CN109520436B - A machine vision-based automatic measuring system for three-dimensional size of butterfly spring and its measuring method - Google Patents

Abstract

The invention discloses a machine vision-based automatic measuring system for the three-dimensional size of a butterfly spring, which includes a hardware part, the hardware part includes a detection device, and the detection device includes a base, the upper side of the base is symmetrical through a corner connecting piece Two vertical rods are fixedly installed, and a horizontal rod is fixedly installed between the two vertical rods near their upper ends. When performing on-site measurement, it is only necessary to place the butterfly spring in the area to be measured. When testing multiple butterfly springs It can be replaced at any time, realizes automatic judgment and detection, is easy to operate, simple, and has high detection efficiency; the invention also discloses a machine vision-based automatic measuring system for three-dimensional size of a butterfly spring and a measuring method, including three steps. The machine vision measurement method, through the non-contact measurement method, reduces the damage of the butterfly spring in the manual measurement process, especially the smaller size of the butterfly spring.

Description

A machine vision-based automatic measurement system for three-dimensional size of butterfly spring and its measurement quantitative method

technical field

The invention relates to the technical field of optical three-dimensional size detection, and more particularly, to an automatic measuring system for three-dimensional size of a butterfly spring based on machine vision and a measuring method thereof.

Background technique

Part size detection has high repeatability and high intelligence. It is affected by random errors in parts processing. Part size detection must be artificially judged. However, manual detection has a large workload, slow speed, easy fatigue, poor stability, and high cost. For complex, tiny, mass-produced parts, manual inspection is labor-intensive and difficult.

Butterfly spring is a kind of mechanical parts widely used in aerospace, mechanical equipment industry and other fields. Because the butterfly spring has the shape of a bowl and dish, it is difficult to measure the size. The existing detection technology is to measure with general gauge, stop gauge and plug gauge. Each butterfly spring needs to be measured 6 times to measure the full size and judge whether it is qualified or not. The workload is large, the speed is slow, and the naked eye is easily fatigued. Parts are easily damaged during manual inspection.

The existing machine vision-based part size inspection is not applied to the size inspection of butterfly springs. The existing machine vision-based part inspection can only obtain two-dimensional dimensions, but only convert the three-dimensional size into a two-dimensional plane for measurement. Industrial cameras The installation and detection devices are relatively complex. The detected parts have relatively regular shapes and cannot be used for 3D measurement of butterfly springs.

Due to the high requirements of detection accuracy and efficiency of butterfly springs in small size and mass production, an automated, high-precision, non-contact, and high-efficiency dimensional measurement and evaluation system is of great significance. Therefore, a butterfly spring based on machine vision is proposed. A spring three-dimensional dimension automatic measuring system and its measuring method.

SUMMARY OF THE INVENTION

In view of the deficiencies in the prior art, the purpose of the present invention is to provide a machine vision-based automatic measuring system for the three-dimensional size of a butterfly spring and a measuring method thereof, so as to solve the problems raised in the above-mentioned background art.

For achieving the above object, the present invention provides the following technical solutions:

A machine vision-based three-dimensional automatic measurement system for butterfly springs includes a hardware part, the hardware part includes a detection device, the detection device includes a base, and two sides are symmetrically and fixedly installed on the upper side of the base through a corner connector. A vertical rod is installed between the two vertical rods near their upper ends, and the middle of the two vertical rods is respectively fixed with a camera CL and a camera CR through a connecting seat, and the bottom of the middle of the horizontal rod is fixed. A camera CT is fixedly installed through the fitting seat, a common optical lens is installed on the camera CL and the camera CR, a telecentric optical lens is installed on the camera CT, a backlight light source is installed on the bottom of the base, and A frosted glass is installed on the upper part of the base above the backlight light source, a butterfly spring to be measured is placed at the center of the upper part of the frosted glass, and the upper part of the frosted glass is located on the outer periphery of the butterfly spring with several marks point.

Further, the marking points are ordinary circular marking points with outer black and inner circle.

Further, it also includes a software part, the software part includes CT, CL, CR camera display window and CT, CL, CR camera calibration interface and display window for measuring size, butterfly spring size standard setting window, butterfly spring size Whether the window is qualified;

The CT, CL, CR camera display windows are used for real-time display of images collected by the cameras;

The CT, CL, and CR camera calibration interfaces are used for camera calibration to prepare for the size detection of the butterfly spring;

The measurement size display window is used for real-time display of measurement results;

The described butterfly spring size standard setting window is used to interactively input standard size and allowable error;

The whether the butterfly spring is qualified or not is used to display whether the size of the detected butterfly spring is qualified.

Further, the measurement results include: the inner diameter, outer diameter, and height of the butterfly spring, the detection error of inner diameter, outer diameter, and height, whether the inner diameter, outer diameter, and height conform to the standard size and tolerance requirements. To the butterfly spring, the test results are all 0.

The present invention also proposes a machine vision-based automatic measuring method for the three-dimensional size of the butterfly spring, comprising the following steps:

The first step, each time the camera CT collects a frame of image, it is compared with the previous frame image, and the grayscale information of the front and rear images automatically enters the detection program within a certain threshold range;

The second step is to use the telecentric lens to detect the inner and outer diameters of the butterfly spring, and obtain the contour information of the butterfly spring to be detected through real-time image processing;

The third step is to use the automatic measurement system to measure the three-dimensional size of the butterfly spring, distinguish the mark point and the butterfly spring, reproject the mark point to the world coordinate system for plane fitting, and obtain the butterfly spring of the camera CL according to the ellipse fitting method. The coordinates of the center point of the upper ellipse and the coordinates of the center point of the ellipse on the butterfly spring of the camera CR, according to the principle of left and right stereo vision, are re-projected to the world coordinate system after left and right matching, to obtain the upper ellipse center of the butterfly spring, and calculate the center of the upper ellipse to The distance to the fitted plane calculates the height.

To sum up, the present invention mainly has the following beneficial effects:

1. The present invention realizes the automatic detection of the inner diameter, outer diameter and height of the butterfly spring in three dimensions, reduces the workload of using the traditional plug gauge measurement method, improves the detection efficiency and accuracy, and ensures that the absolute error of the inner and outer diameters is within 0.02. -0.1mm, the absolute error of height is 0.02mm;

2. In the present invention, the detection device is an optical measuring instrument based on machine vision. When performing on-site measurement, it is only necessary to place the butterfly spring in the area to be measured, and it can be replaced at any time when detecting multiple butterfly springs to realize automatic judgment and detection. , easy to operate, simple, high detection efficiency;

3. The present invention adopts the method of telecentric lens and backlight measurement, and adopts the improved image contour search technology, so that the boundary detection of the butterfly spring in the image is more accurate, and the measurement accuracy is improved;

4. In the present invention, when the height is measured, the marker points can be placed movably, and the measurement plane can be changed in position, so that the measurement plane is not restricted and the measurement is more flexible;

5. In the present invention, the method of similar distinction is adopted to distinguish the marker point and the target butterfly spring, which ensures that the butterfly springs within the field of view of the telecentric lens can be detected, and the stability of the measurement system is guaranteed;

6. In the present invention, the machine vision measurement method adopted, through the non-contact measurement method, reduces the damage in the manual measurement process of the butterfly spring, especially the butterfly spring with smaller specifications;

7. The present invention develops an easy-to-operate software interface, which displays the field of view of the camera, the inspection size, and the mark of whether it is qualified or not in real time, which is suitable for the use of ordinary inspection workers. Sorting provides supporting data;

8. The present invention has the characteristics of high precision, high stability and high efficiency for the three-dimensional dimension detection of the butterfly spring. In addition, the present invention can be extended to the three-dimensional dimension measurement of a class of thin-walled parts, apertures, and stamping parts.

Description of drawings

1 is a schematic structural diagram of a hardware part of an embodiment of the present invention;

2 is a schematic diagram of a software main interface according to an embodiment of the present invention;

3 is a schematic structural diagram of a butterfly spring according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of the reconstruction mark point and the center of the ellipse on the butterfly spring according to an embodiment of the present invention;

FIG. 5 is a processed image of a 31.5*16.3*2.45mm butterfly spring collected by a camera CL according to an embodiment of the present invention;

FIG. 6 is a diagram of a measurement result of a butterfly spring according to an embodiment of the present invention;

FIG. 7 is an error diagram of a measurement result of a butterfly spring according to an embodiment of the present invention.

In the picture: 1-1, corner connector; 1-2, base; 1-3, camera CR; 1-4, connector; 1-5, ordinary optical lens; 1-6, camera CT; 1-7, Mounting base; 1-8, telecentric optical lens; 1-9, camera CL; 1-10, butterfly spring; 1-11, marking point; 1-12, frosted glass; 1-13, backlight source; 1- 14. Pole; 1-15, Crossbar;

2-1 Functional area, 2-2 Setting standard area, 2-3 Prompt area, 2-4 Left camera CL display area, 2-5 Test result display area, 2-6 Camera CT display area, 2-7 Right camera CR display area;

Dimensions to be measured: outer diameter D of the butterfly spring, inner diameter d of the butterfly spring, and height H of the butterfly spring.

Detailed ways

The present invention will be further described in detail below with reference to accompanying drawings 1-7.

Example 1

A machine vision-based three-dimensional automatic measurement system for butterfly springs, as shown in Figure 1, includes a hardware part, the hardware part includes a detection device, the detection device includes a base 1-2, and the Two vertical rods 1-14 are symmetrically and fixedly installed on the upper side through the corner connecting piece 1-1, and a horizontal rod 1-15 is fixedly installed between the two vertical rods 1-14 near the upper end thereof, and the two vertical rods 1-14 are fixedly installed. The camera CL1-9 and the camera CR1-3 are respectively fixedly installed in the middle of the vertical rod 1-14 through the connecting seat 1-4, and the camera CT1- 6. Common optical lenses 1-5 are installed on the cameras CL1-9 and CR1-3, telecentric optical lenses 1-8 are installed on the camera CT1-6, and the bases 1-2 are A backlight light source 1-13 is installed at the bottom, and a frosted glass 1-12 is installed on the upper part of the base 1-2 above the backlight light source 1-13. The upper center of the frosted glass 1-12 is placed to be measured. The butterfly spring 1-10, and the upper part of the ground glass 1-12 is located on the outer periphery of the butterfly spring 1-10 with a number of marking points 1-11.

Preferably, as shown in FIG. 4 and FIG. 5 , the marking points 1-11 are common circular marking points with outer black and inner circle.

By adopting the above technical solution, the common outer black inner circle circular marking point is easy to identify, which is convenient to improve the measurement efficiency.

Preferably, as shown in FIG. 2, it also includes a software part, which includes CT, CL, CR camera display window and CT, CL, CR camera calibration interface and display window for measuring size, and standard setting of butterfly spring size. Whether the fixed window and butterfly spring size are qualified windows;

The CT, CL, CR camera display windows are used for real-time display of images collected by the cameras;

The CT, CL, and CR camera calibration interfaces are used for camera calibration to prepare for the size detection of the butterfly spring;

The measurement size display window is used for real-time display of measurement results;

The described butterfly spring size standard setting window is used to interactively input standard size and allowable error;

The whether the butterfly spring is qualified or not is used to display whether the size of the detected butterfly spring is qualified.

Preferably, the measurement results include: the inner diameter, outer diameter, and height of the butterfly spring, the detection error of inner diameter, outer diameter, and height, whether the inner diameter, outer diameter, and height meet the standard size and tolerance requirements, if not. The butterfly spring is detected, and the detection results are all 0.

The present invention also proposes a machine vision-based automatic measuring method for the three-dimensional size of the butterfly spring, comprising the following steps:

The first step is to determine that the grayscale information of each frame of image and the image of the previous frame are within the threshold range and enter the detection program;

The second step is to use the telecentric lens to detect the inner and outer diameters of the butterfly spring, and obtain the contour information of the butterfly spring to be detected through real-time image processing;

The third step is to use the automatic measurement system to measure the three-dimensional size of the butterfly spring, distinguish the mark point and the butterfly spring, reproject the mark point to the world coordinate system for plane fitting, and reproject the center of the ellipse on the butterfly spring to the world coordinate. Calculate the height from the center of the upper ellipse to the fitted plane.

Example 2

According to the machine vision-based three-dimensional automatic measurement system for butterfly springs built in Example 1, run the machine vision-based three-dimensional automatic measurement system software for butterfly springs on the PC, and connect the camera CL1- 9. CR1-3, CT1-6, perform system calibration, the calibration parameters are stored in the PC, place any butterfly spring on the frosted glass to be tested, and select the "Start Test" button in the 2-1 part of the functional area to enter the test process , the measurement results are displayed in real-time in the detection result display area 2-5, and the inspection personnel can sort the qualified butterfly springs by observing the detection results or the information in the prompt bar.

Example 3

According to the machine vision-based automatic three-dimensional size measurement method of a butterfly spring according to Embodiment 2, the method is: the camera works under a certain frame rate, and each time the camera CT collects a frame of image, it is compared with the previous frame image, and the grayscale is used. The information represents the image attributes of each frame, and the grayscale information of the front and rear images automatically enters the detection program within a certain threshold range.

Example 4

According to the machine vision-based three-dimensional automatic measuring system of the butterfly spring and the measuring method thereof according to the embodiment 1-3, the method is divided into a measuring method for the inner and outer diameters of the butterfly spring (D, d) and the height (H) of the butterfly spring Measurement methods.

(1) The method for measuring inner and outer diameters includes the three steps of telecentric lens camera calibration, image processing, and size calculation;

The camera calibration method is as follows: making a concentric annular calibration plate with a standard size, the actual size of the diameter D ₀ ={D ₀₁ , D ₀₂ , D ₀₃ }, and the annular calibration plate collected in the cameras CT1-6. On the image, find the size D ₀ '={D ₀₁ ', D ₀₂ ', D ₀₃ '} of the diameter corresponding to the ring on the image, then the calibration magnification K is defined as

所以拟合椭圆共有5个参数集，构造的属性矩阵为：P＝(p_ij)_m×5＝[N,W,S]，定义椭圆属性梯度矩阵

其中i＝1,2,...,5,j＝2,3,...,m，则

所对应的行以及相邻的行即为蝶形弹簧在图像上的椭圆，再通过图像位置关系，即可以区分出标记点和目标椭圆；The image processing method includes: image filtering, edge detection, ellipse fitting, and calculating the long axis of the ellipse, the image filtering adopts a Gaussian filter operator, the edge detection adopts the Canny operator, and the ellipse fitting Using the least squares method, since the marking point and the butterfly spring are placed on the upper surface of the frosted glass at the same time, if the butterfly spring and the marking point are in the field of view of the camera at the same time, it is necessary to distinguish the marking point and the butterfly spring. And the method of the butterfly spring is: the attribute table of the fitted circle obtained by fitting the ellipse is p _j ={n _j ,w _j ,S _j },j=1,2,...,m, where m is The number of fitted ellipses, n _j , w _j are the inner ellipse parameters of the jth fitted ellipse, namely the length of the major axis and the minor axis, denoted as N={n _1j ,n _2j }, W={w _1j ,w _2j }, S _j fits the area property of the ellipse, namely

Therefore, the fitting ellipse has a total of 5 parameter sets, and the constructed attribute matrix is: P=(p _ij ) _m×5 =[N,W,S], defining the ellipse attribute gradient matrix

where i=1,2,...,5,j=2,3,...,m, then

The corresponding row and the adjacent row are the ellipse of the butterfly spring on the image, and then through the image position relationship, the marker point and the target ellipse can be distinguished;

After the target ellipse is found, the major axis lengths _{Di img} and d _img of the inner and outer ellipses on the image can be obtained, and then the actual length outer diameter D=KD _img and inner diameter d=Kd _img are calculated.

(2) the described height measurement method comprises the fitting of the measurement plane, the search of the center of the ellipse on the butterfly spring, and the distance calculation;

The fitting method of the measurement plane is as follows: the left and right cameras CL1-9 and CR1-3 simultaneously collect images of the measurement plane, find the marked circles of the left and right images according to the marked point search method, and use the limit constraint method to match the marked points. , using the projection matrices calibrated by the left and right cameras CL1-9 and CR1-3 and the parallax of the left and right cameras, reproject the marked points into the world coordinate system to obtain the three-dimensional coordinates of m marked points in space, and use the least squares method to convert these three-dimensional The coordinate point fitting plane is the representation of the measurement plane in the world coordinate system, such as the plane π in the world coordinate system in Figure 4: Ax+By+Cz+D=0;

The search method in the ellipse of the left and right butterfly springs is as follows: first, perform image filtering to remove noise, especially the noise in the gray part, obtain the region of interest of the butterfly spring, and divide the gray value of the image into three levels: black, gray, and white. , and obtain interest for clustering through machine learning. The specific method is that the gray value of any pixel on the image is f(u, v), and the initial threshold gray value of black, gray and white is given by experiments. B ₀ , G ₀ , W ₀ , the classification idea is to calculate the distance between the gray value of any point on the image and the gray threshold value, and the distance from the pixel belonging to the black area to B ₀ is the smallest,

That is, it has the property |f(u,v)-B ₀ |=min{|f(u,v)-B ₀ |,|f(u,v)-G ₀ |,|f(u,v)-W ₀ |} All pixels belonging to these three categories will be obtained after one classification. In each category, the average pixel gray value is taken to replace the initial three-level threshold, and clustering is performed again until the gray threshold remains stable, that is, The algorithm converges. After clustering, the gray level in the butterfly spring image is initially obtained. The next step is to perform contour search. Again, the idea of clustering is adopted. According to the gray value gradient of the black area transitioning to the gray area, the gray area The grayscale gradient values that transition to the white area and the grayscale gradient values that transition from the black area to the white area are classified by screening the pixels in the transition from the black area to the white area and the pixels transitioning from the black area to the gray area. The boundary is the outline of the upper ellipse, and the boundary formed by the pixels from the gray area to the white area and the pixels from the white area to the white area is the inner circle, and then the center point of the upper ellipse is obtained according to the ellipse fitting method. Image coordinates (u _L , v _L ), and the centers C ₁ , C ₂ , C ₃ , and C ₄ of the four ellipses of the butterfly spring can be obtained at the same time. Perform the same processing on the image collected by the right camera to obtain the coordinates (u _R , v _R ) of the center point of the ellipse on the butterfly spring of the right camera CR. According to the principle of left and right stereo vision, the left and right are matched and then reprojected to the world coordinate system , the world coordinates (x ₀ , y ₀ , z ₀ ) of the center of the circle on the butterfly spring under test are obtained.

The height of the butterfly spring is converted into the distance from the midpoint of the world coordinate system (x ₀ , y ₀ , z ₀ ) to the plane π. Then the measured height of the butterfly spring is:

Example 5

According to Embodiments 1-4, in order to verify whether the detected size of the butterfly spring matches the actual size, a vernier caliper is used for measurement verification. The test results and test errors are shown in Figures 6-7. For the standard size 31.5*16.3*2.45mm butterfly spring, the vernier caliper is used to measure the butterfly spring, the size is 31.37*16.57*2.8mm, and the error between the measurement result of the present invention and the vernier caliper measurement result is very small. At the same time, it can be seen that the butterfly spring is a non-conforming product.

Example 6

According to a machine vision-based three-dimensional automatic measuring system for butterfly springs and a measuring method thereof described in Embodiments 1-5, the functions of each part are as follows:

1. Telecentric lens 1-8: The telecentric lens has an ultra-wide depth of field. Within the depth of field, the size of the camera image remains unchanged, eliminating the perspective error caused by the "near big and far small" imaging of ordinary cameras;

2. Left and right cameras: The left and right cameras can obtain the depth information of the image, which can realize the reconstruction of the marked point and the center of the ellipse on the butterfly spring, so as to obtain the three-dimensional size;

3. Backlight light source: ensure the clear characteristics of the image, make the outline more prominent, and reduce the interference of ambient light on image acquisition, such as reflection;

4. The base, the pole and the cross bar: connect the three cameras, the backlight light source, the corner connector, and connect and support the normal installation and operation of the entire hardware part;

5. Software part: Provide a friendly user interface, which can easily and conveniently perform system calibration and display interface of test results, and display the current test results and the message of whether the butterfly spring is qualified.

Working principle: The machine vision-based three-dimensional automatic measurement system for butterfly spring and its measurement method, firstly connect the camera, adjust the camera pose and parameters, and calibrate the camera. Then turn on the backlight light source, and place the butterfly spring to be tested on the upper surface of the frosted glass. The three-dimensional size measurement is carried out through the three-dimensional size detection and calculation method of the butterfly spring. The measurement results are displayed in real time through the described software interface. The backlit light source makes the outline of the collected image more prominent, improving the image quality and detection accuracy.

The parts not involved in the present invention are the same as the prior art or can be implemented by using the prior art.

This specific embodiment is only an explanation of the present invention, and it does not limit the present invention. Those skilled in the art can make modifications without creative contribution to the present embodiment as required after reading this specification, but as long as the rights of the present invention are used All claims are protected by patent law.

Claims (4)

1. A method for automatically measuring the three-dimensional size of a belleville spring based on machine vision comprises a hardware part, the hardware part comprises a detection device, the detection device comprises a base (1-2), two upright posts (1-14) are symmetrically and fixedly installed on one side of the upper portion of the base (1-2) through corner connecting pieces (1-1), a cross rod (1-15) is fixedly installed between the two upright posts (1-14) and close to the upper end of the two upright posts, the middle portions of the two upright posts (1-14) are respectively and fixedly provided with cameras CL (1-9) and CR (1-3) through connecting seats (1-4), and the bottoms of the middle portions of the cross rods (1-15) are fixedly provided with cameras CT (1-6) through assembling seats (1-7), common optical lenses (1-5) are mounted on the camera CL (1-9) and the camera CR (1-3), telecentric optical lenses (1-8) are mounted on the camera CT (1-6), a backlight light source (1-13) is mounted at the bottom of the base (1-2), frosted glass (1-12) is mounted on the upper portion of the base (1-2) above the backlight light source (1-13), a disc spring (1-10) to be measured is placed at the central position of the upper portion of the frosted glass (1-12), and a plurality of mark points (1-11) are arranged on the periphery of the disc spring (1-10) on the upper portion of the frosted glass (1-12);

the method is characterized in that: the method comprises the following steps:

firstly, comparing each frame of image collected by a camera CT with the previous frame of image, and automatically entering a detection program for the gray information of the previous and next images within a certain threshold range;

secondly, detecting the inner and outer diameter sizes of the belleville spring by using a telecentric lens, and acquiring contour information to be detected of the belleville spring through real-time image processing;

and thirdly, measuring the three-dimensional size of the belleville spring by using an automatic measuring system, distinguishing mark points and the belleville spring, projecting the mark points to a world coordinate system again for plane fitting, obtaining the coordinates of the elliptic central point on the belleville spring of the camera CL and the elliptic central point on the belleville spring of the camera CR according to an elliptic fitting method, performing left-right matching and then projecting the marks to the world coordinate system again according to the principle of left-right stereoscopic vision to obtain the elliptic center on the belleville spring, and calculating the height by calculating the distance from the elliptic center to a fitting plane.

2. The method for automatically measuring the three-dimensional size of the belleville spring based on the machine vision as claimed in claim 1, wherein: the mark points (1-11) are common outer black inner circle round mark points.

3. The method for automatically measuring the three-dimensional size of the belleville spring based on the machine vision as claimed in claim 1, wherein: the device also comprises a software part, wherein the software part comprises a CT, CL and CR camera display window, a CT, CL and CR camera calibration interface, a display window for measuring the size, a belleville spring size standard setting window and a belleville spring size qualified window;

the CT, CL and CR camera display windows are used for displaying images acquired by the cameras in real time;

the CT, CL and CR camera calibration interfaces are used for calibrating cameras and preparing for size detection of the belleville springs;

the measurement size display window is used for displaying the measurement result in real time;

the size standard setting window of the belleville spring is used for interactively inputting the standard size and the allowable error;

and the qualified or unqualified belleville spring window is used for displaying whether the size of the detected belleville spring is qualified or not.

4. The machine vision-based method for automatically measuring the three-dimensional size of the belleville spring according to claim 3, wherein the method comprises the following steps: the measurement result comprises: the inner diameter, the outer diameter and the height of the belleville spring, detection errors of the inner diameter, the outer diameter and the height, and whether the inner diameter, the outer diameter and the height meet the standard size and tolerance requirements or not, if the belleville spring is not detected, the detection result is 0.

Images