Image analysis – Histogram processing – For setting a threshold
Patent
1989-03-29
1993-10-19
Mancuso, Joseph
Image analysis
Histogram processing
For setting a threshold
382 17, 382 30, 382 33, 358101, 358106, G06K 900
Patent
active
052553299
DESCRIPTION:
BRIEF SUMMARY
TECHNICAL FIELD
The present invention relates to a method of automatically detecting a fault of an image to be inspected at high speed as flexibly as a visual inspection in a printed image appearance inspection and to an apparatus for carrying out the method.
BACKGROUND ART
In a conventional printed image appearance inspection using image processing techniques, a printing fault is detected by a "pixel pair comparing method" of comparing corresponding pixels of a reference image and an image to be inspected.
In general, when an image to be inspected is to be imaged and input by a camera, a small level variation is produced due to a positional difference caused by an imaging system, noise of the imaging system or a brightness fluctuation of illumination light. Therefore, in the above simple pixel pair comparing method, a noncoincidence between pixel values is determined by this small level variation. As a result, data which is originally not defective is processed as defective data. In order to prevent such pseudo fault detection, in a conventional fault detecting method, a level value of each pixel of an image to be inspected is compared with those of pixels located within a predetermined range close to a corresponding pixel in a reference image. In this case, if a pixel having a level difference falling within an allowable range is present in the predetermined range of the corresponding pixel in the reference image, it is determined to be normal. Otherwise, the pixel is determined to be defective. This method is called close analogous pixel searching processing [reference literature: Masuda et al., "Printing Quality Inspection System", Mitsubishi Heavy Industries, Ltd. Technical Report, Vol. 23, No. 2, (1986-3), PP. 248 to 252].
In the above conventional fault detecting method, however, if a positional difference allowable amount of input data is .+-.n pixels in both X- and Y-axis directions, a level of each pixel of an object to be inspected must be collated with those of (2n+1).sup.2 pixels in a reference image. That is, a calculation amount is increased in an order of a square with respect to an increase in positional difference allowable value. The number n is practically at least two or more, and therefore level collation must be performed 25 times or more for each pixel. In addition, if a level variation is allowed and allowable levels of level values differ from each other, allowable amount setting must be checked for each pixel of image data having a large number of pixels, resulting in an increase in processing amount. Also, in order to perform an inspection as flexible as a visual inspection, an allowable limit of a printing distortion such as thickening/thinning of a character must be additionally taken into consideration, and therefore cumbersome processing in which a variety of allowable conditions are referred to must be performed. As a result, since a processing amount is increased, a high-speed fault inspection cannot be performed.
In the above conventional fault detecting method, a processing speed can be increased within a certain range by reducing the size of image data or processing all pixels under the same allowable amount conditions. In this case, however, a resolution of an image or a flexibility for a determination reference is degraded to make it difficult to detect a fault with high precision.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a method and apparatus which eliminate the above conventional drawbacks, i.e., which can perform fault detection/determination processing at high speed even when a mechanical positional difference produced upon inputting of an image to be inspected, a small positional difference of a pixel due to a printing distortion, or a level variation in pixel caused by a disturbance is allowed and complicated allowable conditions are set for each pixel.
The present invention has, for each pixel, a memory table including, as its elements, flags of binary values (ON, OFF) in a number corresponding to the number of densit
REFERENCES:
patent: 4644584 (1987-02-01), Nagashima et al.
patent: 4648053 (1987-03-01), Fridge
patent: 4742556 (1988-03-01), Davis, Jr. et al.
patent: 4841473 (1989-06-01), DePaul
patent: 4872203 (1989-10-01), Asai et al.
patent: 4893346 (1990-01-01), Bishop
Ishii Akira
Meguro Shinichi
Tanimizu Katsuyuki
Mancuso Joseph
Nippon Telegraph and Telephone Corporation
LandOfFree
High-speed fault detecting method and apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with High-speed fault detecting method and apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and High-speed fault detecting method and apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1358342