Facsimile and static presentation processing – Static presentation processing – Data corruption – power interruption – or print prevention
Reexamination Certificate
1998-06-01
2002-04-02
Garcia, Gabriel (Department: 2624)
Facsimile and static presentation processing
Static presentation processing
Data corruption, power interruption, or print prevention
C358S001900
Reexamination Certificate
active
06366358
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a streak-like defect inspecting method for a printed matter and device therefor, an edge region extracting method for use in the inspection and device therefor, a binary image preparing method and a device therefor, an image inputting method for use in printed matter inspection and device therefor, and a plain region selecting method and a device therefor, and more particularly to a streak-like defect inspecting method for a printed matter and a device therefor, which is preferably applied to the detection of a fine streak-like defect caused on the printed matter, an edge region extracting method for use in the inspection and device therefor, which is preferably applied to the image processing carried out in the above inspection, a binary image preparing method and device therefor, which is applied in the same manner, an image inputting method for use in printed matter inspection and device therefor, which is preferably applied to the inspection of the printing condition, with a pattern printed on a running original piece as an inspection subject, based on an image to be inspected input by a camera which is movable in the width-direction of the original piece, and a plain region selecting method and device therefor, which select a plain region from the whole pattern on a printed matter, and which is preferably applied to the inspection of the printing condition based on an image to be inspected in which the selected plain region is input as an inspection point by a partially inputting camera disposed on an inspection device.
BACKGROUND OF THE TECHNIQUE
Generally, a web-like gravure rotary press continuously carrying out the printing by transferring ink adhered to a rotary print drum while passing an original piece through between the print drum and a nip roll pressing it.
In order to realize such a continuous printing, the ink is supplied to the surface of the rotating print drum and then the surplus of the ink is scraped by the doctor blade.
When the doctor blade, however, is minutely broken for some reason, ink is permanently left, although a little, on the print drum contacting to the broken portion, thereby resulting in the continuous, but fine, streak-like printing defect, what is called the doctor streak, on the printing surface of the original piece.
FIG. 1
conceptually shows an image to be inspected input to inspect whether or not such a doctor streak is caused, and the features thereof. FIG.
1
(A) shows an input image corresponding to one example of a longitudinally added image described later in detail. The image has a longitudinally continuous streak portion L and noises N on a substantially flat background. Accordingly, as there are shown in FIG.
1
(B), FIG.
1
(C), and FIG.
1
(D) the graduation values (brightness values) obtained on lateral scan lines (I), (II), (III) shown in FIG.
1
(A), the peaks P
1
are found out at substantially the same position corresponding to the streak portions of all the profiles, and the peaks Pn resulting from the noises N are found out at the different positions.
Conventionally, printed matters have been inspected by inputting an image of a pattern to be printed, by photographing means such as a CCD camera and then detecting printing defects through the image processing of the input image.
There is known that the image processing used for detecting such printing defects includes the application of a variety of differential operators for detecting lines from the input image, and the techniques of the Hough transformation, the pattern recognition, and the like.
However, according to the above-mentioned line detecting method of the conventional image processing, it is difficult to discriminate and then detect the edge of the pattern and the streak-like defect from the image including the input pattern. Moreover, many of streak-like defects, such as the doctor streaks, extending in the same direction are low in the contrast between the streak portion and the surroundings, which provides problems that it is difficult to detect the lines, and more difficult to detect them when noises are included in the image.
Moreover, the image processing for use in such printing defect detection includes the edge extraction for detecting the edge as the density boarder of the image from the input image. This edge extraction is executed by applying a variety of the differential operators to the image data comprising the respective pixel values of the input image, and further such operators include the Prewitt operator and the Sobel operator for a primary differential, the Laplacean operator for a second differential, and the like.
However, the above-mentioned edge detecting method according to the conventional image processing undesirably extracts the edge for all the pattern portion in the input image; therefore, applying it to the printing defect inspection carried out based on the input image results in the misjudgment.
In other words, in order to detect only the printing defects by preparing a mask for excluding the edge portion from the input image to be inspected and then applying the mask to the image to be inspected, applying the above-mentioned operators to the mask preparing processing causes, when the image to be inspected includes streak-like defects such as the doctor streaks, the defects also to be processed as is the case with the pattern. Therefore, applying the prepared mask to the above-mentioned image to be inspected causes the streaks also to be undesirably masked, which disables the streak-like defects to be detected.
The above-mentioned situation will now be described more concretely taking the conventional mask preparing method as an example.
Suppose that the image to be inspected input by the camera is schematically shown by the pattern of FIG.
2
(A). The image shown in FIG.
2
(A) comprises four kinds of patterns having widths of a to d, respectively, which are arranged at regular intervals e, in which the narrowest line having the width of d corresponding to the doctor streak. Further, FIG.
2
(B) shows the graduation values of the respective pixels with respect to the horizontal direction of FIG.
2
(A).
First, as shown in
FIG. 3
as FIG.
3
(C) while showing the image to be inspected in FIG.
3
(A), applying the Laplacean operator or the like to the image causes the edge to be extracted; then the edge regions are set at the both sides of the edge up to several pixels.
Next, as shown in
FIG. 4
while showing the respective edge regions of the above FIG.
3
(C), the expansion-processing causes these regions to be enlarged toward the both sides thereof by several pixels, which brings about the condition shown in FIG.
4
(D). Thereafter, the contraction-processing causes these regions to be returned to the condition which is equal in pixel number to that of FIG.
3
(C), which is made into a mask image. On this occasion, subjecting the pattern of the width c in the image of FIG.
2
(A) to the expansion-processing causes two of the edge regions to be connected as shown in FIG.
4
(D), which connected portion is not subjected to the contraction. FIG.
5
(E) shows the case in which the mask images prepared as described above are overlapped on the image to be inspected of FIG.
2
(A).
As can be seen from
FIG. 5
, when preventing the edges of the pattern from being detected by applying the mask images prepared by the use of the conventional edge extracting method, the edge regions shown in FIG.
3
(C) entirely include the doctor streak of the width d shown in FIG.
2
(A), which disables the doctor streak to be detected.
Moreover, the presence of the streak portion on the input image such as the above-mentioned imaged to be inspected is expected to be judgment-processed based on the binary image prepared by binary-coding the pixel values. As to the binary-coding processing method of preparing the binary image used at that occasion, there is used as the most general method a fixed threshold value processing method of binary-coding the respective pixel values of the image by the us
Hayashi Kenta
Sakata Hideto
Satou Hiroshi
Soeda Masahiko
Garcia Gabriel
Oliff & Berridg,e PLC
LandOfFree
Method and apparatus for detecting stripe defects of printed... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for detecting stripe defects of printed..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for detecting stripe defects of printed... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2847891