JPS6349270B2 - - Google Patents

Description

Detailed Description of the Invention The present invention performs A/D (analog-digital) conversion of a video signal input by a photoelectric conversion scanner,
By using the digital signal level of a specific area of the object as a reference signal and correcting the gradation conversion characteristics, the signal level of the object is kept constant, and then,
The present invention relates to a method and apparatus for detecting defective locations on an object by comparing this signal level with the signal level of a standard pattern stored in advance. The object is positioned using the method, and then the specific area of the positioned object is used as a reference signal level to correct the gradation conversion characteristic curve, thereby detecting changes in contrast due to changes in the object. Furthermore, by determining the difference between the value corrected using this gradation conversion characteristic and the standard pattern stored in advance for each divided small section, the defective section can be corrected. The present invention relates to a method and apparatus for making a determination.

Conventionally, this type of device, for example, a device for inspecting labels or stamps with the registered trademark of a product, scans an optically scanned input video signal with binary values of "0" and "1" at a certain threshold level. A method has been used in which defects are detected by converting the pattern into a digital image and comparing the converted binary image with a standard pattern stored in advance. In this case “0”,
Since a binary image of "1" is used, the influence of variations in illumination unevenness, surface condition, etc. of the target object becomes noticeable, and due to this influence, the standard pattern stored in advance There is a difference between
Misrecognition of defects has been a problem.

In order to eliminate the above-mentioned drawbacks, the present invention converts the input video signal into a multi-level signal obtained by A/D conversion, and furthermore, in order to eliminate the influence of the above-mentioned illumination unevenness, surface condition, etc., the input video signal is converted into a multi-level signal. By performing gradation conversion of the input video signal using the A/D converter output signal level of the converted object as a reference signal, the signal level of the object is always kept constant, and then the signal level of the standard pattern is Defects are determined by determining the difference between the two.

The present invention will be described in detail below with reference to the drawings.

FIG. 1 is a diagram for explaining positioning of an object using a binary image input by a photoelectric conversion scanner (not shown), where 1 indicates the entire scanning area of the object, and 2 shows, as an example, a mark used for positioning an object. The target object is positioned with high precision using a known positioning method that performs mask matching using the positioning marks shown in FIG. 2 or the characteristic regions of the target object in the input binary image from the photoelectric conversion scanner. Perform proper positioning. Next, a standard pattern signal level and a reference signal level for gradation conversion correction are extracted from the positioned object. FIG. 2 is a diagram for explaining how the signal level of the standard pattern is extracted from the object positioned by the method described above. In the figure, the horizontal position Ho indicates the position on the line segment AA' in FIG. 1, and the output video signal level Vo indicates the output signal level of the photoelectric conversion scanner at each position. As a standard pattern, a digital signal obtained by A/D converting the signal level of the object is used as a RAM ( Random Access M) as a grid point sampled at certain intervals in the horizontal and vertical directions.
memory) etc. At the same time, the signal level after A/D conversion of the specific area is extracted and stored, which is used as a reference signal for correcting the gradation conversion characteristics.

Next, the inspection algorithm for the object to be inspected is
As shown in the figure. As shown in the figure, by the above method, 2
The object is positioned using the digitized photoelectric conversion scanner output signal, and then the gradation conversion characteristics are corrected using the positioned specific area of the object. FIG. 4 is a diagram for explaining a method of correcting tone conversion characteristics. In the figure, the horizontal axis indicates the input signal level, which is a digital input signal, and the vertical axis indicates the corresponding digital output signal level. The gradation conversion method uses a known digital gradation conversion that converts digital input into digital output. In the figure, 3 indicates the gradation conversion characteristic when storing the standard pattern, and 4 indicates the reference signal level of a specific area that characterizes this characteristic. 5 indicates the reference signal level for the object to be inspected. In this way, when the same gradation conversion characteristic is used, if the input signal level changes depending on the appearance of the object, the output signal level will also change significantly in accordance with the change. Therefore, as shown in FIG. 7, the gradation conversion characteristics are changed so that the output signal level remains constant regardless of changes in the input signal level, and characteristic curve 6 is used. By changing the gradation conversion characteristics in this manner, an output signal level that is not affected by changes in the input signal level due to changes in the surface condition of the object or illumination can be obtained.

Next, FIG. 5 shows a method of determining the defect position by determining the difference between the output signal level and the standard pattern for each small section of the object. In the figure, 8 indicates the entire scanning area of the object to be inspected, and 9
each indicates a small section in which defect determination is performed. Defect determination is performed by calculating the difference in the signal level of the standard pattern for each of the subsections with respect to the sample values of the grid points that are the same as the method used to capture the standard pattern, and then determining the difference in the signal level of the standard pattern for each subsection. This is done depending on whether the threshold is exceeded. In the figure, numeral 10 indicates an example of a section determined to be defective.
In this way, by using the difference between multilevel signals, erroneous recognition due to noise is reduced compared to the case where the binary signal is used as it is to perform defect determination.

FIG. 6 is a block diagram showing a defect detection apparatus constructed using the defect detection method described above. In the figure, a video signal inputted by a photoelectric conversion scanner 11 is converted into a binary signal by a binarization circuit 12, and positioned by a positioning circuit 13 using a known method. Next, in order to extract the reference signal level from this positioned object, a region is designated by the reference signal region designation circuit 13. On the other hand, the input video signal is converted into a digital signal through the A/D conversion circuit 16, and this converted signal becomes an input signal to the gradation conversion circuit 17, where the gradation is converted. This gradation conversion circuit output signal is controlled by the output of the reference signal area specifying circuit, and the reference signal level extraction circuit 15 extracts the reference signal level. This reference signal level is fed back to the gradation conversion correction circuit 18 to correct the gradation conversion circuit 17 in order to correct the gradation conversion characteristics as described above.

Reference numeral 20 denotes a standard pattern storage circuit composed of a RAM or the like that stores a standard pattern that becomes a reference when determining defects in an object. The control circuit 19 performs signal control when the output signal of the gradation conversion circuit 17 is taken into the standard pattern storage circuit 20 as a standard pattern, and when the difference between the signal levels of the standard pattern and the object is calculated for each small section. .

For defect determination, the gradation conversion output signal and the standard pattern signal level are successively compared for each small section divided by the area division circuit 21, and the difference calculation circuit 22
This is done by finding the difference. The determined difference is compared with a threshold level 25 applied from the outside by a determination circuit 23, and a determination signal 24 is output based on the magnitude relationship with this value. Further, the section signal 25 corresponding to this determination signal is transmitted to the control circuit 1.
It is output from 9.

As described above, according to the present invention, by correcting the gradation conversion characteristic curve using a specific area of the object to be inspected as the reference signal level, it becomes less susceptible to changes in the surface condition of the object or uneven illumination.
In addition, unlike the case of using the conventionally used binary pattern of "0" and "1", A/D converted multi-value level is used, and defect determination is performed by dividing the object. Since this is performed based on the difference between the small section and the standard pattern, it is possible to eliminate erroneous recognition due to the influence of noise.

[Brief explanation of the drawing]

Figure 1 shows the input data from the photoelectric conversion scanner.
A diagram for explaining positioning using a valued image, FIG. 2 is a diagram for explaining how to extract the signal level of a standard pattern from an object positioned by the above method, and FIG. Figure 4 is a diagram showing the inspection algorithm for the object to be inspected. Figure 4 is a diagram explaining the method of correcting the gradation conversion characteristics. Figure 5 is a diagram showing the difference between the signal level of the object to be inspected and the standard pattern. FIG. 6 is a block diagram showing a defect detection apparatus constructed using the defect detection method described above. In the figure, 4 is the reference signal level of a specific region characterizing the characteristic curve, 5 is the reference signal level for the object to be inspected, 6 is the gradation conversion characteristic for the object to be inspected, 7 is the above-mentioned object to be inspected, 11 12 is a photoelectric conversion scanner, 12 is a binarization circuit, 13 is a positioning circuit, 14 is a reference signal area specifying circuit, 16 is an A/D conversion circuit, 17 is a gradation conversion circuit, 18 is a gradation conversion correction calculation circuit, 22 is a difference calculation circuit, 23
2 is a defect determination circuit, 24 is a defect determination signal, and 25 is a threshold signal level for defect determination.

Claims (1)

[Scope of Claims] 1. Positioning of an object is performed using any plurality of specific regions of the binarized image detected by a photoelectric conversion scanner, and the positioning of the plurality of specific regions of the positioned object is performed. The gradation conversion characteristics are corrected using the video signal level as a reference signal level, the corrected gradation conversion characteristics are used to keep the video signal level of the object constant, and then The signal level of the standard pattern that was
Defect detection characterized by determining the difference between the signal level after the gradation conversion for each of the plurality of divided small sections, and determining whether the small section is defective depending on whether the value exceeds a certain threshold. Method. 2. a photoelectric conversion scanner; a binarization circuit that binarizes an image detected by the photoelectric conversion scanner; a positioning circuit that positions an object using the binarized image; and the photoelectric conversion scanner. an analog-to-digital conversion circuit that converts the video signal of the object into a digital signal; an area designation circuit that designates a specific area of the positioned object; a signal detection circuit that determines the signal level within the area; a gradation conversion calculation circuit that calculates a correction value for gradation conversion characteristics from the signal level; a gradation conversion circuit that performs gradation conversion of the video signal; and a standard that stores the signal after gradation conversion in advance. A defect detection device comprising a difference calculation circuit that calculates the difference with a signal level of a pattern for each divided small section, and a defect determination circuit that determines a defect depending on whether the difference calculation output is larger than a certain threshold value.