and executing area PT32 of zero level. |
