JP2003203218A - Visual inspection device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To shorten the inspection time without increasing the mounting scale by collectively performing the visual inspections of two or more shapes. <P>SOLUTION: This visual inspection device for determining the quality of an object of inspection by inputting the image data of the object of inspection and comparing the standard data of the object of inspection with the input image data comprises an adjacent shape integrating means for integrating two or more adjacent inspection object shapes of the input image data to one inspection object shape, and an inspection means. The inspection means compares the standard data with the integrated shape data. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a visual inspection apparatus and a visual inspection method, and in particular, it inspects shapes (objects to be inspected) of IC packages, terminals on printed circuit boards, and the like.
The present invention relates to an appearance inspection device and an appearance inspection method.

[0002]

2. Description of the Related Art Generally, in a visual inspection apparatus, C
The standard data including the shape information of the appearance inspection target and the position information of the shape is generated from the design data such as AD and held, and when the image data to be inspected is input by the imaging system including the camera, the input is performed. Shape data including shape information and shape position information in the image data is extracted.

FIG. 2 shows an example of an outline of extraction of an input image and an inspection target shape in the appearance inspection apparatus. In this example, in the input image from the camera, three terminals of the shape A to be inspected,
In addition to one terminal of shape B, the surrounding unnecessary patterns not included in the inspection target are included. The appearance inspection apparatus generates an area signal to be inspected from standard data held in advance, removes a pattern unnecessary for inspection other than a position where the shape to be inspected should be, and extracts shape data to be inspected.

Next, the extracted shape data is compared with the standard data to detect defects such as chipping / protrusion and disappearance of the shape,
Pass / fail judgment is performed. FIG. 3 shows a defect example to be detected and a small modification example not requiring detection. A slight deformation of the terminal shown on the upper left side does not need to be detected because it has no problem in operation, but a large deformation on the lower left side and a missing (disappearing defect) of the right side terminal need to be detected.

FIG. 4 shows a conventional example, which has three shapes A (p1 to p) shown in FIG. 3 as an input image to be inspected.
3) shows the case where the shape B is one (p4). here,
The area signal generating means generates an area signal of the shape of the inspection object from the standard data in advance, the shape extracting means extracts the shape in the area of the inspection object from the input image data, and compares each shape with the standard data to determine the quality.

[0006]

However, in general, there are a plurality of shapes (terminals) to be inspected in the input image, and in order to execute the inspections at the same time, the number of shape extraction means and area signals as many as the shapes to be inspected. Generating means are needed. In the case of the example of FIG. 4, four sets of shape extraction means and area signal generation means are required. Therefore, when the scale that can be realized is limited due to the mounting conditions of the visual inspection apparatus, the number of shapes of the inspection target that can be processed at the same time is limited, and the inspection may not be possible depending on the inspection target.

On the contrary, when the inspection of each shape is performed in time series, the processing time becomes longer as the number of inspection objects increases.
An object of the present invention is to perform visual inspections of a plurality of shapes at once, and to shorten the inspection time without increasing the mounting scale.

[0008]

SUMMARY OF THE INVENTION The present invention is an appearance inspection apparatus for inputting image data of an inspection object, comparing standard data of the inspection object with input image data, and determining pass / fail of the inspection object. An adjoining shape unifying means for unifying a plurality of adjoining shapes in the image data into one shape, and an inspecting means, and the inspecting means compares the standard data with the integrated shape data. And

FIG. 1 shows the outline of the structure of an appearance inspection apparatus according to the present invention. Similar to the conventional example, standard data including shape information of a visual inspection object and shape position information is previously generated from design data such as CAD and is stored. Adjacent shapes in the image of the input image data are integrated by the adjacent shape integration means 1, and integrated shape data and an area signal of the inspection target are generated.

The inspection means 2 compares the shape data in the area to be inspected with the standard data to judge the quality,
Output the inspection result. In this case, in the area to be inspected, the comparison of a plurality of standard data and the shape data can be collectively processed, so that the processing time can be shortened. The comparison by the inspection unit 2 may be pattern matching between the standard data and the shape data, or may be comparison between the feature extracted from the standard data and the feature of the shape data.

Further, in the above invention, the standard shape unifying means for unifying a plurality of adjacent standard shapes in the standard data into one standard shape is provided, and the integrated standard data and the integrated shape data are compared. Is also good. This makes it possible to refer to a plurality of integrated standard data only once, and the processing time can be shortened.

Further, by extracting the area of the integrated shape as the area to be inspected, it is possible to limit the area to be referred in the inspection processing, and the processing time can be shortened.

[0013]

FIG. 5 shows a first embodiment of the appearance inspection apparatus. The input image data is converted into binarized image data by the binarizing means a. In the binarized image data, adjacent shapes are integrated by the expansion / contraction means b. Figure 6
The operation of expansion / contraction processing is shown. In the expansion / contraction means b,
First, expansion processing is performed a predetermined number of times.

FIG. 7 shows an example of an expansion / contraction processing algorithm. Here, black pixels are "1" and white pixels are "0".
It is represented by. In the expansion processing shown in FIG. 7A, if there is at least one "1" pixel among the eight pixels surrounding the pixel to be processed (target pixel), the processing of replacing the target pixel with "1" is performed for all pixels. Perform sequentially while paying attention. Here, the predetermined number of times is determined under the conditions shown in FIG. 3 in which defects (shapes) to be detected are left, small deformations that do not need to be detected are eliminated, and adjacent shapes to be integrated are connected. For example, 11
When the deformation of more than pixels is detected and the shapes with a distance of 10 pixels or less are integrated, the predetermined number of times is five times.

By the above-described expansion process five times, the shape defect within 10 pixels and the shape gap (white pixel) are filled with black pixels. In this way, shape defects within 10 pixels disappear, adjacent shapes are connected, and the outer shape is thickened by 5 pixels to form one integrated shape. Next, the contraction process is performed a predetermined number of times on the image subjected to the expansion process. For example, as shown in FIG. 7-b, if there is at least one pixel of "0" among the eight pixels surrounding the pixel of interest, the process of replacing the pixel of interest with "0" is sequentially performed while paying attention to all the pixels.

After the contraction process 5 times, the small deformations disappear, the connected shapes remain integrated, and the outer shape is thinned by 5 pixels, that is, the original thickness before the expansion process. The integrated shape data is input to the inspection target shape detection means c, and the integrated shape data of the inspection target is detected. The area signal generating means d generates an area signal of the shape of the inspection object representing the peripheral region from the integrated shape detected by the inspection object shape detecting means c.

FIG. 8 shows an example in which the above-described expansion / contraction processing is performed on the shape to be inspected shown in FIG. 4 to integrate adjacent shapes. In this example, the shape p5 in which the shapes of p1 to p4 are integrated and the area signal of the inspection target shape including p5 are generated. FIG. 8 shows an example in which adjacent shapes are integrated, but when the shapes to be inspected are discretely arranged, a plurality of integrated shapes and area signals are generated.

Next, in the logical product e, the logical product of the area signal and the integrated shape data is taken, the pattern outside the area signal is removed as shown in FIG. 9, and only the integrated shape data is input to the image memory f. To do. The pattern unnecessary for the inspection may be removed before the expansion / contraction processing according to the area signal of the inspection target shown in FIG.

The judging means g reads the integrated shape data from the image memory f, compares it with the standard data, and judges the quality. By providing a plurality of determination means g and performing a plurality of determination processes in parallel, the speed can be further improved.
As shown in FIG. 10, when there is a defect in which some of the shapes of the inspection object have disappeared, the integrated shape is different from the shape assumed from the standard data, and therefore the area signal of the correct inspection object shape is generated. Not done. Therefore, the determination unit g refers to the standard data, searches the image memory f for the position where the shape of the inspection target should be, and determines that there is no shape disappearance defect when the shape data does not exist.

Alternatively, the standard data used for comparison with the shape read from the image memory f is recorded, and unused standard data remains after the comparison is completed for all the shapes in the image memory f. May be determined to be a shape loss defect. For defects that cannot be detected by the integrated shape, the binarized shape is extracted with the logical product h, and the entire function is interpolated by comparing the binarized shape data with the standard data as in the conventional example. May be.

FIG. 11 shows a second embodiment of the appearance inspection processing device provided with the standard shape integrating means. In this embodiment, the standard shape is integrated as well as the shape in the input image data to be inspected, so that the comparison process can be simplified and the speed is further improved.

[0022]

As described above, the shapes of a plurality of adjacent inspection objects are integrated to form one inspection object. The scale of the appearance inspection device can be reduced. 2. The amount of data to be inspected can be reduced and speed can be improved. 3. The process is simplified and speed is increased.

[Brief description of drawings]

[Figure 1] Configuration overview of appearance inspection device

[Fig. 2] Example of outline of extraction of input image and inspection target shape data

FIG. 3 is a defect example to be detected and a small modification example not requiring detection.

FIG. 4 Conventional example

FIG. 5: Example 1

FIG. 6 Operation of expansion / contraction processing

FIG. 7 Algorithm for expansion / contraction processing

[Figure 8] Shape integration

FIG. 9: Extraction of inspection target shape

FIG. 10: Shape loss defect

FIG. 11: Example 2

[Explanation of symbols] 1 Adjacent shape integration means 2 inspection means 3 Standard shape integration means a Binarization means b Expansion / contraction means c Inspection target shape detection means d area signal generation means e, h AND f image memory g Judgment means

Continued front page    (72) Inventor Susumu Haga             4-1, Kamiodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa             No. 1 within Fujitsu Limited F term (reference) 2F065 AA51 BB02 CC01 CC28 DD06                       FF04 FF61 JJ03 JJ26 QQ04                       QQ24 QQ25 QQ31 QQ32 QQ39                       RR08                 2G051 AA62 AA65 AB02 AB20 AC21                       EA11 EA30 ED04                 5B057 AA03 BA29 DA03 DB02 DB08                       DC09 DC33                 5L096 AA07 BA03 EA02 EA43 FA18                       GA22 HA07

Claims (5)

[Claims] 1. An appearance inspection apparatus for inputting image data to be inspected and comparing standard data of the inspected object with the input image data to judge pass / fail of the inspected object, wherein a plurality of adjacent image data in the input image data are included. An appearance inspection apparatus comprising: an adjoining shape unifying means for unifying the shapes of 1 into one shape; and an inspecting means, wherein the inspecting means compares standard data with integrated shape data. 2. The appearance inspection apparatus according to claim 1, further comprising a standard shape unifying means for unifying a plurality of adjacent standard shapes in the standard data into one standard shape. Is an appearance inspection device characterized by comparing integrated standard data with integrated shape data. 3. An appearance inspection apparatus for inputting image data of an inspection object and comparing standard data of the inspection object with the input image data to judge pass / fail of the inspection object, wherein a plurality of adjacent image data in the input image data are included. Has an adjoining shape unifying means for unifying the shapes of 1 into one shape, and an inspecting means, wherein the adjoining shape unifying means extracts an area to be inspected from the integrated shape data. apparatus. 4. The appearance inspection apparatus according to claim 2, wherein the inspection means sequentially compares the integrated standard data with the integrated shape data, and there is no integrated shape data corresponding to the integrated standard data. In this case, the appearance inspection device is characterized by a defect. 5. A visual inspection method for inputting image data of an inspection target and comparing standard data of the inspection target with the input image data to determine whether the inspection target is good or bad, which is a method of adjoining a plurality of adjacent input image data. The inspection target shape of 1
A visual inspection method characterized by integrating two shapes to be inspected and comparing the standard data with the integrated shape data.