JPH0815172A - Visual inspection method - Google Patents

Abstract

PURPOSE:To sufficiently delete a region not to become an object to be inspected by automatically and suitably providing a threshold value for binarizing in response to the object to be inspected. CONSTITUTION:An inspection candidate region including at least a region of an object to be inspected is set in a gray level image obtained by imaging a spatial region including the object to be inspected by imaging means (S1). Then, density threshold values WS, BS in which at least one of a high density side and a low density side is deviated by a specified offset value with respect to the mean density of pixels included in the candidate region is set (S2), and the pixels in the candidate region are binarized by using the values WS, BS (S3). A label is applied to each coupled component of the deleted candidate pixel extracted from the candidate region according to the binarizing (S4). The label is compared with the known label of the object to be inspected, the deleted region is decided (S5), and the region in which the deleted region of the candidate regions is deleted is used as the region to be inspected (S6).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect such as a scratch or stain on the appearance of an object to be inspected or an object to be inspected on the basis of a grayscale image obtained by imaging a spatial region including the object to be inspected by an image pickup means. The present invention relates to a visual inspection method for inspecting a shape and the like.

[0002]

2. Description of the Related Art Generally, in this type of appearance inspection method, a spatial region including an object to be inspected is imaged by an image pickup device such as an ITV camera, and the density of each pixel in the picked-up image (that is, by the image pickup device). The normal portion of the inspection object and the defect or background of the inspection object are identified based on the received light amount). That is, a threshold value is set for the pixel density and binarization is performed using this threshold value to separate the normal portion of the inspection object from other portions. In order to facilitate this separation, it is also common to irradiate the object to be inspected with light from an appropriate direction to cause an illuminance difference between a normal part of the object and other parts. ing.

[0003]

By the way, since the threshold value for binarization is fixedly set in the past, the reflectance of the inspected object is different, or the inspected object has a different reflectance. If the light amount of the irradiated light changes, there is a problem that binarization that correctly separates a normal portion of the inspection object from other portions cannot be performed. In other words, there is a problem that the threshold value must be manually adjusted when the appearance of different inspected objects is inspected or when the irradiation light amount to the inspected object changes. In particular, if the density difference (that is, the contrast) of the portions to be separated is relatively small, there is a problem that it becomes difficult to set the threshold value.

Further, when inspecting the appearance of an object to be inspected, the number of pixels to be processed is reduced and the processing efficiency is improved by removing the area which does not need to be inspected from the image. A mask including only an area is set and the presence or absence of a defect and the shape are recognized only in the mask. However, for an object to be inspected having a complicated contour line including a bent portion, a mask is set to include only the object to be inspected, or a portion corresponding to a hole, a groove, or a protruding portion formed in the object to be inspected. It is difficult to set a mask that removes only the target from the inspection target and leaves scratches and stains as the inspection target, and there is a problem that the processing efficiency cannot be sufficiently enhanced.

The present invention has been made in view of the above problems, and automatically sets a threshold value for binarization in accordance with an object to be inspected and determines a normal portion of the object to be inspected. An object of the present invention is to provide an appearance inspection method in which an appropriate threshold value is set even when the density difference from other portions is relatively small, and a region which is not an inspection target is sufficiently excluded to improve processing efficiency.

[0006]

Claims 1 to 5
In order to achieve the above-mentioned object, the invention sets an inspection candidate area including at least a partial area of the inspection object in a grayscale image obtained by capturing an image of a spatial area including the inspection object by an image capturing unit, and performs inspection. In the method of inspecting the appearance of the object to be inspected in the candidate area, the specified density is shifted to at least one of the high density side and the low density side with respect to the average density of the pixels included in the inspection candidate area. The common configuration is to set the density threshold and binarize the pixels in the inspection candidate area using the density threshold to extract the exclusion candidate pixels from the inspection candidate area.

According to the first aspect of the present invention, a connected component, which is a portion occupied by a region in which only exclusion candidate pixels are connected, is compared with known information, and it is determined for each connected component whether or not the exclusion region is excluded from inspection. Next, an area other than the exclusion area in the inspection candidate area is set as an inspection target area, and the appearance of the inspection object is inspected only in the inspection target area. According to the invention of claim 2, a connected component in which an area of a connected component that is a portion occupied by a region connected with only exclusion candidate pixels is larger than a prescribed area threshold is determined as an excluded region to be excluded from the inspection target,
Next, an area other than the exclusion area in the inspection candidate area is set as an inspection target area, and the appearance of the inspection object is inspected only in the inspection target area.

According to a third aspect of the present invention, a connected component in which the number of pixels included in a connected component, which is a portion occupied by a region in which only exclusion candidate pixels are connected, is larger than a prescribed area threshold, and the connected component is For each pixel on the edge, the density difference between a pair of pixels equally spaced in the direction orthogonal to the direction indicated by the differential direction value of the pixel across the pixel exceeds the specified value. The connected component that exceeds the value is determined as the exclusion area to be excluded from the inspection target, and then the area excluding the exclusion area of the inspection candidate area is set as the inspection target area,
It is characterized in that the appearance of the object to be inspected is inspected only within the area to be inspected.

According to a fourth aspect of the present invention, a connected component in which the number of pixels included in a connected component, which is a portion occupied by a region in which only exclusion candidate pixels are connected, is larger than a prescribed area threshold, and the connected component is The connected component whose sum of the differential values of each pixel on the edge exceeds the specified differential threshold is determined as the exclusion area to be excluded from the inspection target, and then the area excluding the exclusion area among the inspection candidate areas is the inspection target area. In addition, the appearance of the inspection object is inspected only within the inspection target area.

According to a fifth aspect of the present invention, a connected component in which the number of pixels included in a connected component that is a portion occupied by a region in which only excluded candidate pixels are connected is larger than a specified area threshold, and the connected component is When the sum of the differential values of each pixel on the edge exceeds the specified differential threshold, and for each pixel on the edge of the connected component, the pixel is sandwiched in the direction orthogonal to the direction indicated by the differential direction value of that pixel, etc. A connected component in which the density difference between a pair of pixels separated by a distance exceeds a predetermined value exceeds a specified pixel count threshold is determined as an exclusion area to be excluded from the inspection target, and then the exclusion area of the inspection candidate area is determined. The area except for is used as the inspection target area, and the appearance of the inspection object is inspected only within the inspection target area.

[0011]

According to the structure common to claims 1 to 5, the high-density side and the low-density side of the average density of the pixels included in the inspection candidate area including at least a partial area of the object to be inspected. An exclusion candidate pixel is extracted from the inspection candidate area by setting a density threshold value shifted to at least one by a specified offset value and binarizing the pixel in the inspection candidate area using the density threshold value. Therefore, the density threshold value used for binarization is automatically set according to each inspection object. Moreover, the inspection candidate area is set to include at least a part of the inspection object, and the density threshold value is set based on the average density of the inspection candidate area. It is possible to set a density threshold value having an appropriate value so that the two can be distinguished from each other even when the density difference between the other area and the other area is relatively small.

According to the configuration of the first aspect of the present invention, the connected component of the exclusion candidate pixel obtained by binarization is compared with the known information, and it is determined for each connected component whether or not it is an exclusion region to be excluded from the inspection target. However, since the inspection target area is an area excluding the exclusion area from the inspection candidate area, an area other than the inspection object such as a hole or a protrusion of the inspection object or a conveyor that conveys the inspection object is removed from the exclusion candidate pixel. It is possible, and as a result, it is possible to sufficiently exclude the region that is not the inspection target, which leads to improvement in processing efficiency.

According to the configuration of the second aspect of the present invention, the connected component in which the area of the connected component of the exclusion candidate pixel obtained by binarization is larger than the specified area threshold is set as the excluded region to be excluded from the inspection target. The inspection target area is determined and the area excluding the excluded area is set as the inspection target area.Defects such as scratches and stains have an area larger than the holes or protrusions formed on the inspection object or the background of the inspection object. Since the defect is small, defects such as scratches and stains can be left in the inspection target area, but holes, protrusions, and the background can be excluded, which leads to improvement in processing efficiency.

According to the third aspect of the present invention, the number of pixels included in the connected component of the exclusion candidate pixels obtained by binarization is a connected component larger than a specified area threshold, and the connected component is also connected. For each pixel on the edge of the component, the number of pixels where the density difference between a pair of pixels equidistant in the direction orthogonal to the direction indicated by the differential direction value of that pixel exceeds the predetermined value. The connected component that exceeds the threshold value is determined as an exclusion area that is not to be inspected, and since the area excluding the exclusion area in the inspection candidate area is set as the inspection area, not only the area of the connected component of the exclusion candidate pixel, By using the information about the difference in density between the inside and outside of the connected component of the exclusion candidate pixel together, the inspection target area and the exclusion area can be separated more reliably than the invention of claim 2.

According to the structure of the invention of claim 4, the number of pixels included in the connected component of the exclusion candidate pixels obtained by binarization is a connected component larger than a specified area threshold, and the connected component is connected. The connected component whose sum of the differential values of each pixel on the edge of the component exceeds the specified differential threshold is determined as the exclusion area to be excluded from the inspection target area, and the area excluding the exclusion area from the inspection candidate area is the inspection target area. Therefore, not only the area of the connected component of the exclusion candidate pixel but also the information on the differential value on the edge of the connected component of the exclusion candidate pixel is used together, and the inspection target area and the exclusion area are claimed. It becomes possible to separate more reliably than the second invention.

According to the configuration of the fifth aspect of the present invention, the number of pixels included in the connected component of the exclusion candidate pixels obtained by binarization is a connected component that is larger than a specified area threshold, and the connected component is also connected. The direction in which the sum of the differential values of each pixel on the edge of the component exceeds the specified differential threshold and the direction of each pixel on the edge of the connected component is orthogonal to the direction indicated by the differential direction value of that pixel. Connected components in which the density difference between a pair of pixels that are equidistant from each other exceeds a specified value are determined as excluded regions that are not to be inspected, and are excluded from the inspection candidate regions. The region except for is used as the inspection target region, and by adopting both the density difference inside and outside the connected component of the exclusion candidate pixel and the differential value on the edge used in the inventions of claim 3 and claim 4, , Inspection area It is possible to more reliably separate the exclusion area.

[0017]

【Example】

(Embodiment 1) This embodiment corresponds to the invention of claim 1, and FIG.
The area to be inspected is determined by the procedure shown in the flowchart in FIG. When inspecting the appearance of the object to be inspected, the object to be inspected is positioned at a predetermined position, and a spatial region including the object to be inspected is imaged by an imaging device such as an ITV camera. The output of the image pickup device is subjected to A / D conversion, and a grayscale image having a density value corresponding to the amount of received light is stored for each pixel in the frame memory. Here, when the imaging device captures an image of a spatial region including the inspection object, the inspection object is irradiated with light at a predetermined angle. The subsequent processing is performed on the grayscale image stored in the frame memory.

[0018] The object to be inspected Q ₁ groove R ₂ is formed, an image of only the object to be inspected Q ₁ is assumed to be as shown in Figure 2.
First, the inspection candidate area D ₁ as shown in FIG.
Is set (S1 in FIG. 1). The inspection candidate area D ₁ may have a shape that can be easily set, and is set to a rectangular shape here. Further, the inspection candidate region D ₁ may be set to include at least a portion of the object to be inspected Q _1, except when performing a visual inspection for the entire region of the object Q _1,
Although it is not always necessary to include the entire inspection object Q ₁ , it is usually desirable to make the ratio of the area occupied by the inspection object Q ₁ in the inspection candidate area D ₁ sufficiently larger than other areas. Here, since the inspection object Q ₁ is positioned at a predetermined position and the imaging device is fixed at a fixed position, the inspection candidate area D ₁ can be fixedly set regardless of the inspection object. However, the shape of the inspection candidate area D ₁ may be changed according to the inspection object Q ₁ . In the image, as shown in FIG. 3, a conveyor R ₁ that is in contact with the contour line of the inspection object Q ₁ and becomes a part of the background, and a groove R formed in the inspection object Q _1
2 etc. are included. Region to be inspected, the inspection because we want to set the excluded areas such as conveyor R ₁ and groove R ₂ from the candidate region D _1, from the examination candidate region D ₁ by the following procedure conveyor R ₁ and groove R _2, etc. Is excluded as an exclusion area. However, as shown in FIG. 4, when there is a defect R ₀ such as a scratch or stain, the defect R ₀ must be excluded from the exclusion area.

Therefore, first, the average value AV of the densities of the pixels included in the inspection candidate area D ₁ is obtained. This average value AV is
The value should be between the density of the normal area of the inspection object Q ₁ and the density of the area to be excluded. For example, when the background becomes a shadow, the density of the background area becomes lower than the average value AV (the density is higher as the amount of light received by the image pickup device is higher), and the groove R ₂ or Depressed defect R
_It is considered that ₀ is also lower than the average value AV. On the other hand, the average value AV
There is a possibility that a portion having a higher density than the average value AV may be generated depending on how the inspection object Q ₁ is illuminated. Therefore, if the average value AV is binarized as a threshold value for the density, it is considered that the exclusion candidate pixels that should be excluded from the region to be inspected can be extracted. Even a region may be included in the exclusion candidate pixel. In order to prevent such an inconvenience, in the present invention, the offset values WO and BO are set on the high density side and the low density side of the average value AV, and only the offset values WO and BO are set. By shifting the average value AV, the light side density threshold value WS (= AV + WO)
And the dark side density threshold BS = (AV-BO) are set (S2 in FIG. 1).

As described above, since the ratio of the inspection object Q ₁ to the inspection candidate area D ₁ is set sufficiently large, the density of the pixels included in the inspection object area of the inspection object Q ₁ is set. Becomes a value close to the average value AV. Therefore, when the density of the exclusion candidate pixel is relatively high, the pixel having the density higher than the density threshold value WS is set as the exclusion candidate pixel by binarizing by using the light side density threshold value WS. When the density of the candidate pixel is relatively small, binarization is performed by using the dark side density threshold BS, and the pixel having the density smaller than the density threshold BS is set as the exclusion candidate pixel. Alternatively, the exclusion candidate pixel may include both a portion having a relatively high density and a portion having a relatively low density, and in this case, as shown in FIG. 6, between the density thresholds WS and BS. Pixels that are not density are excluded candidate pixels (hatched portion in FIG. 6) (S3 in FIG. 1).

As described above, the concentration thresholds WS, BS
The exclusion candidate pixels obtained by binarizing by using are usually a plurality of pixels adjacent to each other to form a group. In the field of image processing, a portion occupied by a connected region consisting of the same pixel value (0 or 1) in a binary image is called a connected component. That is, since the connected component D ₂ of the exclusion candidate pixel is formed by the above-described binarization and a plurality of connected components D ₂ are formed in the inspection candidate region D ₁ as shown in FIG. A label is given to each D ₂ (S4 in FIG. 1). In this way, each connected component D
_At the stage where the label ₂ is attached, the label is attached to the defect R ₀ as well, as shown in FIG. 4 (the connected component to which the label is attached is surrounded by an alternate long and short dash line).

Since the defect R ₀ must be included in the region to be inspected, in order to obtain the region to be inspected, each connected component D ₂ except the defect R ₀ in the inspection candidate region D ₁ is removed. There is a need. In order to realize this processing, a label set in advance for the inspection object Q ₁ having no defect R ₀ is used as known information, and if there is a connected component D ₂ having a label that does not correspond to the known label, based on position information or the like. It is regarded as a defect R ₀ . Defect R ₀ thus obtained
If the connected component D ₂ of the other exclusion candidate pixels except for is set as the exclusion region (S5 in FIG. 1) and the exclusion region is removed from the inspection candidate region D ₁ , the remaining region becomes the inspection target region D ₃ (FIG. 1
S6). That is, in FIG. 5, the connected component D ₂ enclosed by the alternate long and short dash line in the area inside the solid line is the exclusion area, and the portion excluding the exclusion area is the inspection target area D ₃ . That is, the conveyor R ₁ or the background or the groove R ₂ is excluded from the inspection target area D ₃ as an exclusion area, and the defect R ₀ is not included in the exclusion area and is included in the inspection target area D ₃ . If the image processing according to the inspection item is performed on the inspection target area D ₃ obtained in this way, the processing amount required for the inspection can be reduced and the processing efficiency can be improved.

(Embodiment 2) This embodiment corresponds to the invention of claim 2, and in Embodiment 1, the density threshold value W
A label is given to each connected component D ₂ obtained by binarization using S and BS, and the label of the inspection object Q ₁ having no defect R ₀ is used as known information, and the inspection object to be inspected By comparing the label set for Q ₁ with a known label, the exclusion area to be the conveyor R ₁ or the groove R ₂ was identified from the connected components D ₂ of the exclusion candidate pixels. It is necessary to set the label of the inspection object Q ₁ having no defect R ₀ for each Q ₁ . In this embodiment, a method in which this work is unnecessary will be shown.

In this embodiment, the connected component D of the exclusion candidate pixel is
The procedure until obtaining ₂ is the same as that of the first embodiment, and only the procedure of identifying the excluded region to be the conveyor R ₁ or the groove R ₂ from the connected component D ₂ is different. That is, as shown in FIG. 7, steps S1 to S4 are performed in step S1 of FIG.
The same process as 1 to S4. When the connected component D ₂ is labeled, the number of pixels included in each connected component D ₂ is counted. For example, if a label image (an image in which a numerical value serving as a label is given as a pixel value of each pixel) is used, this work is realized by counting the number of pixels having the same numerical value. Here, the pixel density is the number of pixel density thresholds WS larger connected components D ₂ and MW, concentration of pixels and MB the number of pixel density thresholds BS smaller connected components D _2. The number of pixels included in each connected component D ₂ is approximately proportional to the area of the connected component D ₂ . On the other hand, the conveyor R ₁ and grooves because the area of R ₂ is generally considered sufficiently compared to the defect R ₀ large connected component D corresponding to the conveyor R ₁ and groove R ₂ by the area of the connected component D _2
It is thought that ₂ can be identified. Therefore, the area thresholds MWS and MBS are set to appropriate values that enable this identification, and the connected components D ₂ to which each label is attached are
The number of pixels MW, MB is compared with the area thresholds MWS, MBS (S5 in FIG. 7). MW> MWS or M
For the connected component D ₂ that satisfies B> MBS, the defect R
_When it is determined that the area is not _0, it is determined as an exclusion area (S in FIG.
6) Then, this exclusion area is excluded from the inspection area D ₃ (S7 in FIG. 7). Here, when the connected component D ₂ of the exclusion candidate pixel is extracted using only one of the bright side density threshold value WS and the dark side density threshold value BS, the area threshold values MWS and MBS are also extracted. Only one should be used. The other processing procedure is the same as that of the first embodiment, and the image processing according to the inspection item is performed on the inspection target area D ₃ determined as described above.

As described above, in the present embodiment, the area for each connected component D ₂ of the exclusion candidate pixel is used to identify the exclusion area such as the conveyor R ₁ or the groove R ₂ and the defect R ₀ . If the area thresholds MWS and MBS are appropriately set, it is not necessary to obtain information about the inspection object Q _{1 in} advance, and the area can be set more easily.

(Third Embodiment) This embodiment corresponds to the invention of claim 3, and in the second embodiment, the conveyer R ₁ or the like is calculated based on the area (the number of pixels) of the connected component D ₂ of the exclusion candidate pixel. Although the portion to be the groove R ₂ was identified, in some cases, there is also a defect R ₀ having an area similar to that of the conveyor R ₁ or the groove R ₂ . Therefore, in the following embodiments, a method of more surely determining the exclusion area by adding the density difference between the inside and outside of the connected component D ₂ of the exclusion candidate pixel as the identification condition will be described. Further, even when the density difference is relatively small, the exclusion area can be determined by emphasizing the density difference.

In this embodiment, the connected component D of the exclusion candidate pixel is
_In order to evaluate the density difference between inside and outside of ₂ , the density difference between a pair of pixels sandwiching the edge of the connected component D ₂ of the exclusion candidate pixel is obtained. This procedure will be described below. Step S1 of FIG.
As shown in S5, the procedure for comparing the number of pixels MW, MB of the connected component D ₂ of the exclusion candidate pixel with the area thresholds MWS, MBS is as follows:
This is the same as the procedure shown in S5. Next, the number of pixels MW, MB
Is a connected component D whose is greater than the area thresholds MWS, MBS
_{For 2} , the differential direction value of each pixel on the edge is obtained, and the density difference between a pair of pixels at equal distances from the pixel in the direction orthogonal to the direction indicated by the differential direction value is obtained (S6 in FIG. 8). . The number of pixels on the edge where the density difference exceeds the predetermined value is counted, and the pixel number K for which the density difference exceeds the predetermined value for each connected component D ₂ is compared with a prescribed pixel number threshold KS ( S7 of FIG. 8). When the pixel number K exceeds the pixel number threshold value KS (that is, K> KS), the exclusion area is set (S8 in FIG. 8). As shown in FIG. 9, the differential direction value is a direction orthogonal to the direction in which the density difference in the vicinity of the pixel of interest (for example, in the vicinity of 8) is maximum (pixels are indicated by white circles in FIG. The arrow corresponds to the differential direction value of the pixel on the edge), and the continuous edge corresponds to the direction along the edge. Therefore, a direction code as a direction connecting adjacent pixels on an edge may be used instead of the differential direction value. To obtain the density difference between a pair of pixels x ₁ and x ₂ located at the same distance from the pixel on the edge in the direction orthogonal to this direction, and evaluate the density difference between the inside and outside of the connected component D ₂ by this density difference. Can be done. In addition, by using the number of pixels whose density difference exceeds a predetermined value, the density difference is emphasized even when the density difference is relatively small or the density difference varies depending on the location. Judgment becomes possible.

As described above, since each connected component D ₂ of the exclusion candidate pixel is evaluated as to whether or not it is an exclusion region by using the number of pixels and the density difference, known information for each different inspection object Q ₁ is obtained. Compared with the second embodiment, no work is required, and even if the conveyor R ₁ , the groove R ₂ , and the like have the same size as the defect R ₀ , the two can be distinguished by the density difference, which is more reliable. The exclusion area can be determined. Subsequent processing is the same as that of the first embodiment, and the area excluding the exclusion area from the inspection candidate area D ₁ is the inspection target area D 1.
_{As 3} (S9 in FIG. 8), the required image processing may be performed.

(Embodiment 4) This embodiment corresponds to the invention of claim 4, and in the embodiment 3, as the evaluation of the density difference between the inside and outside of the connected component D ₂ of the exclusion candidate pixel, it is equidistant from the pixel on the edge. Although the difference in density between a pair of pixels positioned at is used, this embodiment shows an example in which the difference in density is evaluated by using the differential value of each pixel on the edge.

That is, steps S1 to S1 in FIG.
S5 is the same as steps S1 to S5 in FIG. 8 of the third embodiment.
The same (that is, the same as in the second embodiment), and
Minute D ₂The number of pixels MW and MB for each
Compare with WS and MBS. Next, the number of pixels MW and MB are
Connected component D larger than product thresholds MWS, MBS₂To
About the connected component D₂Differential value for pixels on the edge of
For all pixels on the edge
to add. This added value C is compared with the specified differential threshold CS
(S6 in FIG. 10), the added value C is the differential threshold CS.
If it exceeds (ie, C> CS), its connected component
D₂Is determined as an exclusion area (S in FIG. 10).
7). In short, the differential value is
Since the degree of change in concentration is shown, the larger the differential value, the more connected
Ingredient D₂It is possible to judge that there is a large difference in density between
The density difference can be evaluated by using the differential value.
Become Noh. Also, the derivative value is applied to all pixels on the edge.
And add and compare the added value C with the differential threshold CS
, The density difference is emphasized, and the connected component D₂And around
Be sure to identify even if the difference in concentration with
Can be done. After the exclusion area is determined, the inspection candidate area
D₁Since the inspection target area is set excluding the exclusion area from
Yes (S8 in FIG. 10), other procedures and actions are described in the embodiment.
Same as 3.

(Embodiment 5) This embodiment corresponds to the invention of claim 5, and in order to evaluate the density difference, the density difference between a pair of pixels equidistant from the edge used in Embodiment 3 has a predetermined value. By evaluating the density difference using both the number of pixels on the edge that exceeds the threshold and the sum of the differential values of the pixels on the edge used in the fourth embodiment, the density change in the vicinity of the edge of the connected component D ₂ is evaluated. When,
By using both the density difference at a portion slightly away from the edge, it is possible to make a more accurate determination than in the case of performing the evaluation alone.

That is, as shown in FIG. 11, steps S1 to S5 are steps S1 to S in FIG. 8 of the third embodiment.
5, which is the same as that of Example 5 (that is, the same as in the second embodiment), and the number of pixels MW and MB for each connected component D ₂ of the exclusion candidate pixel is obtained and compared with the area thresholds MWS and MBS. next,
For the connected component D _{2 in} which the number of pixels MW, MB is larger than the area thresholds MWS, MBS, the differential direction value of each pixel on the edge is obtained, and the pixel is equidistant from the pixel in the direction orthogonal to the direction indicated by the differential direction value. The density difference between a pair of pixels at positions separated from each other is obtained (S6 in FIG. 11). The number of pixels on the edge where the density difference exceeds the predetermined value is counted, and when the number K of pixels for which the density difference exceeds the predetermined value for each connected component D ₂ exceeds the prescribed pixel number threshold KS ( That is, K> K
S), the differential value is obtained for the pixels on the edge of the connected component D ₂ , and this differential value is added for all the pixels on the edge. This added value C is compared with a prescribed differential threshold CS (S7 in FIG. 11), and when the added value C exceeds the differential threshold CS, the connected component D ₂ is determined as an exclusion area. (S8 of FIG. 11). The exclusion area determined in this way is excluded from the inspection candidate area D ₁ to be the inspection target area (S9 in FIG. 11). Other procedures and operations are similar to those in the third embodiment, and the exclusion area can be determined more reliably than in the third and fourth embodiments.

In the fourth and fifth embodiments, it is possible to omit the determination regarding the area.

[0034]

According to the invention of claims 1 to 5, at least one of the high density side and the low density side with respect to the average density of the pixels included in the inspection candidate area including at least a partial area of the inspection object. Since a density threshold value deviated by a specified offset value is set to, and the exclusion candidate pixel is extracted from the inspection candidate area by binarizing the pixels in the inspection candidate area using the density threshold value. There is an advantage that the density threshold value used for binarization is automatically set according to each inspection object, and furthermore, the inspection candidate area is set so as to include at least a part of the inspection object, Since the density threshold value is set based on the average density of this inspection candidate area, it is possible to distinguish between the normal portion of the inspection object and the other portion even if the difference in density is relatively small. Value density threshold can be set It has the advantage that become.

According to the first aspect of the present invention, the connected components of the exclusion candidate pixels obtained by binarization are compared with known information to determine whether or not each of the connected components is an exclusion region to be excluded from the inspection target, and the inspection candidate Since the area excluding the exclusion area from the area is set as the inspection target area, it is possible to remove from the exclusion candidate pixel an area that is not the inspection target, such as a hole or protrusion of the inspection object or a conveyor that conveys the inspection object. Therefore, there is an advantage that the region that is not the inspection target can be sufficiently excluded as a result, which leads to the improvement of the processing efficiency.

According to the second aspect of the present invention, a connected component in which the area of the connected component of the exclusion candidate pixel obtained by binarization is larger than a prescribed area threshold value is determined as an exclusion area to be excluded from the inspection, and the inspection is performed. The inspection area is the area excluding the exclusion area from the candidate area, and defects such as scratches and stains are
Since the area is smaller than the hole or protrusion formed on the inspection object or the background of the inspection object, it is possible to exclude the hole, protrusion or background while leaving defects such as scratches and stains in the inspection target area. It has an advantage that it leads to improvement of processing efficiency. Further, there is an advantage that the area to be inspected can be easily determined because information about the inspected object having no defect is not required in advance.

According to a third aspect of the invention, in addition to the evaluation of the area according to the second aspect of the invention, each pixel on the edge of the connected component is orthogonal to the direction indicated by the differential direction value of the pixel with the pixel sandwiched therebetween. Connected components in which the density difference between a pair of pixels equidistant from each other in a given direction exceeds a specified value and the number of pixels exceeds a specified pixel number threshold is determined as an exclusion area to be excluded from the inspection target and excluded from the inspection candidate areas. Since the region excluding the region is the inspection target region, by using not only the area of the connected component of the exclusion candidate pixel but also the information about the density difference between the inside and outside of the connected component of the exclusion candidate pixel, the inspection target region There is an advantage that the exclusion area can be separated more reliably than the invention of claim 2.

According to a fourth aspect of the present invention, in addition to the evaluation of the area according to the second aspect of the present invention, a connected component in which the sum of the differential values of each pixel on the edge of the connected component exceeds a prescribed differential threshold is inspected. Since it is determined as an exclusion area to be excluded and the inspection candidate area is the area excluding the exclusion area, it is not only the area of the connected component of the exclusion candidate pixel, but also the edge of the connected component of the exclusion candidate pixel. By also using the information on the differential value in 1), there is an advantage that the inspection target area and the exclusion area can be separated more surely than the invention of claim 2.

According to the fifth aspect of the present invention, for the connected component of the exclusion candidate pixel, the area, the density difference at a position separated from the edge by a predetermined distance, and the density change near the edge are evaluated. As compared with the case where the evaluation is performed independently, the exclusion area and the inspection area are discriminated under stricter conditions, and there is an advantage that the inspection area and the exclusion area can be more surely separated.

[Brief description of drawings]

FIG. 1 is a flowchart showing a procedure of a first embodiment.

FIG. 2 is a diagram showing an object to be inspected used for explaining an example.

FIG. 3 is a diagram showing a relationship among an object to be inspected, an inspection candidate area, a connected component of exclusion candidate pixels, and the like in an example.

FIG. 4 is a diagram showing a state in which a connected component of exclusion candidate pixels is included in the embodiment.

FIG. 5 is a diagram showing a state in which exclusion areas are determined in the embodiment.

FIG. 6 is an operation explanatory diagram illustrating a relationship between a density threshold value and an exclusion candidate pixel according to the embodiment.

FIG. 7 is a flowchart showing the procedure of the second embodiment.

FIG. 8 is a flowchart showing the procedure of the third embodiment.

FIG. 9 is a diagram illustrating a relationship between a differential direction value and a pixel for which a density difference is obtained according to the third embodiment.

FIG. 10 is a flowchart showing the procedure of the fourth embodiment.

FIG. 11 is a flowchart showing the procedure of the fifth embodiment.

[Explanation of symbols]

D ₁ Inspection candidate area D ₂ Connected component D ₃ Inspection area Q ₁ Inspected object R ₁ Conveyor R ₂ Groove R ₀ Defect

Claims (5)

[Claims] 1. An inspection candidate area including at least a partial area of an object to be inspected is set in a grayscale image obtained by imaging a spatial area including the object to be inspected by an image pickup means, and the object to be inspected in the inspection candidate area is set. In a method of inspecting the appearance of an object, a density threshold value deviated by a specified offset value is set on at least one of a high density side and a low density side with respect to an average density of pixels included in an inspection candidate area. In addition, after extracting exclusion candidate pixels from the inspection candidate area by binarizing the pixels in the inspection candidate area using the density threshold value, the connected component that is the portion occupied by the area in which only the exclusion candidate pixels are connected Is compared with known information to determine whether each connected component is an exclusion area to be excluded from the inspection target, and then the inspection candidate area excluding the exclusion area is set as the inspection area, and only within the inspection area. Cover An appearance inspection method characterized by inspecting the appearance of an inspection object. 2. An inspection candidate area including at least a partial area of the inspection object is set in a grayscale image obtained by capturing an image of a spatial area including the inspection object by an image pickup means, and the inspection object in the inspection candidate area is set. In a method of inspecting the appearance of an object, a density threshold value deviated by a specified offset value is set on at least one of a high density side and a low density side with respect to an average density of pixels included in an inspection candidate area. In addition, after extracting exclusion candidate pixels from the inspection candidate area by binarizing the pixels in the inspection candidate area using the density threshold value, the connected component that is the portion occupied by the area in which only the exclusion candidate pixels are connected The connected component whose area is greater than the specified area threshold is determined as an exclusion area that is not to be inspected, and then the area excluding the exclusion area in the inspection candidate area is set as the inspection area and An appearance inspection method characterized by inspecting only the appearance of an object to be inspected. 3. An inspection candidate area including at least a partial area of the inspection object is set in a grayscale image obtained by imaging a spatial area including the inspection object by an image pickup means, and the inspection object in the inspection candidate area is set. In a method of inspecting the appearance of an object, a density threshold value deviated by a specified offset value is set on at least one of a high density side and a low density side with respect to an average density of pixels included in an inspection candidate area. In addition, after extracting exclusion candidate pixels from the inspection candidate area by binarizing the pixels in the inspection candidate area using the density threshold value, the connected component that is the portion occupied by the area in which only the exclusion candidate pixels are connected Is a connected component in which the number of pixels included in is larger than a specified area threshold value, and for each pixel on the edge of the connected component, a direction orthogonal to the direction indicated by the differential direction value of the pixel with the pixel sandwiched A connected component in which the density difference between a pair of pixels that are equidistant from each other exceeds a prescribed value exceeds a specified pixel number threshold is determined as an exclusion area that is not to be inspected, and next, among the inspection candidate areas, An appearance inspection method, wherein an area excluding an exclusion area is set as an inspection target area, and the appearance of the inspection object is inspected only in the inspection target area. 4. An inspection candidate area including at least a partial area of the inspection object is set in a grayscale image obtained by capturing an image of a spatial area including the inspection object by an image pickup device, and the inspection object in the inspection candidate area is set. In a method of inspecting the appearance of an object, a density threshold value deviated by a specified offset value is set on at least one of a high density side and a low density side with respect to an average density of pixels included in an inspection candidate area. In addition, after extracting exclusion candidate pixels from the inspection candidate area by binarizing the pixels in the inspection candidate area using the density threshold value, the connected component that is the portion occupied by the area in which only the exclusion candidate pixels are connected The connected component whose number of pixels contained in is larger than the specified area threshold and the sum of the differential values of each pixel on the edge of the connected component exceeds the specified differential threshold is excluded from the inspection target. become An appearance inspection method characterized by determining an area to be excluded as an inspection area, then selecting an area of the inspection candidate area excluding the exclusion area as an inspection area, and inspecting the appearance of the inspection object only in the inspection area. 5. An inspection candidate area including at least a partial area of the inspection object is set in a grayscale image obtained by imaging a spatial area including the inspection object by an image pickup means, and the inspection object in the inspection candidate area is set. In a method of inspecting the appearance of an object, a density threshold value deviated by a specified offset value is set on at least one of a high density side and a low density side with respect to an average density of pixels included in an inspection candidate area. In addition, after extracting exclusion candidate pixels from the inspection candidate area by binarizing the pixels in the inspection candidate area using the density threshold value, the connected component that is the portion occupied by the area in which only the exclusion candidate pixels are connected Is a connected component in which the number of pixels contained in is larger than the specified area threshold, and the sum of the differential values of each pixel on the edge of the connected component exceeds the specified differential threshold, and the connected component edge For each of the above pixels, the threshold value is the number of pixels for which the density difference between a pair of pixels that are equidistant in the direction orthogonal to the direction indicated by the differential direction value of that pixel exceeds the specified value. Connected components exceeding 10 are determined as exclusion areas that are not to be inspected, and then the inspection candidate area is the inspection area except the exclusion area, and the appearance of the inspected object is inspected only within the inspection area. Appearance inspection method characterized by.