JPH11351835A - Method for extracting contour line of cutting edge by processing taken image of tool - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for extracting the contour line of a cutting edge of a tool to obtain the contour line of the cutting edge, using multilevel picture data of a cutting edge photographed image for measuring the cutting edge wear degree of a tool. SOLUTION: Three or more line pixel pickup lines K11 -K14 intersecting in the cutting edge direction are extracted according to the cutting edge's orientation on a cutting edge multilevel picture of a tool, line pixel data on each line pixel pickup line are differentiated to obtain boundary candidate points between the tool cutting edge and background, a combination of points which forms a straight line most closely is selected among combinations of boundary candidate points as a guide straight line LG for the cutting edge, a differentiated picture according to the cutting edge on the cutting edge multilevel picture is obtained, and pixels having higher differentiated values are taken out along the guide straight line to obtain the cutting edge contour line of the object.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for extracting a cutting edge contour line by processing a captured image of a tool. More specifically, in a measuring device that images a blade edge portion of a tool and detects a degree of wear of the blade edge portion from the captured image, an imaging that aims to expand the applicable range by eliminating the influence of illumination unevenness and a background portion for imaging. The present invention relates to an improvement in a method for extracting a cutting edge contour line by image processing.

[0002]

2. Description of the Related Art Generally, when measuring the degree of wear of a cutting edge using an image of the cutting edge portion of a tool, an image of the cutting edge is extracted once after binarizing the image data of the cutting edge. We are taking a method.

[0003] The binarization processing of the image data of the cutting edge portion is performed by converting the captured image of the cutting edge into a white portion (corresponding to the tool edge portion, luminance value 255) and a black portion (corresponding to the background portion, luminance value 0). By clearly dividing the image signal, unnecessary noise components are removed from the image signal, and the image recognition of the tool edge and the extraction of the edge contour are performed effectively and well.

[0004] As another measurement method, a method of obtaining a cutting edge contour line by pattern matching between the picked-up images of the cutting edge can be considered. However, in this method, it is necessary to previously register the specified basic tool edge image pattern as the comparison target, and furthermore, the shape and position of the tool itself as the measurement target to be imaged are registered with the registered image. If they are different, the two may not be able to be compared with each other, so that the applicable range is limited and is not always effective.

[0005]

By the way, when the above-mentioned image data of the cutting edge is binarized to recognize the contour of the cutting edge, a white portion and a black portion in the picked-up image and, consequently, a tool cutting edge. The determination of the threshold value of the luminance for clearly distinguishing the relevant part on the part side from the relevant part on the background part becomes a problem.

That is, when there is such a large difference as to be able to be distinguished between the luminance values in the respective portions,
There is no problem in recognizing it, but in the normal case, including the uneven illumination for imaging for each corresponding part of the cutting edge and the background, and the color of each corresponding part, etc., depending on the imaging position Also, since the difference in luminance between the two changes, the threshold value that is always appropriate cannot be set effectively. As a result, the contour of the cutting edge of the tool to be extracted cannot be used for measuring the degree of cutting edge wear.

Conventionally, in order to avoid the above disadvantages, the lighting direction and the luminance have to be considered, and various measures corresponding to the background color have been made. However, such an improvement method increases the cost of the apparatus itself. In spite of this, it is not possible to clearly recognize even a straight line portion specific to the cutting edge of the tool, and in the conventional image processing means using such binarization processing, for example, as shown in FIG. 2 inappropriate for degree measurement
Often a value image.

The conventional problem described above is caused by the fact that the image data of the cutting edge of the tool is simply binarized to extract the contour of the cutting edge. In order to improve this problem, it is considered that a multi-valued image of the cutting edge portion captured image may be used as one means. In this case, the shape recognition (presentation) of the corresponding tool cutting edge portion is also considered. It is not always easy to determine which part of the image corresponds to the blade edge, etc.
There remains a problem that the processing time is relatively long.

In view of the above-mentioned conventional disadvantages, an object of the present invention is to provide
Using the multi-valued image data of the image of the cutting edge, processing it effectively and within the shortest possible time so that the desired cutting edge contour can be obtained. It is to provide a line extraction method.

[0010]

In order to achieve the above-mentioned object, a method for extracting a contour line of a cutting edge by picked-up image processing according to the present invention comprises extracting a plurality of line pixels corresponding to the direction of the cutting edge from a multivalued image of the cutting edge. Extract a line, and obtain a boundary candidate point between the cutting edge portion and the background portion by differentiating the line pixel data on each line pixel extraction line, select a combination of each boundary candidate point and set a guide straight line of the cutting edge portion, Further, a differential image corresponding to the direction of the blade edge of the blade edge multi-valued image is obtained, and a pixel having a large differential value is extracted along a guide straight line to obtain a blade edge contour line.

More specifically, three or more line extraction lines that intersect in the direction of the cutting edge and are parallel to the image coordinate axis in accordance with the direction of the cutting edge are extracted from the multivalued image obtained by imaging the cutting edge of the tool. The process and the line pixel data on the line pixel extraction line are differentiated to determine the boundary candidate points between the tool edge portion and the background portion, respectively, and then the point on the straightest line from the combination of the boundary candidate points And selecting a combination of the two to form a guide straight line of the cutting edge portion, and obtaining a differential image corresponding to the direction of the cutting edge of the multi-valued image of the cutting edge portion, and extracting a pixel having a large differential value along the guide straight line to obtain a cutting edge. And forming at least a part contour line.

Further, in the above-described method for extracting a cutting edge contour line by processing a captured image of a tool, a noise signal is removed from the captured image signal.

Therefore, in the method for extracting a contour of a cutting edge portion by processing a captured image of a tool according to the present invention, three or more cutting edge portions of a multivalued image of a cutting edge portion of a tool that intersect with the cutting edge direction in accordance with the direction of the cutting edge and are parallel to the image coordinate axis. Of the plurality of line pixel extraction lines, and the line pixel data on each line pixel extraction line is differentiated to obtain a boundary candidate point between the tool edge portion and the background portion. Select a combination of points on the straight line to be the guide straight line of the blade edge part, further obtain a differential image according to the direction of the blade edge of the blade edge multi-valued image, take out a pixel with a large differential value along the guide straight line Since the contour of the cutting edge is obtained, the effects of uneven illumination and background for imaging are effectively eliminated. This makes it possible to quickly and easily obtain the intended clearly defined cutting edge contour.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A preferred embodiment of a method for extracting a cutting edge contour line by image processing of a tool according to the present invention will be described below in detail with reference to the accompanying drawings.

As described above, the cutting edge contour line extraction method of this embodiment uses a multi-valued image of the cutting edge to correct the problem caused by the binarization processing of the image of the cutting edge. It is an improvement of the image processing method and takes the following measures.

First, an image of the cutting edge portion of a tool to be measured which is directly illuminated by the illuminating means is taken, and the image data taken together with the direction (either up, down, left or right) data of the corresponding cutting edge portion is calculated by a computer or the like. Take in the device.
In this case, it is assumed that the corresponding blade edge portion closer to the illumination than the background side is expressed in white.

Next, three or more, here, four line pixel extraction lines, which intersect with the direction of the cutting edge corresponding to the direction of the tool edge and are parallel to the image coordinate axis, are extracted. Extract line pixel data. FIG. 2 is a diagram for explaining image coordinates in the case of 512 × 480 pixels where the horizontal axis direction is the X axis and the vertical axis direction is the Y axis.
If the cutting edge of the tool is upward or downward, a line pixel extraction line parallel to the Y-axis direction is extracted, and line pixel data on the line pixel extraction line is extracted. If left or right, a line parallel to the X-axis direction is extracted. A line pixel extraction line is extracted, and line pixel data on the line pixel extraction line is extracted.

FIG. 3 shows four line pixel extraction lines K ₁₁ and K ₁₁ .
_12, and K _13, K _14, described below differentiated boundary candidate points (in Fig. "-" the indicated by) of the linear guide L _G of the tool cutting edge that is and asked shows an actual example is there. In the case of the actual example shown in FIG. 3, since the tool edge is upward, the four line pixels on the Y-axis direction intersecting in the direction of the edge are line pixels K ₁₁ , K ₁₂ , K ₁₃ , and K _14. Try to capture data. At this time, the X coordinate of each line pixel data may be arbitrary. In the following description,
Although only the case where the cutting edge of the tool is upward is specified and described, the case where the cutting edge is in other directions is also compliant.

Next, the differential value ΔP of each line pixel data is calculated and calculated by the following equation (1).

[0020]

(Equation 1)

Here, P _{K, Y} : luminance value of the pixel of the Y coordinate value of the line pixel data K (0 to 255,8 bits) n: number of line pixels Y: Y coordinate value.

From this calculation result, the differential value ΔP of the line pixel data shown at the left end of FIG. 3 becomes, for example, a waveform shown in FIG.

Subsequently, since the brightness of the tool edge portion in FIG. 3 is higher than the background, for each line pixel data, several peaks belonging to a positive portion with a differential value are extracted in ascending order. (X, Y) are recorded.

Here, X is a known X coordinate in the line pixel data K. In this case, the tool edge that is directly illuminated is expressed in white because it is relatively bright.
On the other hand, the background portion is expressed as black because it is darker. Therefore, the differential value at the boundary between the cutting edge and the background is positive when the cutting edge is upward or left, and negative when the cutting edge is downward or right.

Next, in order to obtain the guide straight line L _G of the tool cutting edge in FIG. 3, as shown in FIG. 5, making the combination of the plurality of three or more from the each line pixel data. That is, referring to FIG. 5, in the combination line pixel data, X
The line pixel data having the minimum coordinate X ₀ is K ₀ , the line pixel data having the maximum X coordinate X ₁ is K _1, and the X coordinate X _i ,
Let the line pixel data of X _i ′ be K _i , K _i ′.

The coordinates of the boundary candidate value D _{K0, e0} selected by differentiating the minimum X coordinate X ₀ are defined as (X ₀ , Ye ₀ ).
Here, e ₀ = 1 to n _k0 and n _k0 are the number of boundary candidate values on k ₀ . Similarly, the boundary candidate value D on the X coordinate maximum X _{1
K1, e1} coordinates of (X _1, Ye _1), sequentially selected as e ₁ = 1~n _k1. Then, determined by calculating the Y-coordinate Y _i on the X-coordinate X _i of the intermediate portion line pixel K _i on the straight line passing between these two points in the following equation (2).

[0027]

(Equation 2)

Also, the boundary candidate values D _Ki , _Yi that are on the line pixel data K _i in the middle and are closest to the calculated value Y _i ^* of equation (2).
, And the difference E _Ki between the boundary candidate values D _Ki , _Yi and the Y coordinate Y _i is calculated by the following equation (3) or (3 ′).

[0029]

(Equation 3)

Further, for all other intermediate line pixel data K _i ′ between each line pixel data K ₀ and K ₁ , the difference value E _Ki ′ is similarly obtained, and the total value E _{k is obtained.}
Is calculated by the following equation (4).

[0031]

(Equation 4)

After the values E _K for all the combination line pixel data have been obtained in this way, the combination having the minimum value E _K among them is obtained. Straight by the combination corresponds to the guide straight line L _G of the tool cutting edge illustrated in FIG.
If the number r of combinations of line pixel data is different,
What is necessary is just to change the weight of the value E _{K according to} the number. For example, if the number r of combinations is small, the weight of the value E _K is set to r / n.

Next, a differential image of the picked-up image data of the tool edge is obtained to perform edge enhancement (that is, in the differential image, the luminance of the pixel edge portion is enhanced and increased). The differential direction at this time is changed to the Y direction when the blade edge is upward or left, and to the X direction when it is downward or right, as described above, corresponding to the direction of the corresponding blade edge. FIG. 6 is an example of a differential image of the captured image data of the tool edge.

[0034] Then, in the grounds of the guide straight line L _G, the pixel coordinates large differential value of the above differential image of near, that is, extracts high luminance pixel edge portions to each, is the extracted When connecting the respective pixel edge portions, as shown in FIG. 7, the cutting edge contour line L _R of the tool to be finally Seek is obtained. That is, the so-called chipping state and the amount thereof, such as the wear state of the cutting edge of the tool, can be easily determined based on the cutting edge contour measured in this manner.

In addition, at the time of differentiation of the captured image data, smoothing is performed to remove a noise signal, and three or more pixels are extracted centering on the target pixel, and the luminance is extracted. The average value can be used as the luminance value of the target pixel.

[0036]

As described above in detail, according to the cutting edge part contour extraction method of the present invention, a multi-valued image of a cutting edge part of a tool includes a plurality of three or more lines intersecting in the cutting direction in accordance with the direction of the cutting edge. After extracting the line pixel extraction line of, after differentiating each line pixel data to obtain a boundary candidate point between the tool edge and the background,
From the combinations of candidate boundary points, select the combination of the points on the straightest line and use it as the guide straight line for the cutting edge.
In addition to obtaining a differential image corresponding to the direction of the cutting edge of the multi-valued image of the cutting edge and extracting pixels with a large differential value along the guide straight line, the influence of uneven illumination and background for imaging of the tool cutting edge is obtained. Can be satisfactorily eliminated, and consequently, the intended clearly divided cutting edge contour can be obtained quickly and easily.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating an example of an image state in a case where a captured image of a tool edge is binarized and recognized in a conventional manner.

FIG. 2 is an explanatory diagram of image coordinates used in the present invention.

FIG. 3 is a composite diagram showing four line pixels according to the present invention, a differentiated boundary candidate point, and an actual example of a guide straight line of a tool cutting edge obtained.

FIG. 4 is a waveform chart showing a differential example of the line pixel data shown in FIG.

FIG. 5 is an explanatory diagram for obtaining a straight guide shown in FIG. 3;

FIG. 6 is a diagram illustrating an example of a differential image after image processing according to the present invention.

FIG. 7 is a view showing a cutting edge contour of a tool obtained by the present invention.

[Explanation of symbols]

K ₁₁ ~K ₁₄ ... linear pixel retrieving lines _{K 0, K 1, K i} ... line pixel data L _G ... guide straight line L _R ... cutting edge contour

Claims (2)

[Claims] A step of extracting three or more line pixel extraction lines that intersect in the direction of the cutting edge according to the direction of the cutting edge and are parallel to the image coordinate axis for a multivalued image obtained by imaging the cutting edge of the tool; After differentiating the line pixel data on each line pixel extraction line to determine the boundary candidate points between the tool edge and the background, respectively, the combination of the points on the straightest line from the combinations of the boundary candidate points is determined. A step of selecting a guide straight line of the blade edge portion, and obtaining a differential image corresponding to the direction of the blade edge of the blade edge portion multi-valued image, extracting a pixel having a large differential value along the guide straight line, and obtaining a blade edge portion contour line. And a method of extracting a cutting edge contour line by processing a captured image of a tool. 2. A method according to claim 1, wherein a noise signal is removed from a captured image signal. 2. A method according to claim 1, wherein the noise signal is removed from the captured image signal. Method.