Board classification method and apparatus, a substrate program for executing the classification method in the computer, and a computer readable recording medium storing the program

摘要

PROBLEM TO BE SOLVED: To provide a method and a device of classifying substrates by which, according to knowledge information concerning a known defect distribution, a group of substrates having the corresponding defect distribution can be accurately classified, and even substrates wherein an unknown defect distribution occurs can be automatically classified.;SOLUTION: One or more closed areas wherein any defect exists on each of substrates are extracted, and a group of defects occupying the closed areas is determined as a target defect cluster (S101). Knowledge information concerning the known defect cluster is prepared. The plurality of target defect clusters obtained in S101 are classified into a known group similar to the known defect cluster and an unknown group not similar to the known defect cluster (S102). The groups of substrates are classified into a known group and an unknown group (S103). The target defect clusters classified as the unknown group are classified into a plurality of sub-groups, so that the groups of substrates classified as the unknown group are classified into a plurality of sub-groups (S105).;COPYRIGHT: (C)2009,JPO&INPIT