JPH08247730A - Edge detecting method of dimension measuring instrument - Google Patents

Abstract

PURPOSE: To provide an edge detecting method of dimension measuring instrument by which the binarized edge of an object can be detected in a short time without using any large-scale hardware. CONSTITUTION: A dimension measuring instrument measures the dimension of an object to be measured by receiving a light beam from the object by means of a CCD image sensor and binarizing the output signal of the sensor, and then, detecting edges from the binarized signals and measuring the dimension of the object from the edge sections. All picture element data of the output signal of the image sensor which outputs 5,000 picture element data are divided into 10 areas of 500 data each and the leading edge data of each area is extracted and binarized. Then one binary data indicating bright '0' or dark '1' is obtained from each area from the 10 binarized data. In addition, a leading side area where a bright area and dark area exists adjacently to each other is further divided into a plurality of subareas and narrowing down the subareas in the same way and, finally, picture elements having abnormal binarized values, namely, edges are detected from the output signal of the image sensor.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention irradiates a measuring object with a light beam, receives the passing light thereof by a one-dimensional light receiving sensor, and binarizes the output signal of the measuring object. The present invention relates to a method for detecting an edge of a binarized signal in a dimension measuring device for detecting a position.

[0002]

2. Description of the Related Art Generally, an object to be measured is irradiated with light, and transmitted light or reflected light thereof is received by a one-dimensional light receiving sensor (CCD image sensor), and the light receiving signal is binary with a threshold value held in advance. There are some that measure the outer shape (dimension) of the object to be measured or detect the position by changing the method. Figure 5
Is a transmission type dimension measurement in which the light is emitted from the projector 1 toward the measurement object 2, the transmitted light is received by the light receiver 3 including a one-dimensional CCD image sensor, and the light reception signal is processed by the signal processing unit 4. It is a device.

In FIG. 6, light is emitted from a light projector 1 toward a measurement object 2, the reflected light is received by a light receiver 3 including a one-dimensional image sensor, and the light reception signal is processed by a signal processing unit 4. It is a reflection type dimension measuring device. FIG. 7 shows the output of the CCD image sensor of the transmission type dimension measuring device of FIG.
This is to explain the method of quantifying and measuring, CCD
Threshold value TH that presets the light reception signal Ps of the image sensor
By this, binarization is performed as shown in FIG. The dimension of the measurement target is measured from the relationship between the edge e ₁ (bright → dark edge) and the edge e ₂ (dark → bright edge) in FIG. 7B.

Conventionally, in this type of dimension measuring apparatus, in order to improve the measurement accuracy, there is a method in which a threshold is set for each light receiving area (pixel) of the one-dimensional image sensor and binarized to detect an edge. As a means for realizing that, for example, using hardware that performs pipeline processing,
For example, there is one using software that sequentially reads out accumulated light reception data and binarizes it.

[0005]

The above-described conventional binarization and edge detection method using hardware has a problem in that the scale of the hardware is large, and the method using software processes each data individually. Since the processing is performed, there is a problem that the processing time becomes long. The present invention has been made in view of the above problems, and an object of the present invention is to provide an edge detection method capable of performing binarized edge detection in a short time without using large-scale hardware. I am trying.

[0006]

An edge detecting method of a dimension measuring apparatus according to claim 1 of the present application irradiates a measuring object with a light beam, and primaryally measures the light beam from the measuring object. The original light receiving sensor receives light, binarizes each pixel output of this one-dimensional light receiving sensor, and detects the bright and dark edges of the binarized light receiving signal, thereby measuring the dimension or position of the measurement target. In the above, the pixel data output from the light receiving sensor is divided into a plurality of predetermined areas, one or more pixels in each area are selected, and the pixel data are binarized. "1" or "0" for each area, or binary 1
Data is obtained, and the areas where the binary data of the arranged areas are adjacent to each other are extracted from "1" to "0" or "0" to "1". The pixel data from one selected pixel to the other selected pixel is again divided into a plurality of areas each having a plurality of pixel data, and at least one pixel in each area is similarly selected, and these pixels are selected. Area that is binarized to obtain binary 1 data for each area, and the binary data of each arrayed area changes from "1" to "0" or "0" to "1" in the adjacent areas. Is extracted, and as described above, the area division and the extraction of the binary different values of the adjacent areas are repeated, and finally the adjacent pixels of the pixel data of the one-dimensional light receiving sensor are binary-differentiated. By extracting the pixel that becomes the value,
The edge is detected.

According to a second aspect of the present invention, there is provided an edge detecting method for a dimension measuring apparatus, wherein the pixel data output from the light receiving sensor is divided into a plurality of predetermined plurality of pixels in the same dimension measuring apparatus as in the first aspect. Divide into areas, select several pixels from each area 1
Is it "1" or "0" for each area,
Binary 1 data is obtained, and the area where the binary data of the arranged areas are adjacent to each other is extracted from "1" to "0" or "0" to "1". For pixel data between one selected pixel and the other selected pixel in adjacent areas, select one or more pixels near the center,
Each pixel data is binarized, and it is determined whether or not there is binary data in the binarized pixel that is different from the area side where the binary data of the same value is continuous. If there is, that pixel is an edge. If not, select one to several pixels between the selected pixel and the selected pixel in the area on the different value side,
Edges are detected by repeating similar processing.

According to a third aspect of the present invention, there is provided an edge detecting method for a dimension measuring apparatus, wherein in the dimension measuring apparatus similar to the first aspect, among the pixel data output from the light receiving sensor,
Whether 1 or several pixel data are selected from the vicinity of the beginning, the vicinity of the center, and the vicinity of the tail, and binarized, and whether the values are "1" or "0" near the beginning, the vicinity of the center, and the vicinity of the tail. One binary data is obtained, and for the binary data near the center, which of the two values is different between the beginning and the end is extracted, and then the difference between the center and the center is extracted. The one or several pieces of pixel data in the vicinity of the center between the end portions were selected and binarized to obtain new binary 1 data in the vicinity of the center, and the new pixel data in the vicinity of the center, The vicinity where other pixel data has a different value in binary is extracted, and thereafter, the above extraction processing is repeated between the new center vicinity and the vicinity having a different value. Pixels having different binary values are extracted and the point is detected as an edge.

[0009]

The present invention will be described in more detail with reference to the following examples. FIG. 1 is a block diagram showing the configuration of a dimension measuring apparatus in which the present invention is implemented. This size measuring device includes a light source 2 such as an LED or a halogen lamp, a lens 3, a light projecting unit 1 for irradiating a parallel light beam 4 onto a measurement object 5, and a one-dimensional CCD image sensor 6, and a parallel light source. A light receiving unit 7 that receives the light beam 4, an amplifier 11 that amplifies the light receiving signal, an A / D converter 12 that discretizes the amplified signal, and a data memory 13 that stores the light receiving signal.
A threshold value memory 14 for storing a threshold value for each pixel, and a timing control circuit 1 for applying a timing signal to each constituent circuit.
5, a light source lighting circuit 16, a CPU 17, a binary memory 18, and an output memory 19.

In this dimension measuring apparatus, when the light source lighting circuit 16 is driven by the signal from the timing control circuit 15, the light source 2 is turned on, and the collimated light beam 4 is irradiated on the measuring object 5. When the measuring object 5 is irradiated, the parallel light beam 4 is blocked by the width of the measuring object 5 and becomes a shadow.
The output of the D image sensor 6 is divided into a region of large received light signal (bright) and a small received light signal (dark). This light reception signal is binarized with a predetermined threshold value, a binary value change edge of the light reception signal is detected, and the edge and the length of the edge are measured.
The dimensions of can be measured. In this dimension measuring device, the amount of light incident on each pixel is different even when the measurement target 5 is not installed. Therefore, the threshold value is stored in the threshold value memory 14 corresponding to each pixel. The binarization and the edge detection are executed by the signal processing unit 8 after the amplifier 11. CCD
When the number of pixels of the image sensor 6 is large, if the binarization processing is sequentially executed for all the pixels, the processing time becomes long, so the following processing is executed.

Now, assume that the number of data (number of pixels) corresponding to the measurement area of the CCD image sensor 6 is 5000. It takes time to read out all the threshold values corresponding to these data sequentially, compare them, binarize them, and detect edges. Therefore, when the measurement object 5 is normally placed in the parallel light 4,
From the head to the tail of the CCD image sensor 6, the outputs are bright, dark, and bright, so this characteristic is utilized to efficiently perform binarization and edge detection. First, as shown in FIG. 2A, the area of all data is divided into 10 areas of 500 data. This division method is
The area may be divided into the same number, or the size may be changed depending on the area.

Next, data of one or more points from each area is read from the data memory 13 and read from the threshold memory 14 at the corresponding position and binarized. For each area, for example, as shown in FIG. 2B, from the number of data of 500, the ten data on the leftmost side of the area is binarized. In this example, if 10 points are dark [binary “1”] for 10 data, the area is a dark area, and if 0 points are dark for 10 data, the area is If the bright area and 1 to 9 points of the 10 data are dark, it is determined as an indeterminate area. However, the data position (point position) to be extracted for each area may be continuous or discrete, and the extraction position may be at the beginning or at the center as shown in FIG. It may be the tail. Further, the threshold value for determining darkness, uncertainty, and lightness may be set to an arbitrary value, and the uncertainty section may not be provided.

When the measuring object 5 is set in the irradiation area of the parallel light beam 4 and the outer diameter of the measuring object 5 is measured, the bright and dark discrimination result of each area is shown in FIG. (A) of FIG. In this example, the bright and dark areas are adjacent to each other in the areas 2 and 3, and the areas 8 and 9, and in FIG. Since it is determined that the beginning of area 2 is bright, the beginning of area 3 is dark, the beginning of area 8 is dark, and the beginning of area 9 is bright. Therefore, the edge position of light → dark is area 2 It is considered that the dark → bright edge position is in the area 8. Therefore, this time, area 2 and area 8 are similarly divided into a plurality of areas (for example, 10 areas of 50 data each), and 1 to several data are extracted from each of the re-divided areas and binarized. Then, the light and dark of the area are discriminated. Then, from the subdivided area group, areas in which edges are likely to exist are extracted. Then, such processing is repeated, and the positions where the adjacent pixel data finally become bright are extracted. This position is the edge position. If the number of areas decreases to about 10 as the area division is repeated, the small areas where edges are likely to exist, and finally all the data,
You may perform a binarization process. It is 10 data that is binarized in each area in the stage from FIG. 2A to FIG. 3A, and the binarization processing can be omitted for the remaining 490 data. Further, the binarization processing time can be shortened.

Other methods of edge detection are shown in FIGS.
The same is true up to (a) of FIG. 3, but for detecting the edge position for area 2 in (a) of FIG. 3, for example, one or more data near the center of area 2 [in (b) of FIG. 10
[Individual], and similarly classify into light / dark / indeterminate. If the result is dark, for example, the bright-> dark edge exists on the left side of the center of the area 2 in FIG. 3B. Therefore, for the left half area of area 2, one to several data are extracted from the data near the center,
Bright / dark / indetermination is performed, and it is determined whether there is an edge on the left side or an edge on the right side. Then, for the area on the side having the edge, similar processing is performed thereafter, and the edge position is narrowed down. Then, when the position where the edge is likely to be present is narrowed down, the edge position detection processing is performed for the area. That is, the received light data and the threshold value data are read from the data memory 13 and the threshold value memory 14 from the beginning of the detection target area, and are binarized to detect the light-> dark edge. If there is an uncertain portion in the middle of narrowing down, narrowing down is stopped and edge detection processing is performed for that portion. Similarly, the dark → bright edge (area 8) position is also detected. When the detection of the light-> dark edge position and the dark-> bright edge position is completed, the outer dimensions of the measurement target can be calculated from the respective edge positions, as illustrated in FIG. 3C, for example.

Next, another method for edge detection will be described. Here, for example, as in the case of detecting the position of the measurement target, the case where there is one edge from the light → dark edge or the dark → bright edge in the received light data will be taken up. The number of data to be measured is 5000 as in (a) of FIG. First, as shown in FIG. 4A, an area including one or more data (for example, 10 pieces) is extracted for each of the vicinity of the beginning, the vicinity of the center, and the vicinity of the tail of the data 5000, Whether each area is bright / dark / indeterminate is determined from the corresponding data and the threshold value.

Here, the determination result is, for example, (b) of FIG.
As shown in the example, if the vicinity of the head is bright, the vicinity of the center is bright, and the vicinity of the tail is dark, it is considered that the light-> dark edge is in the area on the right (tail side) of the center. Therefore, next, as for the area on the right side of the center, one or more data near the center is extracted in the same manner as in (b) of FIG. 3 to determine whether it is bright / dark / indeterminate, and the center thereof is determined. Judge whether there is an edge on the left side of the neighborhood or on the right side,
This process is repeated to further narrow down the edge position.

[0017]

According to the invention of claim 1, in order to binarize the output of the light receiving sensor and detect an edge, the output data is divided into a plurality of areas, and one or more data is selectively extracted in each area. , Each area is binarized, the area including the edge is determined based on the relationship with the adjacent area, the area is further divided into a plurality of pieces, and the same processing is repeated to finally detect the edge of the output data. Therefore, since the binarization is not performed by using all the output data and the threshold value, the edge position of the binarization process can be detected at high speed with a small hardware configuration.

According to the second aspect of the invention, in order to binarize the output of the light receiving sensor and detect the edge, the output data is divided into a plurality of areas, and the edge is included in relation to the adjacent areas. An area is determined, and then one or more pixels near the center of the area are selected, each pixel data is binarized, and the binarized pixel is different from the area where binary data is continuous. Since the position of the edge is narrowed down by determining whether there is binary data, the edge can be detected at high speed by small-scale hardware, and two edges are included in the output signal of the light receiving sensor. It is effective for edge detection when there is.

According to the third aspect of the invention, with respect to the output pixel data, one or several pieces of pixel data are selected from the vicinity of the beginning, the vicinity of the center, and the vicinity of the rear, and binarized, For each of the vicinity of the beginning, the vicinity of the center, and the vicinity of the tail, whether it is "1" or "0", or binary 1 data is obtained, and for the binary data near the center, either the vicinity of the beginning or the vicinity of the tail Is extracted if it is a binary different value, and it is judged that there is an edge on the different value side than the vicinity of the center, and this time, between the vicinity of the center, the vicinity of the center, and the vicinity of the end that was the different value. , 1 to several pixel data around the new center are selected and binarized to obtain binary 1 data around the new center, and the pixel data value around the new center and pixels around other To extract the vicinity where the data value is binary different and repeat this process. The edge is narrowed down, so small-scale hardware can detect the edge at high speed, and if the output signal of the light-receiving sensor has one edge, that is, one edge It is valid.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a dimension measuring apparatus in which the present invention is implemented.

FIG. 2 is a diagram for explaining area division of output data of a light receiving sensor and determination processing of bright / dark / indeterminate of each area in the dimension measuring apparatus of the embodiment.

FIG. 3 is a diagram illustrating an example of bright / dark determination result of the same area and calculation of outer dimensions of a measurement target.

FIG. 4 is a diagram illustrating another binarization process and an edge detection example.

FIG. 5 is a schematic view showing a general transmission type dimension measuring device.

FIG. 6 is a schematic view showing a general reflection type dimension measuring device.

FIG. 7 is a diagram illustrating edge detection of a dimension measuring device using a one-dimensional light receiving sensor.

[Explanation of symbols]

 1 Light Emitting Section 2 Light Emitting Element 3 Lens 4 Parallel Light Beam 5 Object to be Measured 6 CCD Image Sensor 7 Light Receiving Section 8 Signal Processing Section

Claims (3)

[Claims] 1. A measurement object is irradiated with a light beam, a light beam from the measurement object is received by a one-dimensional light receiving sensor, each pixel output of the one-dimensional light receiving sensor is binarized, and a binarized light receiving signal is obtained. In the dimension measuring device for measuring the dimension of the object to be measured by detecting the bright and dark edges of, the pixel data output from the light receiving sensor is divided into a plurality of predetermined areas, Select one or more pixels in the area, binarize the pixel data, obtain "1" or "0" for each area, obtain binary 1 data, and arrange them. In the area where the binary data of the areas are adjacent to each other, the area from “1” to “0” or “0” to “1” is extracted, and then the selected pixel of one of the adjacent areas to the other is extracted. For the pixel data up to the selected pixel, Divide into multiple areas of pixel data, select at least one pixel in each area, binarize these pixels,
One binary data is obtained for each area, and the areas where the binary data of the arranged areas are adjacent to each other are extracted from "1" to "0" or "0" to "1". As described above, by repeating the area division and the extraction of the binary different values of the adjacent areas, the difference between the adjacent pixels of the pixel data of the one-dimensional light receiving sensor is finally determined as the binary different value. Edge detection method for a dimension measuring device, which detects an edge by extracting a pixel. 2. A measurement object is irradiated with a light beam, the light beam from the measurement object is received by a one-dimensional light receiving sensor, each pixel output of the one-dimensional light receiving sensor is binarized, and the binarized light receiving signal is obtained. In the dimension measuring device for measuring the dimension of the object to be measured by detecting the bright and dark edges of, the pixel data output from the light receiving sensor is divided into a plurality of predetermined areas, Select one or more pixels in the area, binarize the pixel data, obtain "1" or "0" for each area, obtain binary 1 data, and arrange them. In the area where the binary data of the areas are adjacent to each other, the area from “1” to “0” or “0” to “1” is extracted, and then the selected pixel of one of the adjacent areas to the other is extracted. About the pixel data up to the selected pixel, 1 near the center By selecting more than one pixel, each pixel data is binarized, and it is determined whether or not there is binary data in the binarized pixel that is different from the area side where the binary data of the same value is continuous. If so, the pixel is an edge, and if not, one to several pixels between the selected pixel and the selected pixel in the area on the different value side are further selected, and similar processing is repeated to obtain an edge. EDGE DETECTION METHOD FOR DIMENSION MEASURING APPARATUS USING DETECTION. 3. A light beam radiated to a measuring object, a light beam from the measuring object is received by a one-dimensional light receiving sensor, each pixel output of the one-dimensional light receiving sensor is binarized, and the binarized light receiving signal is obtained. In the dimension measuring device for measuring the dimension of the measurement object by detecting the bright and dark edges of the, the pixel data output from the light receiving sensor is set to 1 in the vicinity of the head, in the vicinity of the center, and in the vicinity of the tail. Or select several pixel data and binarize them, and obtain 1 binary data of "1" or "0" near the beginning, near the center, and near the tail, and obtain binary data and From the binary data, it is extracted which of the vicinity of the beginning and the vicinity of the tail is a binary different value, and then the 1 near the center between the center and the vicinity of the end and the different value Or select a few pixel data and binarize it Obtaining binary 1 data in the vicinity, extracting the neighborhood where the new pixel data near the center and the pixel data in the other neighborhood are different in binary value, and after that, it becomes different from the new center neighborhood. The edge detection method of the dimension measuring apparatus, which repeats the above-mentioned extraction process between the adjacent pixels and the adjacent pixels, or extracts the last adjacent pixels or pixels having different binary values and detects the points as edges.