JPH11132743A - Defect inspecting device by image signal processing - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a defect with high sensitivity within the dispersion range by detecting edge points of an object to be inspected, connecting the edge points to obtain the shape of the object to be inspected, detecting the abrupt direction change of the edge points, and judging the quality of the shape. SOLUTION: An image input section 2 receives an object as an image and stores image data in an image memory 3. An edge detection section 20 is provided at the next stage of the image memory 3, and it binalizes the image data and detects edges via image preprocessing for extracting black/white change points to obtain edge point images. An edge trace section 21 traces the edge point images to obtain a continued edge image as an outline image. A judgment section 22 obtains the x, y-coordinates of the edge points from the outline image, selects and specifies an edge point near the central object edge point, and calculates the angle with the object edge point from the coordinate values. A defect can be detected from the angle and the interval between the edge points, thereby the defect can be detected with high sensitivity.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect detection apparatus using image signal processing.

[0002]

2. Description of the Related Art In image processing, the shape (outer periphery) of an object
When inspecting, namely, defects, irregularities,
In inspecting a flaw or the like, the following two methods are exemplified. One of them is a method of generating an inspection pattern and analyzing data on the inspection pattern. As shown in FIG. 6A, an inner and outer inspection pattern P with respect to the outer shape O of the object is generated. The image data on the pattern P is compared and determined. FIG. 6B shows a block diagram for this purpose. The external shape of the target object 1 is captured as an image by the image input unit 2, and the obtained image data is stored in the image memory 3. On the other hand, the pattern data from the inspection pattern generator 4 is stored in the pattern memory 5. The comparator 6 compares the image data with the inspection pattern, determines a defect, an irregularity, or a flaw on the outer periphery by the determination unit 7, and outputs the result by the output unit 8.

Another method is a method for detecting a difference between an image of a standard product and an image of an inspection object. As shown in FIG. 7A, a difference D between an inspection object O and a standard product (template) P is obtained. Is used to make the determination. FIG. 7B shows a block diagram for this purpose. The external shape of the target object 1 is captured as an image by the image input unit 2, and the obtained image data is stored in the image memory 3. On the other hand, the standard template 9
Is stored in the pattern memory 5.
The comparator 6 compares the image data with the pattern data,
The difference image unit 10 obtains a difference image, the determination unit 7 determines loss, unevenness, and scratches from the difference image, and the output unit 8 outputs the result.

[0004]

In the defect inspection in the above two examples, when analyzing image data on the inspection pattern shown in FIG. 6, it is necessary that a correct outer shape is known. It is necessary that there is no variation in the outer shape of the object. However, it is difficult for the outer shape to have no variation. Therefore, an inspection is performed by setting a variation range in advance. However, it is difficult to set the variation range itself, and if this range is widened in advance, the defect detection sensitivity is lowered, and even if an accurate variation range is determined, if an inspection pattern is generated outside the range, it is difficult to detect a defect. Have a problem.

In addition, when taking the difference between the standard product and the inspection object shown in FIG. 7, the position of the object needs to be known,
In addition, it is necessary that there is no variation in the outer shape of the object, but in this example, since the difference in the image is taken, all the variations in the outer shape are detected. In this case, too, it is difficult to detect a defective portion, and a dead zone is provided. In such a case, the sensitivity decreases. Therefore, even if the variation range is set in advance in the same manner as described above and the difference is obtained, the defect detection sensitivity decreases as in the above-described example.

Originally, as shown in FIG. 8, in an object such as an industrial part, a variation exists as a tolerance, and it is desired to detect a defect or the like within the tolerance. However, it is difficult to detect as described above.

The present invention has been made in view of the above problems, and provides a defect detection apparatus by image signal processing capable of detecting a defect with high sensitivity even within a range of variation such as a tolerance and performing a desired defect detection.

[0008]

The present invention that achieves the above object has the following matters specifying the invention. (1) An edge detection unit for detecting an edge point of the object to be inspected, an edge trace unit for connecting the edge points to obtain the shape of the object to be inspected, and analyzing the edge points to detect a sharp change in direction of the edge And a determining unit for determining whether the shape of the edge is good or bad. (2) In the above (1), the determination unit detects an angle between the target edge point as a center and the edge point selected and specified on both sides thereof, and determines the angle between the angle and the selected edge point. The presence or absence of a defect is determined based on the interval.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Here, an example of an embodiment of the present invention will be described with reference to FIGS. FIG. 1 shows a block configuration of the present example. An edge detection unit 20 is provided next to an image input unit 2 that captures a target object as an image and an image memory 3 that stores image data of the image input unit 2.
The edge detection unit 20 detects edges by binarization and image preprocessing, and outputs a series of stippled edge points. Here, the edge detection unit 20
Is to binarize image data as shown in FIG. 2, perform image preprocessing to extract a black-and-white change point from the binarized image data, and detect an edge from the change point.
An edge point image will be obtained.

An edge tracing unit 21 is provided next to the edge detecting unit 20. This edge trace unit 21
Is to obtain a continuous edge image by tracing the edge point image obtained by the edge detection unit 20. As shown in FIG. 3, an outline image is obtained by connecting edge points continuously. In this case, as shown in FIG. 3, the tracing method is performed in the order of 0, 1, 2,.
..7, an edge shape is obtained by continuous coordinates by searching for an edge.

A determination unit at the next stage of the edge trace unit 21
In step 22, defect detection is performed.
Take the law. First, as shown in FIG.
The x and y coordinates of each edge point are measured from the outline image.
Next, this edge point is centered on the target edge point.
Select and specify an edge point close to
Calculate the angle with the edge point. For example, an edge point
P_TwoWhen is the target edge point, the irregularities to be detected
Edge point P selected by width of part₁P_ThreeIdentify this point
P₁P_ThreeAnd P _TwoCalculate the angle for. Sand
That is, as shown in FIG._ThreeAnd P_TwoAnd
Angle θ₁And edge point P₁And P_TwoAngle θ_Two,
Angle θ = θ₁+ Θ_TwoGet. Edge point P ₁Coordinate of (X
₁Y₁), P_TwoCoordinate of (X_TwoY_Two), P_ThreeThe coordinates of
(X_ThreeY_Three), Θ₁= Tan^-1(X_Three-X_Two
/ Y_Three-Y_Two), Θ_Two= Tan^-1(X₁-X_Two/ Y₁−
Y_Two), And by adding these, FIG.
Θ can be obtained. Here, the corner of the defect to be detected
Degree is less than θ and width is P_Three~ P₁To set the edge
Detecting defects by calculating θ at each point above
Can be done.

[0012] For example, FIG. 4 edge points as shown in (c) P ₁
In the P ₂ P _3, i.e. the gap is a 3 pixel depth for inspection of defects, such as two pixels, the angle of defects in P ₂ in FIG. Applying the above equation for θ, θ = tan ^{- 1} (2/2) + tan ^-1 (2/2) = 90 °, and a defect can be detected at the set angle of 90 °.

[0013]

As described above, according to the present invention, even if there is a variation in the shape of the target object, it is not necessary to compare the existing pattern with the existing pattern, so that defect detection can be performed with high sensitivity. The edge trace algorithm can be inspected by a common algorithm without being influenced by the size and shape of the object even when performing various contour inspections as shown in FIG. 5, for example, and detection of minute defects within tolerances Is also possible.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an example of an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing functions in an edge detection unit in FIG. 1;

FIG. 3 is an explanatory diagram showing functions of an edge trace unit in FIG. 1;

FIG. 4 is an explanatory diagram showing functions of a determination unit in FIG. 1;

FIG. 5 is an explanatory diagram for inspecting the shape of an irregular-shaped object.

FIG. 6 is an explanatory diagram of an example of a conventional example.

FIG. 7 is an explanatory diagram of another example of the conventional example.

FIG. 8 is an explanatory diagram of a tolerance.

[Explanation of symbols]

 Reference Signs List 20 edge detection unit 21 edge trace unit 22 judgment unit

Claims (2)

[Claims] 1. An edge detecting section for detecting an edge point of an object to be inspected, an edge trace section for connecting the edge points to obtain a shape of the object to be inspected, and a sharp direction change of an edge by analyzing the edge points And a determining unit for determining whether or not the shape composed of the edges is good or bad. 2. A determining unit detects an angle between an edge point selected and specified on both sides of a target edge point as a center, and determines a defect based on the angle and the interval between the specified edge points. 2. The defect inspection apparatus according to claim 1, wherein the presence or absence of the defect is determined.