JP3370608B2 - Image quality inspection device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image quality inspection apparatus, and more particularly, to a semiconductor / component / finished product inspection process for a video IC mounting inspection process, a video device completion inspection process, and a video-related component inspection process. The present invention relates to an image quality inspection device used for function tests.

[0002]

2. Description of the Related Art Conventionally, in order to enhance the reliability of the functions of video equipment, parts, and devices, a mounting inspection has been performed in which an object to be inspected is actually operated. The mounting inspection is to inspect an output image of the mounted inspection object by a method in which a person visually sees an image output from the inspection object. V
FIG. 4 shows a method of mounting inspection of the TR device.

In FIG. 4, an IC 21 to be inspected is an IC to be inspected. The VTR 22 is a device that creates an actual operating environment for inspecting the IC 21 to be inspected. Test head 23
Is a device having contacts and peripheral circuits for pulling out a portion of the IC to be inspected 21 to be incorporated in the VTR 22 and easily exchanging the IC to be inspected 21 for each inspection. Video signal 24
Is a video signal output from the IC 21 to be inspected, the VTR 22, and the test head 23. The monitor 25 is a display device that displays the video signal 24. The human 26 visually observes the image displayed on the monitor 25 by the image signal 24 continuously for an arbitrary time, and inspects for the presence or absence of a defective image caused by the defect of the IC 21 to be inspected.

Conventionally, the object to be inspected has been inspected by visual sensory inspection of the image output from the object to be inspected.

[0005]

However, when the image quality is inspected and measured by the above-mentioned conventional inspection method, the difference in the inspection result due to the variation of the judgment by the person, the difference in the inspection accuracy due to the difference in the skill of the inspection, and the difference in the inspection accuracy. However, there is a problem that efficiency decreases due to monotonous labor and cost increases due to personnel expenses.

The present invention is to solve the above problems by automating such a conventional human visual inspection method, and it is an object of the present invention to realize an image quality inspection apparatus capable of detecting defective images in the same manner as humans. And

[0007]

In order to solve the above-mentioned problems, according to the present invention, an image quality inspection apparatus is provided with a contour evaluation processing unit for performing evaluation by differentiating a difference between image data to be inspected and reference image data. , And a level comparison processing section for performing integration and evaluation. With this configuration, it is possible to automatically perform the image quality inspection, which is almost the same as the visual inspection of the image by the human senses.

Further, the image quality inspection apparatus separates the composite signal, which is the image to be inspected, into RGB signals, digitally quantizes the RGB signals, and differentiates the difference from the previously quantized and stored reference image data. A contour evaluation processing section for performing evaluation and a level comparison processing section for performing evaluation by integration are provided. With this configuration, it is possible to automatically perform the quality inspection of the image in the composite signal format.

Further, the image quality inspection apparatus is digitally quantized for an RGB signal which is an image to be inspected, and a contour evaluation processing section for performing evaluation by differentiating the difference from the previously quantized reference image data, and integrated and evaluated. And a level comparison processing unit for performing. With this configuration, it is possible to automatically perform the quality inspection of the image in the RGB signal format.

[0010]

BEST MODE FOR CARRYING OUT THE INVENTION The invention according to claim 1 of the present invention is
Image to be inspected, which is image data output from the inspected device
Regarding the data string and the standard image data string that is the standard for inspection
Te, each optionally inspection image data stream is divided into a reference image data stream and the difference calculation unit calculating the difference for each corresponding image data, before and after the data sequence of any of the output data from the difference calculation unit an averaging filter unit for averaging at width, a differential processing unit for calculating a difference obtained by subtracting from the output of the data processed by the averaging filter unit from the difference calculation unit, any output of the differential processing unit a contour comparison processing section having a function of comparing the first threshold set, and the reference image average value calculating unit for calculating an average value of the reference image data stream, checking for calculating the average value of the inspection image data stream Picture
An image average value calculation unit, a division unit that divides the output of the reference image average value calculation unit by the output of the inspection image average value calculation unit , and a data of the image data string to be inspected at the output of the division unit.
Subtract the data of the reference image data string from the value multiplied by
Arbitrarily setting the square sum calculator for calculating the sum of squared obtained was data value difference, the output of the square sum calculating section was
An image quality inspection apparatus comprising: a level comparison processing unit having a function of comparing with the second threshold value; and an evaluation unit that evaluates image quality based on a comparison result of the level comparison processing unit and the contour comparison processing unit. By simultaneously inspecting the difference in the waveform contour information and the difference in the waveform level, the level of the image quality inspection can be brought close to the level of the inspection by human vision, and the image quality inspection can be automated.

According to a second aspect of the present invention, the device to be inspected is
Image data string that is the image data output from the
And a circuit for separating a composite signal, which is an image to be inspected, into RGB with respect to a reference image data string serving as an inspection reference ,
An A / D converter that digitally quantizes the separated RGB analog signals or the RGB analog signals that are images to be inspected, a storage circuit that stores the converted RGB data, and an image that serves as an inspection reference. RGB
A memory circuit for storing the respective data, each corresponding image of the subject <br/>査image data stream and the reference image data sequences read from the memory circuit for each horizontal line delimiting each RGB
A difference calculation unit that takes a difference for each image data, an averaging filter unit that adds and averages data strings before and after each output data from the difference calculation unit in an arbitrary width, and the averaging filter
The data processed by the section from the output from the difference calculation section
A differential processing unit that takes the subtracted difference and the output of the differential processing unit
A contour comparison processing unit having a function of comparing the first threshold set arbitrarily, and the reference image average value calculating unit for calculating an average value of the reference image data stream for each RGB, the inspection image
Inspection image average value calculator for calculating the average value of the image data sequence
When the division unit for dividing the output of the reference image average calculator at the output of the test images average calculator, the division unit
From the value obtained by multiplying the output by the data of the inspection image data string
And square sum calculator for calculating the sum of the serial reference image data stream squared obtained was data values the difference obtained by subtracting the data of the
An image quality based on the comparison result of the level comparison processing unit having a function of comparing the output of the square sum calculation unit with a second threshold value set arbitrarily, and the level comparison processing unit of each of the RGB and the contour comparison processing unit. An image quality inspection apparatus including an evaluation unit for evaluating the image quality. By simultaneously inspecting the difference in the waveform contour information and the difference in the waveform level of the image of the composite signal format or the RGB format, the level of the image quality inspection is visually inspected by human eyes. It has the effect of making it possible to automate the image quality inspection by approaching the standard.

FIG. 1 shows an embodiment of the present invention.
~ It will be described in detail with reference to FIG.

(First Embodiment) In the first embodiment of the present invention, the difference between the image data sequence to be inspected and the reference image data sequence is differentiated and the result is compared with a contour threshold to obtain a contour comparison result. Further, the image quality inspection apparatus evaluates the image quality based on the result of level comparison by comparing the result of integral processing of the difference with the level threshold value.

FIG. 1 is a block diagram of an image quality inspection apparatus according to the first embodiment of the present invention. In FIG. 1, the reference image data sequence 1 is reference image data. The inspected image data sequence 2 is image data output from the inspected device. The contour comparison processing unit 3 is means for differentiating the difference between both image data and performing threshold processing. The level comparison processing unit 4 is means for performing threshold processing by integrating the difference between both image data. The comprehensive evaluation unit 5 is means for performing comprehensive evaluation based on the determination results of both processing units.

The operation of the image quality inspection apparatus of the first embodiment having the above-described configuration will be described below. Reference image data sequence 1 and inspection image data sequence 2
Is input to the contour comparison processing unit 3 and undergoes a differentiating process as shown by the following equation.

[Equation 1]

However, STD (i) is the i-th data of the reference image data sequence 1. DUT (i) is the i-th data in the inspection image data sequence 2 . STD
(J) is the j-th data of the reference image data sequence 1. DUT (j) is the j-th data in the inspection image data sequence 2 . W is a differential processing constant. H (i)
Are output values of the i-th contour processing section for the reference image data sequence 1 and the inspected image data sequence 2.

The contour comparison processing is performed by the i-th data of the reference image data sequence 1 and the i-th data of the inspection image data sequence 2.
Difference from the second data

[Equation 2] Thus, the contour information of the inspected image data sequence 2 is removed from the contour information of the reference image data sequence 1 .
Furthermore, the jth {j = (i−W) to (i +) of the reference image data sequence 1 within the ± Wth centered on the i th respectively.
W)} data and the j-th {j = (i−W) to (i + W)} data of the inspected image data string 2 are averaged.

[Equation 3] And the differential processing constant W
While excluding low frequency components that indicate the overall brightness of the image in the contour information of the inspection image data sequence 2,
It is possible to extract contour shape difference information. Further, H (i) obtained by the equation (1) is compared with an arbitrary threshold value for each data element, and if any data is larger than the threshold value, the image quality is rejected. If not, the image quality is determined to be good and the image is output. The above processing contents are shown in the form of a graph in FIG.

The differential processing constant W can increase or decrease the sensitivity of the difference between the contour shapes to be extracted. When the differential processing constant W is small, it becomes blunt for the extraction of contour shape difference information, and when the differential processing constant W is large, it becomes sensitive to the extraction of contour shape information. The ability to extract shape difference information can be adjusted.

At the same time, the reference image data sequence 1 and the inspected image data sequence 2 are input to the level comparison processing unit 4 and subjected to integration processing as shown by the following equation.

[Equation 4]

[Equation 5]

[Equation 6]

However, STD (i) is the i-th data of the standard image data sequence 1. DUT (i) is the i-th data in the inspection image data sequence 2 . N is the number of data of the reference image data sequence 1 and the inspected image data sequence 2. E is the reference image data sequence 1 and the inspection image data sequence 2
Is an output value of the level comparison processing unit 4 for.

In this level comparison process, the coefficient obtained by dividing the level average value of the reference image data sequence 1 by the level average value of the inspection image data sequence 2 is the i-th coefficient of the inspection image data sequence 2. By integrating the difference between the value multiplied by the data and the i-th data of the reference image data sequence 1, the level difference between the reference image data sequence and the inspected image data sequence is calculated as the overall brightness of each image. It is possible to correct the difference in height and integrate and extract. Further, E obtained by the equation (6) is compared with an arbitrary threshold value, and if it is larger than the threshold value, the image quality is judged to be negative, and if not, the image quality is judged to be good and the E is output.

Next, the judgment output from the contour comparison processing unit 3 and the judgment output from the level comparison processing unit 4 are input to the comprehensive evaluation unit 5. Only when each judgment output determines that the image quality is good, it is determined that the image quality is good, and otherwise, it is determined that the image quality is bad.

As described above, in the first embodiment of the present invention, the image quality inspection apparatus is configured to perform comparison processing by paying attention to both the contour and level information of the image to be inspected. , It is possible to make a judgment that correlates with the human visual sense and to calculate an evaluation value.

(Second Embodiment) In a second embodiment of the present invention, an analog signal obtained by separating a composite signal which is an image to be inspected into RGB or an analog signal of each RGB which is an image to be inspected is digitally quantized. , compared to the threshold differential of the difference of the inspected image data stream and the reference image data string to a contour comparison result, the level comparison result the square sum of the difference is compared with a threshold, the image quality on the basis of both the comparison result Is an image quality inspection device for evaluating.

FIG. 3 is a block diagram of an image quality inspection apparatus according to the second embodiment of the present invention. In FIG. 3, the inspected image composite signal 11 is a composite video signal output from the inspected device. The RGB separator 12 is
This is a circuit for separating the composite signal into RGB signals.
The A / D converter 13 is a circuit that digitally quantizes an analog signal separated into RGB signals. The frame memory 14 is a memory which stores each digitally quantized RGB data. The reference image memory 15 is a memory that stores each RGB data that is reference image data. The contour comparison processing unit 16 is means for differentiating the difference between the two image data for each RGB and performing threshold processing. Level comparison processing unit 17
Is means for performing threshold processing by integrating the difference between both image data for each RGB. The comprehensive evaluation unit 18 is means for performing comprehensive evaluation based on the determination results of both processing units.

The operation of the image quality inspection apparatus according to the second embodiment having the above-described structure will be described below.

The inspection image 11 which is a composite signal is
It is input to the RGB separator 2 and separated into RGB signals.
Next, the RGB signal output from the RGB separator 12 is A
The data is input to the / D converter 13 to be digitally quantized data. It is stored in the frame memory 14 which stores each of RGB. In addition, the reference image data memory 15 stores in advance a reference image in the same manner as the inspection image 11 and stores it.

Next, the reference image data sequence and the inspected image data sequence read out for each horizontal line from the RGB frame memory 14 and the reference image memory 15 are R
Each of the GBs is input to the contour comparison processing unit 16 and subjected to a differentiating process as shown by the following equation.

[Equation 7]

However, STD (i) is the i-th data of the reference image data string. DUT (i) is the i-th data in the inspection image data sequence. STD (i) is the j-th data in the reference image data string. DUT (j)
Is the j-th data in the inspection image data sequence. W
Is a differential processing constant. H (i) is an output value of the i-th contour processing section for the reference image data sequence and the inspected image data sequence.

These contour comparison processing units 16 are provided for the i-th data of the reference image data string and the i-th data of the inspected image data string.
Difference from the second data

[Equation 8] Thus, the contour information of the inspected image data string is removed from the contour information of the inspected image data string. Furthermore, the jth {j = (i−W) to (i + W)} of the reference image data sequence within the ± Wth centered on the i th respectively.
Data and the jth {j = (i-
W) to (i + W)} average value of the difference

[Equation 9] And the difference information of the contour shape while removing the low frequency component indicating the overall brightness of the image in the contour information of the inspection image data sequence 2 by the differential processing constant W. Can be extracted. further,
H (i) for each RGB obtained by the equation (1) is compared with an arbitrary threshold value prepared for each RGB for each data element, and there is at least one data larger than the threshold value. Determines that the image quality is not good, and otherwise determines that the image quality is good and outputs it. The differential processing constant W can increase or decrease the sensitivity of the difference in the contour shapes to be extracted. When the differential processing constant W is small, it is blunted for the extraction of the difference information of the contour shape, and when the differential processing constant W is large, it is sensitive to the extraction of the information of the contour shape. Therefore, the contour shape can be correlated with the human visual sense. It is possible to adjust the difference information extraction ability of the.

At the same time, the reference image data sequence and the inspected image data sequence read out for each horizontal line from the RGB frame memory 14 and the reference image memory 15 are R
It is input to the level comparison processing section 17 of each GB and subjected to integration processing as shown by the following equation.

[Equation 10]

[Equation 11]

[Equation 12]

However, STD (i) is the i-th data of the reference image data string. DUT (i) is the i-th data in the inspection image data sequence. N is the number of data in the reference image data sequence / inspected image data sequence. E
Are output values of the level comparison processing unit 4 for the reference image data sequence and the inspected image data sequence.

In this level comparison processing, the coefficient obtained by dividing the level average value of the reference image data sequence by the level average value of the inspection image data sequence is multiplied by the i-th data of the inspection image data sequence. By correcting the difference between the value and the i-th data of the reference image data string, the level difference between the reference image data string and the inspected image data string is corrected, and the difference in the overall brightness of each image is corrected. Then, it can be integrated and extracted. Further, E of each RGB obtained by the equation (6) is compared with an arbitrary threshold value prepared for each RGB, and if it is larger than the threshold value, the image quality is rejected, and if not, the image quality is good. Judge and output.

The above contour comparison processing 16 and level comparison processing 17
Is repeated for an arbitrary number of horizontal lines forming one screen.

Next, the judgment output for any number of horizontal lines from the contour comparison processing unit 16 for each of RGB and the judgment output for any number of horizontal lines from the level comparison processing unit 17 for each of RGB are comprehensively evaluated. Input to the part 18. Only when each determination output determines that the image quality is good, it is determined that the image quality is comprehensively good.

As described above, in the second embodiment of the present invention, the image quality inspection apparatus uses the R image having the image to be inspected.
Since the comparison processing is performed by focusing on both the contour and level information of the GB component, it is possible to calculate an evaluation value that is correlated with the sensory test by human vision, and to automatically judge the quality of the image. Can be done.

[0037]

As described above, according to the present invention, the image quality inspection apparatus compares the contour of the image to be inspected with a reference image and judges and extracts the contour, and the level information of the image to be inspected. Since the image quality is determined by using the two comparison results of the level comparison processing that extracts and determines by comparing with the reference image, the image quality determination that correlates with the sensory test by human vision is automatically performed. The effect that the image quality inspection device to be performed can be easily configured is obtained.

[Brief description of drawings]

FIG. 1 is a block diagram of an image quality inspection device according to a first embodiment of the present invention,

FIG. 2 is an explanatory diagram of data processing in a contour comparison processing unit according to the first embodiment,

FIG. 3 is a block diagram of an image quality inspection device according to a second embodiment of the present invention,

FIG. 4 is a schematic view of a conventional image quality inspection device.

[Explanation of symbols]

1 Standard image data string 2 Inspection image data string 3 contour comparison processing unit 4 level comparison processing unit 5 Comprehensive evaluation department 11 Inspected video 12 RGB separator 13 A / D converter 14 frame memory 15 Reference image data memory 16 Contour comparison processing unit 17 Level comparison processing unit 18 Comprehensive Evaluation Department 19 outputs 21 IC to be inspected 22 VTR (Video Tape Recorder) 23 test head 24 video signal 25 monitors 26 human

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) G01J 3/46-3/51 G01J 1/00 G01J 1/42-1/44 G01R 31/02 G01R 31 / 28 G01M 11/00 G01N 21/84-21/958 G06T 7/00-7/60 H01L 21/66 H04N 17/00-17/06

Claims (2)

(57) [Claims] 1. Image data output from a device under inspection
A certain image data sequence to be inspected and a reference image data to be used as an inspection reference.
Respect chromatography data sequence, a difference calculation section that takes a <br/> differences for each corresponding image data of arbitrarily divided inspection image data stream and the reference image data stream, for each output data from the difference calculation unit an averaging filter unit for averaging the data sequence before and after in any width, and the differential processing unit the data that is processed by the averaging filter unit takes the difference obtained by subtracting from the output from the difference calculation section, the differential Set the output of the processing unit arbitrarily
Starring the contour comparing unit, and the reference image average value calculating unit for calculating an average value of the reference image data stream, the average value of the inspection image data string having a function of comparing a first threshold and
An inspection image average value calculating unit, a division unit that divides the output of the reference image average value calculating unit by the output of the inspection image average value calculating unit, and an output of the division unit to the inspection image data
Data multiplied by the data of the reference image data
And square sum calculator for calculating the sum of the difference obtained by subtracting the over data squared obtained was data value, the output of the square sum calculating section
A level comparison processing unit having a function of comparing with a second threshold value set arbitrarily; and an evaluation unit for evaluating image quality based on a comparison result of the level comparison processing unit and the contour comparison processing unit. Image quality inspection device 2. Image data output from the device under inspection
A certain image data sequence to be inspected and a reference image data to be used as an inspection reference.
Regarding the data sequence, a circuit that separates the composite signal that is the image to be inspected into RGB, an A / D converter that digitally quantizes the separated analog signal of each RGB or the analog signal of each of RGB that is the image to be inspected, and conversion a storage circuit for storing RGB respective data, a memory circuit for storing the RGB data of each of the image as a reference of the inspection, the each RGB read out from the memory circuit for each horizontal line delimiter A difference calculation unit that takes a difference for each corresponding image data of the inspection image data sequence and the reference image data sequence, and averaging that averages the data sequences before and after each output data from the difference calculation unit in an arbitrary width. a filter unit, a differential processing unit for the treated with averaging filter unit data taking the difference obtained by subtracting from the output from the difference calculation unit, said differential processing unit A contour comparison processing section having a function of comparing the first threshold value which is arbitrarily set the output for each RGB
Be computed reference image average value calculating unit for calculating an average value of the reference image data stream, the average value of the inspection image data stream
Inspection image average value calculator, and the reference image average value calculator
Of the output of the inspection image average value calculation unit , and the output of the division unit, the inspection image data string
From the value obtained by multiplying the data of
And square sum calculator for calculating the sum of squared obtained was data values the difference obtained by subtracting the arbitrary output of the square sum calculating section
A level comparison processing unit having a function of comparing with a second threshold value set to the above, and an evaluation unit that evaluates image quality based on a comparison result of the level comparison processing unit of each of the RGB and the contour comparison processing unit. An image quality inspection device characterized by.