WO2015115499A1

WO2015115499A1 - Tool inspection method and tool inspection device

Info

Publication number: WO2015115499A1
Application number: PCT/JP2015/052396
Authority: WO
Inventors: 石崎　義公; 優宮本; 智仁服部; 延偉陳; 康介宮脇
Original assignee: 株式会社タカコ
Priority date: 2014-02-03
Filing date: 2015-01-28
Publication date: 2015-08-06
Also published as: TW201543001A; JP2015145819A; JP6293505B2

Abstract

This tool inspection method is provided with a grayscale image generation step for imaging a tool under inspection and generating a grayscale image; a topographical image generation step for processing the grayscale image using a first filter and generating a topographical image indicating a brightness gradient; a marker image generation step for processing the grayscale image using a second filter, extracting a characteristic portion within a wear area, setting a marker for the extracted portion, and generating a marker image; a wear area extraction step for using the topographical image and marker image to extract a wear area through maker-based watershed area segmentation; and a determination step for determining the state of the tool under inspection on the basis of the surface area of the wear area extracted in the wear area extraction step.

Description

Tool inspection method and tool inspection apparatus

The present invention relates to a tool inspection method and a tool inspection apparatus.

JP 1997-21080A discloses a method for detecting a shape defect of a tool by scanning a line entering a minute amount from the outer peripheral shape drawn by a ridge line with a laser focus displacement meter and detecting irregularities on the scanning line. Yes.

In JP1998-96616A, as a method of inspecting a chip of a tool tip, a tool of a milling machine is transferred by a traveling robot and fixed to a jig on a chip inspection / exchange work table, and a slit is projected from a projection window of a structure light unit. It is disclosed that light is projected onto a tool tip, a camera shot of the tool tip is taken and analyzed by an image processing device, and index data representing the size of the tool tip defect is obtained to determine the necessity for replacement. . Regarding the necessity of tool tip replacement, the Y-axis direction and Z-axis direction spread of the recess formed by wear progressing from the edge portion of the tool tip is recognized as index data, and the index data is a preset allowable value. It is determined that there is a need for replacement when exceeding

The method described in JP1997-21830A is for inspecting a tool for defects by scanning with a laser beam, so that the apparatus becomes large and the time required for the inspection becomes longer.

The method described in JP1998-96616A is a method that pays attention to a change in the wear region and is a method that indirectly measures the wear region, and it is difficult to say that the state of the tool can be inspected with high accuracy.

An object of the present invention is to provide a tool inspection method and a tool inspection apparatus capable of inspecting a tool with a simple method with high accuracy.

According to an aspect of the present invention, there is provided a tool inspection method for imaging a target tool to be inspected, generating a grayscale image, and processing the grayscale image with a first filter. A terrain image generation step for generating a terrain image showing a luminance gradient, and processing the grayscale image with a second filter to extract a feature portion in the wear region, and a marker is placed on the extracted feature portion A marker image generation step for setting and generating a marker image, a wear region extraction step for extracting a wear region by region division by a Maker Based Watershed method using the topographic image and the marker image, and the wear region extraction step A determination step of determining a state of the target tool based on the extracted area of the wear region.

According to another aspect of the present invention, there is provided a tool inspection apparatus that captures a target tool to be inspected and generates a grayscale image, and a grayscale image generated by the first filter. A terrain image generation unit for processing to generate a terrain image showing a luminance gradient; and processing the grayscale image with a second filter to extract a feature portion in a wear region; and marking the extracted feature portion with a marker A marker image generation unit that generates a marker image by setting the wear region extraction unit that extracts a wear region by region division using a Maker Based Watershed method using the topographic image and the marker image, and the wear region extraction step A determination unit that determines a state of the target tool based on the extracted area of the wear region.

FIG. 1 is a schematic configuration diagram of a tool inspection apparatus according to a first embodiment of the present invention. FIG. 2 is a flowchart showing the procedure of the tool inspection method according to the first embodiment of the present invention. FIG. 3 shows an inspection target image. FIG. 4 shows a grayscale image. FIG. 5 is a diagram for explaining the Watershed method. FIG. 6 shows a topographic image. FIG. 7 shows a marker image. FIG. 8A shows an image obtained by extracting a wear region. FIG. 8B is a partially enlarged view of FIG. 8A. FIG. 9 is a graph showing a change between the automatically extracted wear area and the manually extracted wear area. FIG. 10 shows a topographic image in the second embodiment of the present invention.

<First Embodiment>
First, a first embodiment of the present invention will be described.

The tool inspection apparatus 100 according to the first embodiment of the present invention is an apparatus that inspects the state of a tool. In this embodiment, a case where the inspection target is a throw-away tip of various tools will be described.

As shown in FIG. 1, the tool inspection apparatus 100 includes a camera 1 as an image acquisition unit that acquires an inspection target image by photographing a cutting edge of a throw-away tip mounted on a processing apparatus, and an image acquired by the camera 1. And a computer 2 that performs image processing of data and determines the state of the throw-away chip. The camera 1 is attached to a processing apparatus. The computer 2 may be provided adjacent to the processing device, or may be provided at a location away from the processing device.

The computer 2 includes a display 21 as a display unit capable of displaying image data, and a keyboard 22 and a mouse 23 as input units capable of inputting an instruction from a user.

Next, the method for inspecting the throw-away chip will be described in detail with reference to FIG.

In step 11, the cutting edge of the throw-away tip mounted on the processing apparatus is photographed using the camera 1, and an inspection target image (FIG. 3) is acquired (image acquisition process). In the inspection target image shown in FIG. 3, a wear portion is displayed at the lower left of the image. The inspection target image is acquired as a color image.

In step 12, a gray scale image (FIG. 4) is generated from the inspection target image acquired in step 11 (gray scale image generation step). Instead of generating a grayscale image from the color inspection target image acquired in step 11, the grayscale image may be generated directly by the camera 11.

In step 13, as shown in FIG. 4, the gray scale image acquired in step 12 is positioned. Specifically, the upper linear portion 41 of the edge contour is detected, and the gray scale image is positioned using affine transformation so that the linear portion 41 is horizontal. In this way, all the inspection target images acquired in step 11 are positioned at the same angle.

In Steps 14 to 18, the wear area of the blade tip of the throw-away tip is extracted using the Watershed method. The watershed method is a method for dividing an image using a density gradient. Region division is performed using the region growth method with the obtained minimum value of the absolute value of the gradient as a seed. As a result, since the region grows from a portion with a small gradient to a portion with a large gradient, a boundary of the region is generated at the edge portion. More specifically, with reference to FIG. 5, the luminance gradient of the pixel of the image is regarded as a topographic structure, and the minimum value and the maximum value of the luminance value are regarded as a valley and a ridge, respectively. First, the minimum value of the luminance value (

minimum values

1 and 2 in FIG. 5) is searched, and water is poured into the valleys of the

minimum values

1 and 2 to become maximum values (

maximum values

1, 2, and 3 in FIG. 5). Water until water is found. When a local maximum value sandwiching the local minimum value is found, the region between the local maximum values sandwiching the local minimum value becomes one region, which is divided into two regions (regions 1 and 2) in FIG.

The watershed method has a problem that the image is divided into the number of seeds. Therefore, in the present embodiment, region division is performed using a Maker Based Watershed method that solves this problem by providing a seed marker in advance. In the Maker Based Watershed method, it is necessary to manually provide a marker, but in this embodiment, a marker is automatically generated using a Gabor filter to be described later, and the wear area of the tip of the throw-away tip is extracted automatically. . This will be described in detail below.

In step 14, the grayscale image acquired in step 12 is processed by the Gabor filter as the first filter and then the luminance gradient is calculated to generate a terrain image (FIG. 6) indicating the luminance gradient (terrain image generation step). ). The topographic image indicates the brightness gradient of an image that can be regarded as a topographic structure.

The Gabor filter is defined by the product of a Gaussian function and a sine wave, and emphasizes those that match the filter characteristics and smoothes those that do not match. The filter characteristics are set by adjusting the width of the Gaussian function and the phase and amplitude of the sine wave.

As shown in FIG. 4, in the gray scale image, the worn portion is displayed white (high luminance value), and the non-weared portion is displayed black (low luminance value). Therefore, in step 14, the edge portion of the grayscale image is emphasized using a Gabor filter to enhance the white portion. Thereby, as shown in FIG. 6, the boundary of the wear region is displayed in white in the topographic image. The portion displayed in white in the topographic image is regarded as a ridge when the Maker Based Watershed method is applied, so by generating a topographic image that emphasizes the white part of the grayscale image using the Gabor filter, Maker Based The wear area extraction accuracy is improved when the watershed method is applied.

In addition, since the Gabor filter emphasizes the filter that matches the filter characteristics and smoothes the filter that does not match, the Gabor filter emphasizes the white portion corresponding to the wear area and wears small scratches and the like. A portion (noise) that is not related to the noise can be smoothed and removed.

In step 15, the grayscale image acquired in step 12 is processed with a Gabor filter as a second filter to emphasize features in the wear region, and further binarized to extract features in the wear region. To do. And the marker 71 is set to the extracted feature part and a marker image (FIG. 7) is produced | generated (marker image production | generation process). The features in the wear region are extracted using a plurality of Gabor filters having different characteristics. Specifically, a portion corresponding to the maximum value is extracted from the filter values calculated by the filter processing using a plurality of Gabor filters. The characteristic portion extracted by the Gabor filter is an edge portion or a region having a high luminance value having a luminance value larger than a certain value.

Since the portion corresponding to the maximum value is extracted from the filter values calculated by the filter processing using a plurality of Gabor filters, the feature region in the wear region is also set as a marker. there is a possibility.

Therefore, in step 16, a restriction process for restricting the marker is performed (marker restriction process). Specifically, weighting is performed according to the distance from the cutting edge contour of the throw-away tip to the marker, and the marker is limited based on the weighting. That is, the marker set in a portion far from the edge contour is deleted.

In step 17, as shown in FIG. 7, the area restriction marker 72 is set so as to surround a certain area around the marker set in the marker image (restriction marker setting step). In step 16, the marker is limited according to the distance from the edge contour to the marker, and by setting the region limiting marker 72 in step 17, the wear region is extracted by region division by the Maker による Based Watershed method in the next step. The extraction accuracy when doing so can be improved.

In step 18, the wear area is extracted by area division by the Maker Based Watershed method using the topographic image (FIG. 6) and the marker image (FIG. 7) (wear area extraction step). Specifically, the region is expanded by applying the Maker マーカ Based Watershed method to the topographic image using the

markers

71 and 72 set in the marker image, and the wear region is extracted.

FIG. 8A shows an image obtained by extracting the wear region. FIG. 8B is a partially enlarged view of FIG. 8A. 8A and 8B, the boundary of the extraction region is indicated by reference numeral 81. In FIG. 8B, the marker 71 is indicated by hatching. Note that FIG. 8A shows an image whose orientation has been restored to the same angle as the inspection target image shown in FIG. As can be seen from FIG. 8B, the wear region extracted by the Maker８１Based Watershed method (the region surrounded by the boundary indicated by reference numeral 81) is almost coincident with the wear portion displayed in white, and the wear region is accurately extracted. I was able to confirm that it was possible.

In step 19, the area of the wear region extracted in step 18 is calculated, and the state of the tip of the throw-away tip is determined based on the area (determination step). For example, the state of the blade tip of the throw-away tip is determined by the following method. When the area of the wear region extracted in step 18 reaches a predetermined area, the throw-away tip is determined to have a lifetime. Further, deterioration levels 1 to 5 are set in advance according to the size of the area of the wear region, and when the deterioration level 5 is reached, the throw-away tip is determined to have a lifetime. In addition, the correlation between the area of the wear area and the number of times of use of the throw-away tip is set in advance, and the number of times of use is estimated according to the area of the wear area extracted in step 18 to determine the state of the cutting edge of the throw-away tip. Determine. In addition, when the area of the wear region extracted in step 18 changes abruptly at a predetermined rate or more, it is determined that the throw-away tip is abnormal.

The processes in steps 12 to 19 described above are automatically executed by software stored in the computer 2. As a result, if it is determined that the throw-away tip has reached the end of its life, a notification that prompts replacement is issued.

Next, examples will be described.

Each time it was processed, the side surface of the tip of the throw-away tip was imaged to obtain the inspection target image. The image size of the inspection target image is 1024 × 960 pixels. The obtained image to be inspected is subjected to image processing in steps 12 to 18 in FIG. 2 to automatically extract the wear region, and the area of the extracted wear region is calculated. Further, as a comparative example, every time machining was performed, the side surface of the blade tip of the throw-away tip was imaged, the wear region was manually extracted, and the area of the extracted wear region was calculated.

Fig. 9 shows the change between the automatically extracted wear area and the manually extracted wear area. In FIG. 9, the area of the automatically extracted wear region is indicated by a solid line, and the area of the manually extracted wear region is indicated by a broken line. As can be seen from FIG. 9, the change in the area of the automatically extracted wear region generally coincided with the change in the area of the manually extracted wear region. From this, it was confirmed that the wear region can be extracted with high accuracy by using the image processing in steps 12 to 18. Therefore, it can be said that the state of the blade tip of the throw-away tip can be accurately determined by performing image processing in steps 12 to 18 to automatically extract the wear region and evaluating the area of the extracted wear region.

According to the above 1st Embodiment, there exists the effect shown below.

In the present embodiment, since the state of the target tool is automatically determined based on the area of the wear area extracted by area division by the Maker Based Watershed method using the topographic image (FIG. 6) and the marker image (FIG. 7), The tool can be inspected with high accuracy by a simple method.

In this embodiment, a Gabor filter is used to emphasize features in the wear region of the grayscale image, and markers used in the Maker Based Watershed method are set there. In this way, markers can be set automatically instead of manually.

In addition, the marker is limited according to the distance from the edge contour to the marker, and the area restriction marker 72 is set so as to surround a certain area around the marker, so the wear area is extracted by dividing the area by the Maker Based Watershed method. The extraction accuracy when doing so can be improved.

Second Embodiment
Next, a second embodiment of the present invention will be described. Hereinafter, only differences from the first embodiment will be described.

In the second embodiment, the topographic image obtained in step 14 in FIG. 2 is different from the topographic image obtained in the first embodiment.

A topographic image (FIG. 10) showing a luminance gradient is obtained by black-and-white inversion processing of a grayscale image (FIG. 4) with a Gabor filter.

In the topographic image (FIG. 10) subjected to the black-and-white reversal process, the worn portion is displayed in black (low luminance value) and the non-weared portion is displayed in white (high luminance value). The marker of the marker image (FIG. 7) is set corresponding to a low luminance value in the terrain image. Therefore, when the region is divided by applying the Watershed method, the region grows from the valley (low luminance value) to the ridge (high luminance value), so that the wear region extraction accuracy is improved.

Also in the second embodiment described above, the same operational effects as in the first embodiment are obtained.

The embodiment of the present invention has been described above. However, the above embodiment only shows a part of application examples of the present invention, and the technical scope of the present invention is limited to the specific configuration of the above embodiment. Absent.

This application claims priority based on Japanese Patent Application No. 2014-018512 filed with the Japan Patent Office on February 3, 2014, the entire contents of which are incorporated herein by reference.

Claims

A tool inspection method,
A gray scale image generation step of imaging a target tool to be inspected and generating a gray scale image;
A terrain image generation step of processing the grayscale image with a first filter to generate a terrain image indicating a luminance gradient;
A marker image generation step of processing the grayscale image with a second filter to extract a feature portion in a wear region, setting a marker on the extracted feature portion, and generating a marker image;
Using the terrain image and the marker image, a wear region extraction step of extracting a wear region by region division by the Maker Based Watershed method,
A determination step of determining the state of the target tool based on the area of the wear region extracted in the wear region extraction step;
A tool inspection method comprising:
The tool inspection method according to claim 1,
The tool inspection method, wherein the first filter and the second filter are Gabor filters.
The tool inspection method according to claim 1,
The tool inspection according to claim 1, further comprising a marker limiting step of performing weighting according to a distance from a cutting edge contour of the target tool to the marker, and limiting the marker based on the weighting. Method.
The tool inspection method according to claim 3,
A tool inspection method further comprising a restriction marker setting step of setting an area restriction marker so as to surround the marker set in the marker image.
The tool inspection method according to claim 1,
The terrain image generated in the terrain image generation step is a tool inspection method obtained by performing black-and-white reversal processing on the grayscale image with the first filter.
A tool inspection device,
A grayscale image generation unit that images a target tool to be inspected and generates a grayscale image;
A terrain image generation unit for processing the grayscale image with a first filter to generate a terrain image indicating a luminance gradient;
A marker image generating unit that processes the grayscale image with a second filter to extract a feature portion in a wear region, sets a marker in the extracted feature portion, and generates a marker image;
A wear region extraction unit that extracts a wear region by region division by the Maker Based Watershed method using the terrain image and the marker image;
A determination unit that determines the state of the target tool based on the area of the wear region extracted in the wear region extraction step;
A tool inspection device comprising: