JP2002039948A - Apparatus and method for inspecting appearance - Google Patents

Abstract

PROBLEM TO BE SOLVED: To inspect and decide a non-defectiveness or a defectiveness by imaging an appearance of an outer periphery of an article to be inspected and noted by a character, a symbol or the like on the outer peripheral surface. SOLUTION: A method for inspecting the appearance comprises the steps of rapidly rotating a cylindrical object 20 of a set object to be inspected by a motor 5, photographing the outer peripheral surface of the object 20 by a line sensor camera 1, detecting the head position of an area to be inspected of the object 20 by an image-processing unit 2 from the image data obtained by the camera 1, re-photographing the area to be inspected of the object 20 from the head position by the camera 1, and inspecting the appearance by the processing unit, based on the data assured at its resolution obtained by the camera 1.

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 for performing a visual inspection process by imaging a character string or the like partially printed on the surface of an inspection object with a line sensor.

[0002]

2. Description of the Related Art In mass production of industrial products, automation of various production methods and product inspection methods has been promoted in order to streamline production. Among these automations, visual inspections of products were mainly performed by visual inspection by humans. However, with the spread of electronic computers in recent years, television cameras (hereinafter, referred to as "TV cameras") have been widely used.
Automatic inspection has been performed by image processing in which an image captured by a "camera" is processed by an electronic computer.

In such a visual inspection apparatus using image processing, an image picked up by a camera is an analog electric signal, and after converting this analog signal into a digital signal, a discrimination inspection is performed by a computer.

As a method of discriminating inspection by an electronic computer, generally, an image of a standard product or a superior product (hereinafter, simply referred to as a “standard image”) previously taken into the electronic computer and an image of a product to be inspected (hereinafter referred to as “standard image”). (Hereinafter, simply referred to as “image to be inspected”), and if a point or part different from the standard image is present in the image to be inspected, the object to be inspected in the image to be inspected is determined to be defective, If there is a match, there is a method of determining a good product. According to the method using such a standard image, even if characters and symbols are written on the outer peripheral surface of the inspection object, the inspection can be performed in consideration of the characters and symbols.

In the method of comparing the standard image and the image to be inspected, the start point and the end point of the inspection image must always coincide with each other in accordance with the image range of the standard image. The outer diameter of the product to be inspected must be polygonal and the top, bottom, left and right must be constant.

Here, when an inspection is performed on the external appearance of a cylindrical object based on an image captured by a camera, it is necessary to capture an entire image of the outer periphery of the object. In the method of inspecting the appearance of a cylindrical object as described above, when imaging the outer periphery of the object to be inspected, the camera is fixed and the object to be inspected is rotated on a conveyor or the object is fixed and the object is fixed. There is a method of rotating the camera around the outer circumference of the camera. However, the method of rotating the camera requires a complicated mechanism for mounting and rotating the camera. Therefore, the camera is fixed and the object to be inspected is rotated. Is more preferred.

Conventionally, in order to recognize a character or pattern position having a shading difference on the surface of an object, a target image is fetched (FIG. 16), and a certain threshold value Lth is provided for the luminance level of the image (FIG. 17). Generally, a method of distinguishing by determining whether the brightness is higher (bright) or lower (dark) than the threshold Lth (FIG. 18).

[0008] This method is effective when the obtained image is stable. However, in the actually captured image, there is unevenness or unevenness in brightness over the entire image, and dust adheres. Erroneous recognition is easy. For example, FIG.
As shown in FIG. 20, when there is unevenness in the background,
As shown in the figure, misrecognition occurs at the spots of brightness,
Further, when dust adheres as shown in FIG. 21, as shown in FIG. 22, erroneous recognition occurs at a portion where the dust adheres.

Conventionally, there is also known a method of adding image differentiation (difference) or the like as preprocessing instead of evaluating the image data itself. This method is effective for an image having a spot of brightness on the background as shown in FIG. 19, but is effective for an image to which dust adheres as shown in FIG. 21 as shown in FIG. However, misrecognition also occurs. Also, it is difficult to determine whether the target is bright or dark based only on the differential value.

[0010]

By the way, in the method of taking in the entire image of the outer periphery of the cylindrical object and performing the appearance inspection, the starting point of taking the image to be inspected does not always coincide with the starting point of the standard image. Is difficult to use. In other words, in mass production of cylindrical objects, it is necessary to perform the appearance inspection while automatically inspecting the inspected object by a conveyor or the like, and it is not always necessary to make the image start positions of all the inspected objects the same during the appearance inspection. ,
This is difficult due to the structure of the transfer device, and solving this problem not only results in a complicated transfer structure, but also causes a reduction in the inspection processing speed.

Therefore, an object of the present invention is to provide a character
It is an object of the present invention to provide a visual inspection apparatus and a visual inspection method for an inspection object, in which a single camera is used to image the external appearance of an object to be inspected on which symbols and the like are written, so that a good / bad judgment inspection can be performed. .

[0012]

According to the present invention, there is provided an appearance inspection apparatus comprising: a line sensor camera for photographing a set inspection object; a relative driving means for relatively moving the inspection object and the line sensor camera; Image processing means to which image data obtained by the line sensor camera is supplied,
And control means for controlling these operations, the set inspection object and the line sensor camera are relatively moved at high speed by the relative driving means, and the outer peripheral surface of the inspection object is photographed by the line sensor camera, From the image data obtained by the line sensor camera, the image processing means detects the head position of the inspection target area of the inspection target object, and moves the inspection target object and the line sensor camera at the relative speed at a speed that ensures the required resolution. Relative movement by the driving means, the inspection object area of the inspection object is again photographed from the top position by the line sensor camera, and the image processing means based on the image data obtained by the line sensor camera and ensuring the resolution. It is characterized by performing a visual inspection process.

Further, in the appearance inspection method according to the present invention, the inspection object is set, the set inspection object and the line sensor camera are relatively moved at a high speed, and the outer peripheral surface of the inspection object is moved to the line sensor camera. The head position of the inspection target area of the inspection object is detected from the image data obtained by the line sensor camera, and the inspection object and the line sensor camera are relatively moved at a speed that ensures a necessary resolution. The inspection object area of the inspection object is photographed again from the top position by the line sensor camera, and the appearance inspection processing is performed based on the image data obtained by the line sensor camera and having the ensured resolution.

[0014]

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

The present invention is applied to, for example, a visual inspection apparatus 10 having a configuration as shown in FIG. This visual inspection device 10
Is a line sensor camera 1 for photographing the outer peripheral surface of a cylindrical object 20 to be inspected, an image processing unit 2 to which an image signal photographed by the line sensor camera 1 is supplied, and a process performed by the image processing unit 2. A central processing unit (CPU) 3 to which the obtained image data is supplied, a motor controller 4 connected to the CPU 3, and a motor 5 for rotating the cylindrical object 20 as the inspection object.

In the appearance inspection apparatus 10, a cylindrical object 20 to be inspected is supplied and set by a supply / discharge unit (not shown), and the inspected cylindrical object 20 is processed.
Is discharged by the supply / discharge unit.

Here, the outer peripheral surface 21 of the cylindrical object 20 to be inspected is shown in FIG.
As shown in a developed view of the outer peripheral surface in FIG.
It is composed of a character portion 21A on which a character "ABCDEF" is printed and a gap portion 21B where no character is printed. In the cylindrical object 20 supplied and set by the supply / discharge unit, the character portion 21A is randomly positioned with respect to the line sensor camera 3 every time.

The CPU 3 in the visual inspection device 10
As shown in the flowchart of FIG. 3, when the cylindrical object 20 supplied by the supply / discharge unit is set (step S1), the outer peripheral surface 21 of the cylindrical object 20 is set.
An image search process (step S2) for detecting the head position of the character portion 21A is performed, an image for character recognition is fetched from the head position of the character portion 21A detected by the image search process, and the inspection process of the character portion 21A is performed. Execute (Step S3), and supply the inspected cylindrical object 20
It is discharged by the discharge unit (step S4).

In the image search process in step S2,
The motor controller 4 is controlled so that the cylindrical object 20 is rotated at a high speed by the motor 5, and the entire outer peripheral surface 21 of the cylindrical object 20 is imaged in a short time by the line sensor camera 1. In this image search process, since the cylindrical object 20 is rotated at a high speed and the entire outer peripheral surface 21 is imaged in a short time, a dense image signal is not obtained so that the character shape of the character portion 21A can be understood. The head position of the character portion 21A is detected from the image signal by the processing according to the procedure shown in the flowchart.

In the image search processing, the image picked up by the line sensor camera 1 is taken in at a high speed, so that the shape of the character is not clear and the character is slightly blurred as shown in FIG. 5, for example. It becomes a patchy mass. If this image was evaluated at a fixed threshold, characters,
It is difficult to distinguish uneven density, dust and the like.

Therefore, in the image search processing in the visual inspection apparatus 10, first, as shown in FIG. 6, a gray level variation is obtained for each line in the image (step S21). The variation referred to here is obtained by calculating the difference between each pixel data from the total data average in the line and averaging the difference by squaring them.

Next, weighting is performed by multiplying each pixel data in the line by the variation value obtained in step S21 (step S22).

It is determined whether or not the weighting process in step S22 has been performed for all the lines in the image (step S23). If there is a line to be weighted, the process returns to step S22 and returns to step S22. And step S23 are repeated to execute the weighting process on all the lines in the image.

In the image in which the weighting process has been executed for all the lines, as shown in FIG. 7, the area with characters is emphasized by the weighting process. Executing the weighting process on all the lines in this manner is more effective than the method of evaluating only the variation value in the sense that the extraction target is limited to the white (large gray data value) portion. Only areas having density variations in a large and narrow range will be highlighted.

Next, in each line, an average value of gray data weighted in the area is calculated as one area of about 3 lines on the left and right (step S2).
4). The obtained result is represented as the center value. That is, as shown in FIG. 8, for example, assuming that a part of the image is enlarged, when considering the case of six lines a to f in the image, the representative value of each line obtained as an average value of three lines is: b ′ = (a + b + c) / 3 c ′ = (b + c + d) / 3 d ′ = (c + d + e) / 3 e ′ = (d + e + f) / 3

It is determined whether or not the average value calculation processing in step S24 has been executed for all the lines in the image (step S25). If there is a line for which the average value calculation processing is to be performed, the processing proceeds to step S24. Return and step S24
And step S25 are repeated to execute the average value calculation processing for all the lines in the image.

The head position of the character portion 21A is detected based on the representative value of each line as a result of executing the average value calculation processing for all the lines in the image (step S26).

In step S26, the presence / absence of characters is determined for a representative value of each line at a certain threshold value.
If three or more lines for which the character class is determined to be “continuous” are regarded as a character block, that part is regarded as a character block. In addition, when one “absence” line is inserted while a line whose character class is determined to be “presence” continues, that line is also regarded as “presence” and character blocks are assumed to be continuous. By executing the above processing, the character block in the image can be grasped, and thus the head position of the character block is recognized as the head position of the character part 21A.

Here, in the image search processing for detecting the leading position of the character portion 21A on the outer peripheral surface 21 of the cylindrical object 20, the outer peripheral surface 21 of the cylindrical object 20 is imaged by the line sensor camera 1 and taken in at a high speed. Examples of the results of calculating a) average value, b) line MAX, c) MAX average of eight lines, d) variation of each line, and e) line MAX × variation of each line for the obtained image data. 9 to 14.

The visual inspection apparatus 10 performs an image search process for detecting the head position of the character portion 21A on the outer peripheral surface 21 of the cylindrical object 20 in a short time, and the head position of the character portion 21A detected by the image search process. Then, an image for character recognition is fetched from the device, and the character portion 21A is inspected.
The image for character recognition can capture only the necessary range, that is, only the character portion 21A, at a slow speed at which the character shape can be captured finely. This allows you to scan the whole area more slowly from the start,
The external processing of the cylindrical object 20 can be efficiently performed by shortening the total processing time.

It should be noted that the present invention is not applied only to the appearance inspection of the cylindrical object 20, but can be applied to a general plane image. That is, for example, as shown in FIG. 15 (A), when a pattern lump in a plane having gentle density unevenness or illumination unevenness partially exists in a small range, as shown in FIG. 15 (B). By dividing the plane image into a plurality of areas A to F and performing the above-described processing in each of the areas A to F,
The position of the cluster can be efficiently detected.

[0032]

As described above, according to the present invention, the inspection object is set, the set inspection object and the line sensor camera are relatively moved at a high speed, and the outer peripheral surface of the inspection object is moved along the line. The head position of the inspection target area of the inspection object is detected from the image data obtained by the sensor camera based on the image data obtained by the line sensor camera, and the inspection object and the line sensor camera are relatively moved at a speed that ensures the required resolution. By moving, the inspection target area of the inspection target is again photographed from the head position by the line sensor camera, and by performing the appearance inspection processing based on the image data that has secured the resolution obtained by the line sensor camera,
Compared to the method of scanning the whole at a low speed from the beginning, the total processing time can be shortened, and the appearance inspection of the inspection target area of the inspection target can be efficiently performed.

Therefore, according to the present invention, the object to be inspected is such that a single camera is used to image the external appearance of the object to be inspected, on which characters, symbols, etc. are written on the outer surface, so that a good / bad judgment inspection can be performed. An object appearance inspection apparatus and an appearance inspection method can be provided.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a visual inspection apparatus to which the present invention is applied.

2A and 2B are an external view and a development view of an outer peripheral surface of a cylindrical object to be inspected by the visual inspection apparatus.

FIG. 3 is a flowchart showing an inspection processing procedure of the appearance inspection apparatus.

FIG. 4 is a flowchart showing a procedure of an image search process executed by a CPU of the visual inspection apparatus.

FIG. 5 is a diagram showing an image captured by a line sensor camera in the image search processing.

FIG. 6 is a diagram showing an example of gray level variation for each line in an image obtained in the image search processing.

FIG. 7 is a diagram showing an image in which weighting processing has been performed on all lines in the image search processing.

FIG. 8 is a diagram illustrating an example of a calculation result of calculating an average value of weighted gray data.

FIG. 9 shows an average value, a line MAX value, a MAX average value of eight lines, a variation of each line, and image data obtained by capturing an outer peripheral surface of a cylindrical object with a line sensor camera at a high speed in the image search processing. In addition, 8
FIG. 9 is a diagram showing a comparison between the results of calculating the line MAX × variation of each line.

FIG. 10 is a diagram showing an average value of the image data.

FIG. 11 is a diagram showing a line MAX value of the image data.

FIG. 12 is a diagram showing a MAX average value of eight lines of the image data.

FIG. 13 is a diagram showing a variation of each line of the image data.

FIG. 14 shows a line MAX × 8 lines of the image data.
It is a figure showing the value of the variation of each line.

FIG. 15 is a diagram illustrating an example of planar image processing.

FIG. 16 is a diagram illustrating an example of an image captured in conventional image processing.

FIG. 17 is a diagram showing a luminance level of the image.

FIG. 18 is a diagram showing evaluation based on threshold values in conventional image processing.

FIG. 19 is a diagram showing an example of an image having a brightness spot on the background.

FIG. 20 is a diagram illustrating erroneous recognition in an image having a brightness unevenness in the background.

FIG. 21 is a diagram illustrating an example of an image to which dust adheres.

FIG. 22 is a diagram illustrating an example of erroneous recognition in an image to which dust has adhered.

FIG. 23 is a diagram illustrating another example of erroneous recognition in an image to which dust has adhered.

[Explanation of symbols]

1 line sensor camera, 2 image processing unit, 3
CPU, 4 motor controller, 5 motor, 10
Visual inspection device, 20 cylindrical object

Continued on the front page F-term (reference) 2F065 AA49 BB06 BB15 BB27 FF04 JJ02 JJ09 JJ25 MM04 PP15 QQ31 RR06 UU05 2G051 AA90 AB01 AB07 AB11 CA03 CA04 CB01 DA08 EA12 EA14 EA26 EC03 ED01 5B007 A0203 CA14 CA16 EA14 EA39 FA32 FA81 HA07

Claims (4)

[Claims] 1. A line sensor camera for photographing a set inspection object, relative driving means for relatively moving the inspection object and the line sensor camera, and an image supplied with image data obtained by the line sensor camera Processing means; and control means for controlling these operations. The set inspection object and the line sensor camera are relatively moved at high speed by the relative driving means, and the outer peripheral surface of the inspection object is moved by the line sensor camera. The head position of the inspection target area of the inspection object is detected by the image processing means from the image data obtained by the line sensor camera, and the required resolution between the inspection object and the line sensor camera is secured. Relative movement by the relative driving means at a speed, and moves the inspection area of the inspection object from the top position to the line Again taken by Nsakamera, appearance inspection apparatus and performs a visual inspection process by the image processing means based on image data obtained by securing a resolution obtained by the line sensor camera. 2. The image processing means according to claim 1, wherein in the head position detection processing for detecting the head position of the inspection target area of the inspection object from the image data obtained by the line sensor camera, each line in the image is grayed out. A level variation is obtained, a weighting process of multiplying the obtained variation value by each pixel data in the line is performed on all the lines in the image, and the weighted gray data in the image is centered on each line. The average value of the plurality of lines is calculated for all the lines, and based on the representative value of each line as a result of executing the average value calculation process on all the lines in the image, the inspection target area of the inspection target is 2. The visual inspection device according to claim 1, wherein the head position is detected. 3. The inspection object is set, the set inspection object and the line sensor camera are relatively moved at a high speed, and the outer peripheral surface of the inspection object is photographed by the line sensor camera. The head position of the inspection target area of the inspection target is detected from the obtained image data, and the inspection target and the line sensor camera are relatively moved at a speed that ensures a required resolution. Is again photographed from the head position by the line sensor camera, and a visual inspection process is performed based on the image data obtained by the line sensor camera and having the ensured resolution. 4. A head position detection process for detecting a head position of an inspection target area of the inspection object from image data obtained by the line sensor camera, wherein a gray level variation is obtained for each line in the image. A weighting process of multiplying the obtained variation value by each pixel data in the line is performed on all the lines in the image, and for the weighted gray data in the image, an average value of a plurality of lines around each line is calculated. Calculating for all lines and detecting the top position of the inspection target area of the inspection target based on the representative value of each line as a result of executing the average value calculation processing for all the lines in the image The appearance inspection method according to claim 3, wherein: