Image processing method, substrate inspection method, substrate inspection apparatus and method of generating substrate inspection data

A substrate inspection method of obtaining a target image by causing a computer of an image processor for processing color images from a color image taking device to execute a program stored on a computer-readable memory medium and taking an image of a target substrate to be inspected, said target image being a color image which is a combination of color data corresponding to a plural n-number of different hues, extracting target component parts from the obtained target image and carrying out a specified inspection, said method comprising:

• a first step of obtaining a reference image from a substrate without having said target component parts thereon;
• a second step of thereafter setting a target area to be inspected at a position corresponding to one of said target component parts on said target image;
• a third step of considering those of said color data corresponding to specified less than n of said different hues and extracting difference both in magnitude and in direction in the degree of intensity of said considered color data relative to said combination of color data corresponding to n different hues between each pixel contained in said target area and a corresponding pixel on said reference image;
• and
• a fourth step of extracting those of said pixels of said target image for which the difference extracted in said third step is greater than a specified threshold value and is in a specified direction and carrying out image processing for said specified inspection on an image area comprised of the extracted pixels;
• each of said first step, said second step, said third step and said fourth step being carried out for each of said target component parts.

A substrate inspection apparatus comprising:

• an image generating part for generating a color image of a target substrate for inspection, said color image being a combination of color data corresponding to a plural n-number of different hues;
• an image processor for extracting a target component part for inspection from said color image of said target substrate generated by said image generating part and carrying out an image processing for inspection;
• a judging part for judging whether said target substrate is good or defective based on results of said image processing by said image processor;
• and
• an output part for outputting results of judging by said judging part;
• wherein said image processor has the functions of:
• color data processing by considering those of said color data corresponding to specified less than n of said different hues and calculating the degree of intensity of said considered color data relative to said combination of color data corresponding to n different hues for each pixel;
• registering in a memory a reference image obtained by said image generating part from a substrate not provided with any target component part or results of said color data processing function on said reference image;
• setting a target area to be inspected, when a target image of said target substrate for inspection is generated by said image generating part, at a position corresponding to a target component part in said target image;
• first extraction for having an image in said set target area processed by said color data processing function and using results of the process by said color data processing function and data stored by said memory to extract difference both in magnitude and in direction in the degree of intensity of said considered color data relative to said combination of color data corresponding to n different hues between each pixel in said target area and a corresponding pixel on said reference image;
• second extraction for extracting those of the pixels of said target image at which the change extracted by said first extraction function is greater than a specified threshold value and is in a specified direction;
• and
• recognizing an image area formed by the pixels extracted by said second extraction function as said target component part.

A method of generating inspection data for an automatic inspection of a substrate by causing a computer of an image processor for processing color images from a color image taking device to execute a program stored on a computer-readable memory medium and using a color image which is a combination of color data corresponding to a plural n-number of different hues;

• said method comprising:
• a first step of obtaining a reference image from a substrate without target component part thereon;
• a second step of obtaining a target image for inspection from a substrate having a target component part thereon;
• a third step of considering those of said color data corresponding to specified less than n of said different hues and extracting difference both in magnitude and in direction in the degree of intensity of said considered color data relative to said combination of color data corresponding to n different hues between each pixel of said target image and a corresponding pixel on said reference image;
• a fourth step of extracting those of said pixels of said target image for which the difference extracted in said third step is greater than a specified threshold value and is in a specified direction and recognizing an image area formed by the extracted pixels as said target component part;
• and
• a fifth step of setting a target area for inspection including said target component part recognized in said fourth step and including said set target area in said inspection data for said target component part.