Abstract: A printed pattern is provided with a wiring pattern having a land and a through hole surrounded by the land. The wiring pattern and the through hole are read by a photoelectric image sensor to obtain a pattern image and a hole image, respectively. A pattern gauge image is defined by the pattern image, and a hole gauge image having a ring-like shape is obtained from the hole image. A logical product between the inversion image of the pattern gauge image and the hole gauge signal provides a protrusion region image. Since the hole gauge image has an isotropic contour, the minimum annular width of the land is detected, independent of the direction in which the through hole is deviated from the land.

Abstract: A method and apparatus for detecting neck disconnection between a land and a line in a wiring pattern on a printed board includes a scanner for obtaining images of the wiring pattern and a through hole formed in the pattern. The image of the wiring pattern is converted to a line image representing a wiring line by thinning the pattern image. A window encompassing the image of the through hole is defined, and part of the thinned pattern image overlapping the window is extracted. The number of end points of the thinned pattern image is counted in the window. If the number of the end points is an odd number larger than one, it is determined that neck disconnection is present in the wiring pattern.

Abstract: A frequency distribution or histogram M (L.sub.j-1, L.sub.j) of differences between gradation levels (L.sub.j-1, L.sub.j) of adjacent pixels (P.sub.j-1, P.sub.j) in an image is obtained, and the number of level-changes for each tentative threshold level L.sub.j is determined as a function of the histogram of the level differencies. By using the number of level-changes, an optimum binarization threshold level is set by means of an adaptive thresholding method. Since the number of level-changes is calculated from the histogram of the level-differencies without practical binarization for each threshold level, the amount of calculations is not dependent on the number of pixels but on the number of gradation levels. The calculations for setting the threshold level for the binarization of the image are achieved at high speeds by a small-scale hardware.

Abstract: Respective images of a conductive pattern and through holes of a printed circuit board are obtained. The through holes are classified into normal through holes having a relatively large diameter and mini via holes having a comparatively small diameter. The mini via hole image is enlarged to become an enlarged mini via hole image (SHSm). The image of the conductive pattern has hole parts therein, which are filled with images generated from the normal through hole image and the mini via hole image to obtain a corrected pattern image (CPS). The edge of the corrected pattern image is detected and enlarged to obtain an enlarged edge image (SES). The enlarged edge image is classified into two areas, one of which overlaps with the enlarged mini via hole image (SHSm). These two areas are subjected to inspection in different inspection modes.

Abstract: A printed circuit board (11) is provided with a conductive pattern (P, R) and a through hole (H) thereon. Respective images of the conductive pattern and a through hole are photoelectrically read with an image reader (20). The conductive pattern image is magnified in size to obtain a magnified pattern image (PI.sub.m), while the hole image is magnified at different magnification factors to obtain magnified hole images (HI.sub.n, HI.sub.n-i). A ring image (RP) is defined by the difference of the magnified hole images. The overlapped region (WR) on which the ring image and the magnified pattern image overlap each other is then calculated, and a defect or an opening of the conductive pattern is detected by calculating the aperture angle (.theta.) of the overlapped image.

Abstract: A printed circuit board (11) is provided with a conductive pattern (P, R) and a through hole (H) thereon. Respective images of the conductive pattern and a through hole are photoelectrically read with an image reader (20). The conductive pattern image is magnified in size to obtain a magnified pattern image (PI.sub.m), while the hole image is magnified at different magnification factors to obtain magnified hole images (HI.sub.n, HI.sub.n-i). A ring image (RP) is defined by the difference of the magnified hole images. The overlapped region (WR) on which the ring image and the magnified pattern image overlap each other is then calculated, and a defect or an opening of the conductive pattern is detected by calculating the aperture angle (.theta.) of the overlapped image.

Abstract: A device for inspecting a printed board has linear image sensors. The image sensors read the image of the printed board on respective image reading positions (CA0, CB0, CA1, CB1), while scanning the image. A predetermined tolerance range (RY) for image-reading positions is previously determined so that the respective image reading positions are included in the tolerance range. Image signals obtained in respective image sensors are delayed by respective delay times. The respective delay times are previously determined in proportion to the deviations between the image reading positions and a reference position (RLY1), where the reference position is defined on the rear end (EPY1) or behind the tolerance range in the scanning direction (-Y). Through the delay process, all of the image signals are corrected to compensate for the respective deviations in the image reading positions.

Abstract: A printed circuit board (11) is provided with a conductive pattern (P, R) and a through hole (H) thereon. Respective images of the conductive pattern and a through hole are photoelectrically read with an image reader (20). The conductive pattern image is magnified in size to obtain a magnified pattern image (PI.sub.m), while the hole image is magnified at different magnification factors to obtain magnified hole images (HI.sub.n, HI.sub.n-i). A ring image (RP) is defined by the difference of the magnified hole images. The overlapped region (WR) on which the ring image and the magnified pattern image overlap each other is then calculated, and a defect or an opening of the conductive pattern is detected by calculating the aperture angle (.theta.) of the overlapped image.

Abstract: The periphery of an area of a master pattern is expanded by a required number of pixels to define a plurality of master pattern areas two-dimensionally misregistered pixel by pixel, for performing pattern comparison of the respective master pattern areas with an area of an object pattern. The object pattern is deemed defective when all of the comparisons indicate pattern mismatches. The pattern is deemed not defective when at least one comparison indicates a pattern match. Thus, even if an inspected object includes misregistration errors, pattern defects can be detected with high accuracy.

Abstract: A binarized image data group of an inspected object for a plurality of pixels arrayed in the sub-scanning direction is sequentially created in order along the main scanning direction to determine whether or not the pattern of an inputted image corresponds to a standard through-hole pattern on the basis of the binarized image data group. When the determination is of yes, through-hole diameter masking data are sequentially created in order along the main scanning direction to sequentially perform masking processing on the binarized image data group of the inspected object arrayed in the sub-scanning direction. Thus, the through-hole pattern is automatically detected to be subjected to masking processing, whereby an inspected object having through-holes can be inspected by a pattern matching method.