JPH1021409A - Size measuring method by picture processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily proofread a size value with irreducible minimum necessary units of a picture processor when the size of an object is measured by a picture processing by photographing the test pieces of the measuring object and calculating the size of the arbitrary part of the measuring object from the picture and resolution which is previously measured. SOLUTION: At the time of measuring the size by the picture processing, the positional relation with the camera 1 is set to be equal to that at the time of taking the measuring object and the test pieces 6A-6C of the measuring object are photographed by a camera 1. The pictures are inputted to an arithmetic unit 2 and the number of picture elements in the arbitrary part of the measuring object is measured, and the resolution is calculated from the obtained number of picture elements and the previously measured and registered size. Then the size of the arbitrary part is calculated from the number of the picture elements of the obtained arbitrary part and the previously calculated resolution. Thus, the size measured value is easily proofread.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a dimension measuring method by image processing, and more particularly to an improvement of a calibration method.

[0002]

2. Description of the Related Art When an image from a camera is input to an image processing device to measure a dimension such as a distance, the resolution of the input image (actual size that can be expressed by 1 pixel / 1 mm) is input to the image processing device. Or need to set. Since this resolution changes depending on the magnification and the focal length of the camera lens, it is necessary to calibrate the resolution or the dimensional measurement value by some method.

As a method for setting the resolution, a method using a scale is conventionally known.

In this method, an arbitrary scale is projected by a camera in advance, and the magnification and the focal length of a camera lens are adjusted so that the scale becomes the same size as the vertical or horizontal direction of the monitor screen, and the resolution is set. Is what you do. For example, FIG.
(A) As shown in FIG.
4 so that the number of pixels in the vertical direction of the monitor screen 11 such as a CRT is 51, as shown in FIG.
In the case of two pixels, a scale 12 having a length of 51 mm is displayed on the screen 11.
The magnification and the focal length of the camera lens 13 are adjusted so that the entire image is projected in the vertical direction.

[0005] However, when the image of the object to be measured is actually input, if the focus of the camera lens 13 is shifted from that at the time of setting the resolution, the value of the resolution is lost. Therefore, the dimension measurement value is out of order.

As a method for solving this problem, a calibration method using both a laser distance sensor and a lens with an autofocus mechanism is known. That is, the focus is always adjusted to the measurement object by using the lens with the auto-focus mechanism. Further, by measuring the distance between the camera and the object to be measured by the laser distance sensor, the resolution is calculated from the distance measurement value and the magnification and the focal length of the lens with an autofocus mechanism, and the dimension measurement value is calibrated. .

However, the use of a laser distance sensor and a lens with an auto-focus mechanism increases the cost.
In addition, the camera lens becomes large due to the provision of the autofocus mechanism.

[0008]

SUMMARY OF THE INVENTION It is an object of the present invention to minimize the need for an image processing apparatus when measuring the size of an object by image processing without adding a laser distance sensor and a lens with an autofocus mechanism. It is an object of the present invention to provide a method that can easily calibrate a dimension measurement value with a simple device.

[0009]

According to a first aspect of the present invention, there is provided a dimension measuring method using image processing, wherein at least one of the objects to be measured is a test piece before the dimension is measured by image processing. Measure the dimensions of the specific part in advance by an appropriate method and register it in the image processing device.When measuring the dimensions by image processing, first, the positional relationship with the camera is the same as shooting the measurement object The test piece is photographed by a camera, and the image is input to an image processing device to measure the number of pixels of the specific part of the test piece, and the obtained number of pixels of the specific part is measured in advance. Calculating the resolution from the registered dimensions, then photographing the object to be measured by a camera, inputting the image to an image processing device, and measuring the number of pixels of an arbitrary portion of the object to be measured. If, and, the number of pixels the arbitrary part obtained from the resolution previously calculated, and calculates the size of the arbitrary part.

According to a second aspect of the present invention, there is provided a method for measuring dimensions by image processing, wherein at least one of the objects to be measured is a test piece and the dimensions of a specific portion of the test piece are measured before the dimensions are measured by image processing. Measure in advance with an appropriate method,
To be registered in the image processing device, when measuring the dimensions by image processing, the test piece and the measurement object,
The same positional relationship with the camera is taken by the camera, and each image is input to an image processing device to measure the number of pixels of the specific part of the test piece and the number of pixels of an arbitrary part of the measurement object. And the number of pixels of an arbitrary portion of the measurement object obtained by this measurement, the number of pixels of a specific portion of the test piece, and the dimensions of the specific portion measured and registered in advance, the measurement object The dimensions of the arbitrary part are calculated.

That is, if the test piece and the other object to be measured are successively photographed by the camera and subjected to image processing, the resolution does not change during that time. Therefore, assuming that the size of the specific portion of the test piece is L ₆ and the number of pixels of the specific portion of the test piece is G ₆ , the resolution R ₆ is R ₆ = L ₆
It can be calculated by the / G _6. Accordingly, by using the calculated resolution R ₆ and letting the number of pixels of an arbitrary portion of the measurement target be G _x , the dimension X of the arbitrary portion is X = G _x
- can be calculated by R _6, it is automatically calibrated.

Further, even if the resolution R ₆ is not determined once, the size L ₆ of the specific portion of the test piece, the number of pixels G _{6 of} the specific portion of the test piece, and the number of pixels G _x of the arbitrary portion of the measurement object can be obtained. , The dimension X can be directly calculated by X = G _x · (L ₆ / G ₆ ), and calibration is automatically performed.

In any case, even if the camera is out of focus, calibration is performed at the time of measurement, so that dimensional measurement with a certain degree of accuracy is possible.

[0014]

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

(First Example) FIG. 1 shows an example of the overall configuration of an image processing apparatus to which a dimension measuring method according to claim 1 of the present invention is applied, and a camera 1 and an arithmetic unit for performing image processing and necessary control. 2, a memory 3, a console 4 such as a keyboard,
It comprises a monitor 5, various test pieces 6A to 6C, a measurement stage 7, and a moving mechanism 8. Camera 1,
The memory 3, console 4, monitor 5, and moving mechanism 8 are connected to the arithmetic unit 2.

The moving mechanism 8 is for setting the relative positional relationship between the camera 1 and the object to be measured and the test piece constant, and under the control of the arithmetic unit 2, for example, only the camera 1 Alternatively, only the measurement object and the test piece are moved to a predetermined position on the measurement stage 7.

In the following description, it is assumed that the object to be measured and the test piece are moved, and the positional relationship and posture of the camera 1 with respect to the measurement stage 7 are fixed, but the reverse is also true.

The test piece 6A is one extracted from a type A measurement object, and as shown in FIG.
The length L _6A of a specific side is registered in advance by measuring the length L _6A with an appropriate dimension measuring device and storing the length L _6A in the memory 3 in association with the type A of the measurement object.

Similarly, the test piece 6B is one extracted from the type B measurement object, and measures the length L _6B of a specific side in advance and associates it with the type B in the memory 3. Stored and registered. Test piece 6C is C
One of the types of measurement objects is extracted, and the length L _6C of a specific side is measured in advance, stored in the memory 3 in association with the type C, and registered.

Next, the function of each unit will be described with reference to FIGS. 3 and 4, together with the procedure of calculating the resolution and measuring the dimensions.

In measuring the dimensions, the operator operates the console 4.
Is operated to input the type of the measurement target, thereby specifying the type of the measurement target in the arithmetic unit 2. By specifying this type, the arithmetic unit 2 selects the type of the measurement object and the test piece (step S1 in FIG. 3), and reads out the corresponding registered dimensions from the memory 3 (step S2).

In the following description, type A is selected and registered
Only dimension L_6AIs read. In FIG. 3,
Registered dimensions of test pieces without distinguishing between types A to C
To L ₆, The number of pixels of the test piece is G₆Obtained from these
R₆, The number of pixels of an arbitrary part of the measurement object
_x, The dimensions of the arbitrary portions are represented by X, respectively.

The arithmetic unit 2 first operates in a resolution calculation mode, controls the moving mechanism 8 to move the test piece 6A to a predetermined position on the measurement stage 7, and, as shown in FIG. Input (step S3).

Further, the arithmetic unit 2 performs binarization processing on the input image of the test piece 6A as image processing (step S).
4) The number of pixels G _6A on the specific side whose dimensions have been measured in advance is measured (step S5), and the resolution R _6A is calculated by the formula R _6A = L _6A / G _6A and stored in the memory 3 (step S5).
6). The binarized image is displayed on the monitor 5.

After that, the arithmetic unit 2 shifts to a dimension measurement mode, controls the moving mechanism 8 and places a type A measurement object at a predetermined position on the measurement stage 7 (hereinafter referred to as a measurement object A).
Are sequentially moved, and the images are sequentially input from the camera 1 (step S7). That is, the state of the camera 1 with respect to the test piece 6A and the state with respect to the measurement target A are the same.

The arithmetic unit 2 performs binarization processing on the input image of the measurement object A as image processing (step S8), and obtains the number of pixels G _xA of an arbitrary portion (eg, an arbitrary side) (step S8).
10).

Further, the arithmetic unit 2 calculates the size X of the arbitrary portion.
_A is calculated using the number of pixels G _xA of the measurement object A and the previously calculated resolution R _6A by the formula X _A = G _xA · R _6A (step S11), and the monitor 5 displays the measurement result X _A ( mm)
Is displayed (step S12).

When the binarization processing is performed in the dimension measurement mode, the arithmetic unit 2 determines the presence or absence of an object to be measured from the image after the binarization processing (step S9). If there is an object to be measured, the number of pixels G _xA is determined in the next step S10, but if not, the measurement is terminated.

Although the above description has been made with respect to the type A measurement object, the same applies to other types B and C measurement objects.

(Second Example) Next, a dimension measuring method according to a second aspect of the present invention will be described with reference to FIGS.
In this example, when measuring dimensions by image processing, first,
The test piece is photographed by the camera with the same positional relationship with the camera as that of the object to be measured, and the image is input to the image processing device, and the number of pixels of the specific part of the test piece is measured. , An image of the object to be measured is captured by a camera, the image is input to an image processing apparatus, and the number of pixels of an arbitrary portion of the object to be measured is measured. The size of the arbitrary portion of the measurement target is calculated from the number of pixels of the specific portion of the test piece and the previously measured and registered dimensions of the specific portion. The overall configuration of the image processing apparatus is the same as that of the previous example and is shown in FIG. 1, and includes a camera 1, an arithmetic unit 2 for performing image processing and necessary control, a memory 3, a console 4 such as a keyboard, and a monitor 5 , Test pieces 6A to 6C, and measurement stage 7
And a moving mechanism 8. However, in the procedure of FIG. 3, the resolution is once calculated and stored. However, in this example, the resolution is not calculated. Instead, the number of pixels of a specific portion of the test piece is measured and stored in the memory 3 as described later. I do.

Also in this example, the camera 1, the memory 3, the console 4, the monitor 5, and the moving mechanism 8 are connected to the arithmetic unit 2, and under the control of the arithmetic unit 2, the moving mechanism 8 By moving only 1 or only the measurement object and the test piece to a predetermined position on the measurement stage 7,
The relative positional relationship between the camera 1 and a measurement object or a test piece is set to be constant. Further, in the following description, as in the previous example, only the measurement target and the test piece are moved, and the positional relationship and the posture of the camera 1 with respect to the measurement stage 7 are fixed. The same applies to the test pieces 6A to 6C, and one is extracted from each of the measurement objects of each type A to C, and the length L _6A of any one side,
L _6B and L _6C are measured in advance by an appropriate dimension measuring instrument, and stored in the memory 3 in correspondence with the types A to C of the measurement object, and registered.

Next, with reference to FIG. 5, the function of each part will be described together with the procedure of dimension measurement.

In measuring the dimensions, the operator operates the console 4.
Is operated to input the type of the measurement target, thereby specifying the type of the measurement target in the arithmetic unit 2. By specifying this type, the arithmetic unit 2 selects the type of the measurement object and the test piece (step S21 in FIG. 5), and reads out the corresponding registered dimensions from the memory 3 (step S22).
In the following description, it is assumed that the type A is selected and the registered dimension _L6A is read, as in the previous example. Also in FIG. 5, as in FIG. 3, the types A to C are not distinguished, the registered size of the test piece is L ₆ , the number of pixels of the test piece is G ₆ , and the number of pixels of an arbitrary portion of the measurement object is G. _x and the dimension of the arbitrary portion are represented by X, respectively.

The arithmetic unit 2 controls the moving mechanism 8 to move the test piece 6A to a predetermined position on the measuring stage 7 and input the image from the camera 1 (step S2).
3), as an image processing an image of the test piece 6A to binarization processing (step S24), and obtains the number of pixels G _6A particular sides previously dimension measurement (step S25). The number of pixels G _6A is stored in the memory 3 instead of the resolution.

Next, the arithmetic unit 2 controls the moving mechanism 8 to
The measurement object of type A (hereinafter, referred to as measurement object A) is sequentially moved to a predetermined position on the measurement stage 7 and the camera 1
Are sequentially input (step S26), the image of the measurement target A is binarized as image processing (step S27), and the number of pixels _GxA of an arbitrary portion, for example, an arbitrary side is obtained (step S29).

[0036] Then the dimensions X _A of any part of the operation unit 2 is the measurement object, and the number of pixels G _xA of the arbitrary part, the number of pixels G _6A particular side of the test piece 6A measured previously, the test piece X _A = G _xA using the registered dimension L _{6A of 6A}
· Calculated by the formula (L _6A / G _6A ) (step S3
0), the measurement result X _A (mm) is displayed on the monitor 5 (step S31). Before _{obtaining the} number of pixels _GxA in step S29, the presence or absence of a measurement target is determined in step S28 from the image after the binarization processing. If there is no measurement target, the measurement ends.

Although the above description has been made with respect to the type A measurement object, the same applies to other types B and C measurement objects.

As can be seen from the above description, as a method of calibrating the resolution in the image processing, at least one of the measurement objects is used as a test piece before the dimensions of the measurement object are measured by the image processing. The dimensions of the part are measured in advance by an appropriate method, registered in the image processing apparatus, the test piece is photographed by the camera in the same manner as the photographing of the measurement object with the positional relationship with the camera, and the image is taken. Is input to the image processing apparatus to measure the number of pixels of the specific part of the test piece, and the resolution is calculated from the obtained number of pixels of the specific part and the previously measured and registered dimensions. Thus, the resolution can be automatically calibrated by the function of the image processing apparatus itself.

[0039]

According to the present invention, at least one of the objects to be measured is used as a test piece, and a laser distance sensor and a lens with an autofocus mechanism are added only by measuring the dimensions of a specific portion of the test piece in advance. Without doing
With the function of the image processing apparatus itself, it is possible to automatically calibrate the resolution or the dimension measurement value.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an example of an apparatus configuration for dimension measurement by image processing according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a test piece.

FIG. 3 is a diagram showing an example of a procedure of dimension measurement by image processing according to the embodiment of the present invention.

FIG. 4 is a diagram showing a shooting state of a test piece.

FIG. 5 is a view showing another example of the procedure of dimension measurement by image processing according to the embodiment of the present invention.

FIG. 6 is a diagram showing a resolution setting method using a conventional scale.

[Explanation of symbols]

 Reference Signs List 1 camera 2 arithmetic unit 3 memory 4 console 5 monitor 6A to 6C test piece 7 measuring stage 8 moving mechanism

Claims (2)

[Claims] In a dimension measuring method by image processing for measuring a dimension of a measurement object from an image of the measurement object obtained by a camera, at least one of the measurement objects is measured before the dimension is measured by the image processing. Measure the dimensions of a specific part of the test piece in advance using an appropriate method and register it in the image processing device. When measuring the dimensions by image processing, first measure the positional relationship with the camera. Shooting the test piece with a camera in the same way as shooting an object, inputting the image to an image processing device and measuring the number of pixels of the specific portion of the test piece, Calculating the resolution from the number of pixels and the dimensions that have been measured and registered in advance. Measuring the number of pixels of an arbitrary part of the measurement object, and calculating the dimensions of the arbitrary part from the obtained number of pixels of the arbitrary part and the previously calculated resolution, Dimension measurement method by processing. 2. A dimension measuring method by image processing for measuring dimensions of a measurement object from an image of the measurement object obtained by a camera, wherein at least one of the measurement objects is measured before the dimension is measured by the image processing. As a test piece, the dimensions of a specific part of the test piece are measured in advance by an appropriate method and registered in an image processing device. When measuring the dimensions by image processing, the test piece and the object to be measured are separated by a camera. The same positional relationship as above is taken by a camera, and each image is input to an image processing device to measure the number of pixels of the specific part of the test piece and the number of pixels of an arbitrary part of the measurement object. , And the number of pixels of an arbitrary part of the measurement object and the number of pixels of a specific part of the test piece obtained by this measurement, and the previously measured and registered dimensions of the specific part From dimension measurement method by the image processing, characterized in that, to calculate the size of the arbitrary portion of the measurement object.