Novel human-computer interaction video extensometer measuring system and method thereof

Abstract

The invention discloses a novel man-machine interaction video extensometer measurement system and a method thereof, and the method comprises the steps: installing a test workpiece on a universal testing machine, and installing and debugging an image collection device; starting the image acquisition equipment to acquire an initial reference image of the test workpiece, then starting the universal testing machine to load the test workpiece, and continuously acquiring a deformation image of the test workpiece in the loading process until the stretch rate of the test workpiece reaches a preset threshold value of the resistance strain gauge; and transmitting the acquired image to a video extensometer measurement system, calculating the transverse strain, the longitudinal strain and the Poisson's ratio of the test workpiece according to the displacement of the deformation point in the deformation image relative to the reference standard point in the initial reference image, and displaying the transverse strain, the longitudinal strain and the Poisson's ratio through a system interface. The device and the method can not only adapt to measurement of deformation and strain of objects of various sizes, but also reduce errors caused by out-of-plane displacement of the objects. Software and hardware are combined, and accurate and efficient measurement of the video extensometer is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of video extensometer measurement, and in particular to a novel human-computer interactive video extensometer measurement system and method thereof. Background Art

[0002] Among the commonly used material mechanical property testing methods, there are two types: electrical measurement and optical measurement. The electrical measurement method has high measurement accuracy and is relatively simple to operate, and is widely used in various engineering fields. However, the electrical measurement method is a contact measurement method, and the sensor used for measurement needs to be in direct contact with the object being measured. Compared with the electrical measurement method, the optical measurement method can achieve non-contact measurement. The traditional laser interferometer method has the characteristics of high accuracy and full-field measurement, but it is easily affected by factors such as temperature field radiation, light intensity, and airflow during the measurement process.

[0003] Using the principles of electrical and optical measurement, extensometers are widely used in tensile and fatigue tests of materials, and have become one of the main tools for experimental mechanical measurements. The research on the hardware and software of video extensometers has been continuously developed and improved. In terms of hardware, the CCD / CMOS sensor technology has developed rapidly, and the hardware requirements for the development of video extensometers have been met in terms of accuracy and sensor imaging speed; many more effective algorithms and solutions have also been proposed in terms of software, but due to the influence of efficiency and accuracy, the research on domestic video extensometers is still in the primary experimental stage, and its stability, measurement speed and measurement range still cannot meet the existing experimental and engineering requirements.

[0004] To this end, we designed a new human-computer interactive video extensometer measurement system and method to solve the above problems. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art that the research on domestic video extensometers is still in the primary experimental stage due to the influence of efficiency and accuracy, and its stability, measurement speed and measurement range still cannot meet the existing experimental and engineering needs. A new type of human-computer interactive video extensometer measurement system and method are proposed.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: A novel human-computer interactive video extensometer measurement method includes the following steps: Equipment debugging preparation: install the test workpiece on the universal testing machine, install and debug the image acquisition equipment, paste the resistance strain gauge on the back of the test workpiece on the universal testing machine, and connect it to the resistance strain gauge, and the resistance strain gauge is connected to the video extensometer measurement system; Experimental test: After starting the image acquisition device to acquire the initial reference image of the test workpiece, start the universal testing machine to load the test workpiece, set the shooting frame rate of the image acquisition device, and continuously acquire the deformation image of the test workpiece during the loading process until the stretching rate of the test workpiece reaches the preset threshold of the resistance strain gauge; Calculation and result output: The collected initial reference image and deformation image are transmitted to the video extensometer measurement system, and the deformation image is numbered. The transverse strain, longitudinal strain and Poisson's ratio of the test workpiece are calculated according to the displacement of the deformation point in the deformation image relative to the reference standard point in the initial reference image, and displayed through the program interface of the video extensometer measurement system.

[0007] As a further preferred solution of the present invention, before the test workpiece is installed on the universal testing machine, white paint is first sprayed on the surface of the test workpiece as a base color, and then black paint dots are sprayed to form a random speckle pattern, and a marking line is engraved on the test workpiece; the test workpiece is installed on the universal testing machine, the fixture is first fixed to the upper and lower sides of the universal testing machine, and then the test workpiece is placed in the fixture and clamped.

[0008] As a further preferred solution of the present invention, when the resistance strain gauge is pasted on the back of the test workpiece, the same test workpiece is prepared and the resistance strain gauge is pasted on the same position without loading, and is connected to the resistance strain gauge as a temperature compensation sheet; Installing and debugging image acquisition equipment includes: A telecentric lens is installed at the front end of the camera, a filter is arranged in front of the telecentric lens, a monochromatic annular light source is installed in front of the filter, a tripod is placed in front of the universal testing machine, a slide rail is installed on the upper part of the tripod, the slide rail and the workpiece to be tested are on the same vertical plane, the camera and the monochromatic annular light source are both slidably arranged on the slide rail, and the monochromatic annular light source is located between the camera and the universal testing machine; Install the tripod for fixing the camera, adjust the position of the camera and the monochromatic ring light source on the slide rail, and adjust the input power of the monochromatic ring light source to adjust the light intensity of the light source.

[0009] As a further preferred solution of the present invention, the camera adopts a CCD camera with a target surface size of 2 / 3 inches; the shooting frame rate of the image acquisition device is set to capture an image every 2 seconds, and the preset threshold of the resistance strain gauge is set to the resistance strain gauge elongation ≤5%.

[0010] As a further preferred solution of the present invention, a marked point is selected in the initial reference image as a reference standard point, a deformation image with a corresponding number is selected when the universal testing machine is in elastic deformation during the loading phase, and a deformation point corresponding to the reference standard point in the deformation image is determined; The FAST feature detection algorithm is used to detect the reference standard points selected in the initial reference image, and the SIFT feature description algorithm is used to describe the feature points detected by the FAST algorithm. The process is as follows: (1) Direction matching of feature points: Input the initial reference image and the deformed image and perform grayscale processing; Select pixels in the initial reference image and the deformed image , compare its brightness difference with surrounding pixels; If the brightness difference of N pixels exceeds the threshold, the judgment is the feature point; Output the feature points in the initial reference image and the deformed image and ; The FAST feature detection algorithm is used to find the position of the reference standard point in the initial reference image, and the feature point is described by the feature description algorithm provided by the SIFT feature description algorithm: Input the feature point position coordinates; The SIFT feature description algorithm uses the feature points of the DoG pyramid obtained in the FAST feature detection algorithm to collect the Gaussian pyramid image where it is located. Gradient and directional feature distribution of pixels in the neighborhood window, gradient modulus and direction The calculation of is as follows: ; ; In the formula, ( ) represents the coordinates of the reference standard points in the initial reference image, Indicates the scale space value of the feature point, sampled according to the scale Principle: Sampling is performed with a certain neighborhood radius as the sampling window; (2) Feature description of feature points: For the feature points found by the FAST feature detection algorithm, two important information are given to the feature points: position and direction, and a description vector is established for each feature point; Descriptor set generation: Divide the neighborhood into × Sub-region, feature points are described in × Calculation within the window Gradient learning in each direction, obtaining ( × × ) vector representation of dimensions; Output the descriptor set for each feature point and ; Establish the gradient modulus and direction of the feature points and use the FLANN algorithm for matching: Input the descriptor set of feature points in the initial reference image and the deformed image, Use the FLANN library to generate a set of descriptors for feature points in the initial reference image Build a KD tree index, Approximate nearest neighbor search: a set of descriptors for each feature point in the deformed image , search for the most similar k neighbors in the KD tree; Filter the matching pairs, take a feature of the initial reference image, and search for the two features closest to it in the deformed image. The distances are recorded as and , filter false matches by nearest neighbor distance ratio (NNDR): ; If the above conditions are met, the matching feature points are retained and .

[0011] As a further preferred solution of the present invention, the process of calculating the transverse strain, longitudinal strain and Poisson's ratio of the test workpiece according to the displacement of the deformation point relative to the reference standard point is as follows: Select the length D to be measured in the X direction of the test workpiece surface. Direction Select the length L to be measured. When the test workpiece is loaded, the length D in the X direction and The length L in the direction is deformed, resulting in lateral displacement and longitudinal displacement , the transverse strain and longitudinal strain are shown as follows: Transverse strain: ; Longitudinal strain: ; Poisson's ratio: ; In the above formula, is the lateral strain of the test workpiece when the load is applied, is the longitudinal strain of the test workpiece when the load is applied, represents the Poisson's ratio of the test workpiece material, ( , ) is the coordinate of the reference standard point, ( , ) are the coordinates of the deformation point; The program interface of the video extensometer measurement system displays the Poisson's ratio of the test workpiece in the deformation image under the corresponding number, as well as the displacement-time curve and stress-time curve.

[0012] A video extensometer measurement system for a novel human-computer interactive video extensometer measurement method, the video extensometer measurement system comprising: The camera is used to collect the image of the test workpiece. A telecentric lens is installed on the camera. A filter is arranged in front of the telecentric lens. A monochromatic annular light source is installed in front of the filter. The slide rail and the test workpiece to be tested are on the same vertical plane. The camera and the monochromatic annular light source are both slidably arranged on the slide rail. The monochromatic annular light source is located between the camera and the universal testing machine. The slide rail is installed on the upper part of the tripod. Computer, install and run the video extensometer measurement system program, establish data transmission connection with the camera and the universal testing machine, receive the camera captured images and the universal testing machine loading data, calculate the transverse strain, longitudinal strain and Poisson's ratio of the test workpiece according to the captured images, and display them in real time through the interface; The universal testing machine applies a loading force to the test workpiece to cause the test workpiece to deform and obtain the mechanical performance parameters of the test workpiece.

[0013] As a further preferred solution of the present invention, the camera adopts a CCD camera with a target surface size of 2 / 3 inches; the monochromatic annular light source adopts a monochromatic LED light source with a wavelength of 465-470nm, and the irradiation angle of the monochromatic annular light source is 60° with the horizontal ground; the filter adopts a filter with a central wavelength of 465nm, a half-width of 10nm, and a transmittance of 80%.

[0014] As a further preferred solution of the present invention, the telecentric lens is a bilateral telecentric lens, and the bilateral telecentric lens adopts a small-field-of-view bilateral telecentric lens or a large-field-of-view bilateral telecentric lens. The magnification of the small-field-of-view bilateral telecentric lens is 0.22, and the field of view range is 40×30mm; the magnification of the large-field-of-view bilateral telecentric lens is 0.0716, and the field of view range is 122.9×92.2mm.

[0015] As a further preferred solution of the present invention, the program interface of the video extensometer measurement system installed and run on the computer includes a control area and a parameter area, the control area is a function control area, and the parameter area is a parameter display area.

[0016] The video extensometer measurement system proposed in the present invention can adapt to the displacement and strain measurement under various working conditions, and a stable illumination scheme is designed. The influence of ambient light can be ignored by the cooperation of a monochromatic light source and a filter. A CCD camera with a 2 / 3 target surface is used as an imaging camera, and a bilateral telecentric lens with different field of view ranges is selected. It can not only adapt to the measurement of deformation and strain of objects of various sizes, but also reduce the error caused by the off-plane displacement of the object. In addition, the present invention designs a human-computer interaction interface, which can select virtual marking points, output the displacement, strain and Poisson's ratio of the sample, and generate corresponding displacement time diagrams and strain time diagrams. The combination of software and hardware realizes accurate and efficient measurement of the video extensometer, breaking through the limitations of traditional extensometers. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of the composition of a novel human-computer interactive video extensometer measurement system proposed by the present invention; Figure 2 A schematic diagram of the camera composition of a novel human-computer interactive video extensometer measurement system proposed by the present invention; Figure 3 A schematic diagram of a human-machine interaction interface of a novel human-machine interactive video extensometer measurement system proposed by the present invention; Figure 4 This is a flow chart of a novel human-computer interactive video extensometer measurement method proposed by the present invention; Figure 5 A schematic diagram of a histogram generated by feature points during the gradient calculation process of feature points in this embodiment; Figure 6 It is a schematic diagram of a comparison curve between the actual strain value obtained by the strain gauge in an embodiment of the present invention and the strain value measured by the video extensometer.

[0018] In the figure: 1. camera; 2. telecentric lens; 3. monochromatic annular light source; 4. filter; 5. tripod; 6. slide rail; 7. slider; 8. computer; 9. universal testing machine; 91. upper fixture; 92. lower fixture; 10. test workpiece. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0020] The working principle of the video extensometer is different from that of the ordinary mechanical extensometer, but both are designed to track the deformation of the specimen, so that the displacement and strain of the specimen can be measured. The same video extensometer also has different measurement methods, such as based on the markings, punctuation points on the specimen or the characteristic points on the test piece, but their measurement principles are the same, which is to simulate the contact points of the mechanical extensometer by tracking the marking points.

[0021] The video extensometer measurement system is mainly composed of hardware and software. At present, the hardware and software of the video extensometer measurement system still have certain limitations. There are mainly the following aspects: Light source: In order to efficiently track the object under test, the video extensometer usually needs to shoot at a high frame rate. However, the ambient light and ordinary indoor light sources cannot meet the exposure requirements for high frame rate shooting. The quality of the image is also directly related to the light source. Poor lighting quality will increase the noise in the image, reduce the contrast, weaken the confusion, and fail to clearly capture the characteristics of the image. Therefore, a good light source system is essential to obtain stable imaging quality.

[0022] Sensor: For image acquisition and system imaging, there are two conventional sensor types to choose from: Charge Coupled Device (CCD) and Complementary Metal Oxide Semiconductor (CMOS). CCD sensors are composed of photosensitive elements arranged in a plane array or a linear array. After a light source shines into the photosensitive element, the photons are converted into electrons. These electrons are transmitted to the video capture card through a time-series register to become a digital image signal. CMOS sensors have a high degree of integration, and the distance between the components is close, resulting in serious interference between them, resulting in high imaging noise.

[0023] Lens: In order to adapt to the measurement of object deformation under different working conditions, different lenses need to be selected, which is similar to the focusing function of human eyes. Several parameters need to be considered. Lens focal length: that is, the light enters the lens in parallel, is refracted by the lens, and gathers into a light spot on the other side. The focal length affects the distance of the object being photographed; Lens working distance: The working distance of the lens is affected by the focal length. Once this distance is exceeded, a clear image cannot be captured; Field of view: The field of view is the coverage range of the lens. The objects to be measured are of different sizes. Only lenses with different fields of view need to be replaced to measure objects of different sizes; Lens depth of field: The imaging of the lens for an object at a certain working distance. Whether the object to be measured can form a clear image on the sensor is determined by the axial range of the depth of field. A clear image cannot be obtained outside the depth of field range; Aperture: The aperture is used to control the amount of light entering the lens. The larger the aperture, the more light enters. When measuring, you need to choose a suitable aperture. If the aperture is not suitable, overexposure or dark images may occur.

[0024] In summary, we have developed an adaptive multi-condition video extensometer suitable for various engineering practical application measurements. Then we built a hardware system for the video extensometer, which consists of a CCD camera, a bilateral telecentric lens, a suitable light source, a filter, etc., and assembled it into a video extensometer that can measure various actual working conditions. For specimens of different ranges, we only need to replace different lenses. For actual application scenarios with large influence of ambient light such as high temperature, we optimize the light source and imaging, eliminate the influence of ambient light and off-plane displacement to achieve high-precision measurement, and can measure Poisson's ratio.

[0025] This embodiment proposes a novel human-computer interactive video extensometer measurement system, which mainly includes: a camera 1 (image acquisition device) for collecting image data, a computer 8 for processing data, and a universal testing machine 9 for loading test materials. Specifically, the composition of the video extensometer measurement system is as follows: Figure 1 and Figure 2 As shown: a telecentric lens 2 is installed on a camera 1, a filter 4 is arranged in front of the telecentric lens 2, a monochromatic annular light source 3 is installed in front of the filter 4, a slide rail 6 and a test workpiece 10 to be tested are in the same vertical plane, the camera 1 and the monochromatic annular light source 3 are both slidably arranged on the slide rail 6, and the monochromatic annular light source 3 is located between the camera 1 and the universal testing machine 9, the slide rail 6 is horizontally installed on the upper part of the tripod 5, and the telecentric lens 2 in the present invention is a bilateral telecentric lens.

[0026] The resolution of camera 1 directly affects the clarity of the image. The higher the resolution of the sensor, the clearer the image it captures, which has a great impact on the high-precision tracking of the mark point. The larger the sensor size, the more photosensitive elements can be installed in the sensor to achieve higher clarity. In the case where the number of pixels remains unchanged, the larger the sensor size, the stronger the control of noise will be, because if the photosensitive elements are too close, they will interfere with each other and form noise on the image. The signal-to-noise ratio of the sensor, the typical value of the signal-to-noise ratio is 46dB, the higher the signal-to-noise ratio, the smaller the noise. For example, a signal-to-noise ratio of 50dB will have a certain amount of noise. If it reaches 60dB, the image quality will be better and there will be no noise. Based on the above limiting factors, the present invention comprehensively considers and selects a CCD camera with a target surface size of 2 / 3 inches for camera 1.

[0027] The computer 8 is installed with and runs the video extensometer measurement system program, establishes a data transmission connection with the camera 1 and the universal testing machine 9 via a data line, receives images collected by the camera 1 and loading data of the universal testing machine 9, calculates the lateral strain, longitudinal strain and Poisson's ratio of the test workpiece 10 according to the collected images, and displays them in real time through the interface.

[0028] The universal testing machine 9 applies a loading force to the test workpiece 10 to deform the test workpiece 10 and obtain the mechanical performance parameters of the test workpiece 10. The universal testing machine 9 includes an upper loader and a lower loader, and a fixture for fixing the test workpiece 10. The fixture consists of an upper fixture 91 and a lower fixture 92. Both ends of the test workpiece 10 are fixed on the upper fixture 91 and the lower fixture 92.

[0029] In this embodiment, a stable light source system is designed, and a monochromatic ring light source 3 with a wavelength of 465-470nm is used. Because it only accounts for 1.25% of the entire visible light wavelength, light of this wavelength has a good reflection effect on most test pieces, providing high-quality light for the sensor. The monochromatic ring light source 3 can use an LED light source to replace an ordinary white light source.

[0030] By accurately matching the emission wavelength of the light source, the filter 4 can enhance the light intensity in the target area, making the speckle pattern clearer, improving the image contrast, and facilitating the subsequent acquisition and calculation of deformation displacement and strain.

[0031] To eliminate the influence of stray light, in a complex lighting environment (such as thermal radiation in a high-temperature scene) or a multi-light source system, the filter 4 can filter out spontaneous radiation or reflected light of non-target wavelengths to ensure that the measurement is only for the light source signal of the monochromatic ring light source 3.

[0032] Improve measurement stability, reduce the impact of lighting condition fluctuations (such as ambient light changes or light source drift) on imaging, and ensure the consistency of digital image correlation (DIC) analysis, especially in long-term or dynamic experiments. The use of an annular light source can save the installation space of the video extensometer, does not require more brackets to support the light source, and solves the shadow problem of diagonal illumination. The annular light source of the monochromatic annular light source 3 can achieve high-brightness and high-uniformity diffuse lighting, that is, there is an angle between the straight light of the lamp bead and the ground. The preferred solution is that the angle between the illumination angle of the monochromatic annular light source 3 and the horizontal ground is 60°; this can reduce the shadow and highlight the edge features of the target. In order to make the system compact, the present invention installs the light source in a parallel position of the camera 1, and adopts a high-angle illumination method to reduce the shadow and highlight the edge features of the tested workpiece 10. The ambient light changes that may occur during the test can be ignored, and stable and uniform light is provided to the sensor.

[0033] In addition, a filter 4 (bandpass filter) is added in front of the telecentric lens 2. The filter 4 allows light in a specific wavelength range to pass through, and can only pass light with a wavelength of 450-480nm, while blocking light of other wavelengths, effectively suppressing ambient light interference. The filter 4 allows light that matches the system light source (such as an LED light source with a specific wavelength) to pass through, while blocking ambient light of other wavelengths (such as indoor lights or natural light), thereby reducing background noise and improving the image signal-to-noise ratio. In particular, it solves the problem that traditional video extensometers cannot work at high temperatures. Because there is a large amount of infrared light around high-temperature objects, the noise of the image is greatly increased, and the object features cannot be highlighted. The filter 4 only allows light of wavelengths within the bandwidth to enter, which effectively isolates the influence of the self-luminescence of the high-temperature furnace on the sensor imaging. The specific parameters of the filter 4 are shown in Table 1.

[0034] Table 1 Filter parameters

[0035] In order to obtain a high-fidelity digital image of the test workpiece 10 and lay the foundation for subsequent high-precision strain measurement, the present invention has established a hardware imaging system with high stability and near perfection, combining active imaging with a high-quality telecentric lens. (Active imaging refers to the technology of actively emitting a specific light source (such as laser, LED structured light, etc.) to illuminate the test workpiece 10 and using reflected or transmitted light for imaging. Compared with passive imaging (which relies on ambient light), active imaging can accurately control the lighting conditions.

[0036] Telecentric lenses are not affected by slight changes in object distance and imaging distance, and there is no need to consider the influence of deformation perpendicular to the camera caused by edgeless deformation. Therefore, compared with the use of traditional lenses, imaging systems equipped with telecentric lenses perform better in minimizing virtual displacement and strain caused by inevitable out-of-plane motion (translation or rotation) of the specimen surface. In order to meet the measurement range of various ranges, the present invention selects two types of lenses to cope with a variety of different measurement needs, namely a small field of view bilateral telecentric lens and a large field of view bilateral telecentric lens. The two lenses cooperate with camera 1 to achieve micron-level measurement accuracy. The specific parameters of the two lenses are shown in Table 2 below.

[0037] Table 2 Lens parameters

[0038] As the software control part, the program interface of the video extensometer measurement system installed and run on computer 8 includes a control area and a parameter area. The control area is a function control area, and the parameter area is a parameter display area. In addition, the interface also displays the collected image and the output image number and data curve. The display effect of the interface is as follows Figure 3 shown.

[0039] In this embodiment, based on the above novel human-computer interactive video extensometer measurement system and based on the digital image correlation (DIC) theory, a novel human-computer interactive video extensometer measurement method is further proposed.

[0040] The theory of digital image correlation (DIC) method is as follows: The surface of an object usually has natural or artificial random speckles with unique information, which appear as physical microstructures or optical speckle fields. They are used to study the shape changes of physical surfaces. The two-dimensional surface images of objects during deformation can be recorded by digital cameras or high-speed cameras, and the speckle fields can be calculated and analyzed. The basic principle of the DIC method is to track the spots in the collected images, use the image before deformation as the initial reference image, and the image during deformation as the deformed image, perform a correlation matching search on the points of interest in the two images, and calculate the deformation field.

[0041] Assuming that camera 1 can obtain continuous two-dimensional images and contain the same total amount of grayscale, it can be considered that the difference between the images before and after deformation lies in the different spatial distribution of scattered speckles. Taking pixels as coordinates, a certain size of the area of ​​interest to be calculated is selected in the initial reference image as the reference sub-region (Reference image subset), and the same size of the area of ​​interest is selected in the deformed image as the deformed sub-region (Deformed image subset). The two sub-regions are matched for correlation, and the position change of the sub-region with the highest matching coefficient is the required displacement.

[0042] Define the reference sub-region center point as , the center point of the deformed sub-region is defined as , then: ; in, , Mark the points along , The displacement in the direction.

[0043] In actual deformation, the sub-area will not only undergo overall rigid body translation, but also undergo complex deformations such as rotation, tension, compression, and shear. Therefore, the above formula cannot simply represent the changes of each point in the sub-area. , Respectively as about , function, establish any point in the reference sub-area With deformation sub-area The corresponding relationship is: ; The above formula can be rewritten into different forms according to the deformation characteristics of the sub-area. When only rigid body translation exists: ; When there is rotation, tension, compression or shear deformation: for , The translation displacement in the direction, and for , The displacement component in the direction.

[0044] ; When more complex deformations occur: ; In the formula, is the first-order displacement gradient, is the second-order displacement gradient.

[0045] Therefore, refer to Figure 4 Based on the above principle, the novel human-computer interactive video extensometer measurement method proposed in this embodiment includes the following steps: Step 1, equipment debugging preparation: run the program of the video extensometer measurement system, install the test workpiece 10 on the universal testing machine 9, install and debug the image acquisition device, paste the resistance strain gauge on the back of the test workpiece 10 on the universal testing machine 9, and connect it to the resistance strain gauge, and the resistance strain gauge is connected to the video extensometer measurement system.

[0046] Speckle treatment of the test workpiece 10: Under normal circumstances, the natural texture and spots on the surface of the test piece are used as tracking points to calculate the two-dimensional displacement field of the test piece surface. Generally speaking, the surface of the test workpiece 10 should have a high-quality, high-contrast random speckle pattern as a carrier of deformation information to ensure the uniqueness and accuracy of the DIC method deformation measurement. However, there are some common defects in natural speckle patterns, such as low contrast and poor durability under large deformation, which may lead to information loss and produce low correlation and inaccurate strain measurement data. In order to prevent the above phenomenon from affecting the results. Before the test workpiece 10 of the present invention is installed on the universal testing machine 9, white paint is first sprayed on the surface of the test workpiece 10 as a base color, and then black paint dots are scattered to form a random speckle pattern with high contrast for displacement field calculation, and a marking line is engraved on the test workpiece 10; the test workpiece 10 is installed on the universal testing machine 9, and the clamps are first fixed on the upper and lower sides of the universal testing machine 9, and then the test workpiece 10 is clamped in the upper clamp 91 and the lower clamp 92.

[0047] When the resistance strain gauge is pasted on the back of the test workpiece 10, prepare the same test workpiece 10 and paste the resistance strain gauge at the same position without loading, and connect it to the resistance strain meter as a temperature compensation gauge; the preset threshold value of the resistance strain gauge is set to the resistance strain gauge elongation ≤5%.

[0048] The purpose of sticking the resistance strain gauge on the back of the test workpiece 10 is to form a control group with the video extensometer to verify the accuracy of the video extensometer measurement. The principle of the video extensometer is to measure the deformation (displacement) of the test piece under the tension or compression of the electronic universal testing machine. Through the principle of non-contact shooting, the image of the test piece at rest is first taken as the initial reference image. During the loading process, the test piece will deform, and the deformation image of the test piece is taken at intervals. All images are imported into the software in chronological order. Select an area in the initial reference image as the target area (the minimum unit of the selected area is 1 pixel). Assuming that the initial reference image target is divided into N units (pixels), the deformed image should also have N corresponding units in principle. Find the N units in the deformed image that correspond to the initial reference image. For example, the coordinates of a point in the initial reference image are , at time t, the deformed coordinates become , then the deformation of the coordinate can be calculated by subtracting the two coordinates to get the displacement, and by measuring the displacement, the strain can be calculated. The resistance strain gauge provided in the present invention is used as a control, and its purpose is to verify whether the video extensometer measurement meets the actual needs. During the specimen loading process, it is assumed that the central area of ​​the specimen is selected, and a video extensometer is used to measure the strain value of the central area on one side (the specimen is usually dumbbell-shaped), and a resistance strain gauge is pasted on the center of the other side of the specimen (the traditional contact method for measuring the strain of the specimen), then a comparison is made to see whether the strain gauges of the central area of ​​the specimen measured by the two methods at the same time during the loading process are consistent. Through the horizontal coordinate time and the vertical coordinate strain value, the error of the two methods in measuring strain can be clearly seen.

[0049] Installing and debugging image acquisition equipment includes: Open the tripod 5. The height of the tripod 5 is adjusted up and down according to the height of the test workpiece 10. The tripod 5 can be fixed on the ground by the hinge on the tripod 5. Adjust the level bubble to make the tripod 5 in a horizontal position. Install the slide rail 6 on the tripod 5. The slide rail 6 has a bolt hole. The slide rail 6 is fixed on the tripod 5 by screws and gaskets. The slide rail 6 has a slider 7 that can slide horizontally. The slider 7 has a clamp for clamping the camera 1 and the monochromatic ring light source 3.

[0050] A telecentric lens 2 is installed at the front end of the camera 1, a filter 4 is arranged in front of the telecentric lens 2, a monochromatic annular light source 3 with a wavelength of 465-470nm is installed in front of the filter 4, a tripod 5 is placed in front of the universal testing machine 9, a slide rail 6 is installed on the upper part of the tripod 5, the slide rail 6 and the test workpiece 10 to be tested are on the same vertical plane, the camera 1 and the monochromatic annular light source 3 are both slidably arranged on the slide rail 6, and the monochromatic annular light source 3 is located between the camera 1 and the universal testing machine 9. By adjusting the horizontal bubble, the camera 1 is kept on the horizontal plane, and the center point of the test workpiece 10 and the telecentric lens 2 are kept on the horizontal line. According to the measurement environment and the size of the test workpiece 10, by moving the slider 7 on the slide rail 6, the measurement of the test workpiece 10 can be clearly imaged. The filter 4 is installed in front of the telecentric lens 2 to only allow the same filter to enter the measurement system, and the monochromatic annular light source 3 is installed in front of the filter 4 to make the measurement system in the optimal state.

[0051] Install the tripod 5 for fixing the camera 1, adjust the positions of the camera 1 and the monochromatic annular light source 3 on the slide rail 6, adjust the input power of the monochromatic annular light source 3 to adjust the light intensity of the light source, and keep the telecentric lens 2 at the optimal object distance from the test workpiece 10. The camera 1 uses a CCD camera with a target surface size of 2 / 3 inches.

[0052] Step 2, experimental test: turn on the equipment, set the test parameters, such as the loading force and loading time of the universal testing machine 9 (when the elongation of the resistance strain gauge pasted on the back of the test workpiece 10 is ≤5%), start the image acquisition device (camera 1) to collect the initial reference image of the test workpiece 10, turn on the universal testing machine 9 to load the test workpiece 10 (preheat the universal testing machine 9), and set the universal testing machine 9 to the stretching mode, and the loading method of the test workpiece 10 is vertical upward stretching; set the shooting frame rate of the image acquisition device to shoot an image every 2 seconds, trigger the program operation of the video extensometer measurement system, and execute image acquisition. Continue to collect images of the test workpiece 10 during the loading process until the elongation of the test workpiece 10 reaches the preset threshold of the resistance strain gauge (until the preset threshold of the resistance strain gauge is set to the resistance strain gauge elongation value displayed as 5% and stop taking pictures), and store the captured images in the portable computer 8.

[0053] Step three, calculation and result output: the collected images are transmitted to the video extensometer measurement system, and the deformation images are numbered. According to the displacement of the deformation point in the deformation image relative to the reference standard point in the initial reference image, the elongation of the test workpiece 10 is calculated, and the elongation time curve is plotted. Finally, the calculation results of the transverse strain, longitudinal strain and Poisson's ratio are obtained and displayed on the video extensometer measurement system page.

[0054] A marked point is selected as a reference standard point in the initial reference image, and a deformation image with a corresponding number is selected when the universal testing machine 9 is in elastic deformation during the loading phase to determine a deformation point in the deformation image corresponding to the reference standard point.

[0055] The feature points of the image are detected and matched by combining the FAST feature detection algorithm (Feature from Accelerated Segment Test), the SIFT feature description algorithm (scale invariant feature transform), and the FLANN feature matching algorithm (Fast Library for Approximate Nearest Neighbors).

[0056] The FAST feature detection algorithm quickly detects corner points (feature points) in images with low time complexity, and is suitable for scenarios with high real-time requirements.

[0057] The SIFT feature description algorithm generates high-dimensional descriptors (128-dimensional vectors) for feature points that are invariant to scale, rotation, and illumination, thereby improving the robustness of feature matching.

[0058] The FLANN feature matching algorithm achieves fast approximate matching of high-dimensional feature vectors by constructing a KD tree or prioritizing searching a K-means tree, significantly improving matching efficiency.

[0059] First, the FAST feature detection algorithm is used to detect the feature points in the image (the feature points are the reference standard points selected in the initial reference image), and the SIFT feature description algorithm is used to describe the feature points detected by the FAST algorithm. This method requires detection of both the initial reference image and the deformed image, and the brightness difference between the selected detection point and several surrounding pixels is detected in the initial reference image and the deformed image respectively. The initial reference image and the deformed image generate their own feature points respectively. These feature points are then described and matched.

[0060] (1) Direction matching of feature points: Input: initial reference image and deformed image, grayscaled.

[0061] Testing process: Select pixels in the initial reference image and the deformed image , compare its brightness difference with the surrounding 16 pixels. If there are N (at least 12 in this embodiment) consecutive pixels with brightness differences exceeding the threshold (such as the threshold is ±10% of the brightness, indicating that there is a difference), then determine is the feature point.

[0062] Output: Feature points in the initial reference image and the deformed image and .

[0063] The focus of the description of the detected local features is the robustness and distinguishability of the local features.

[0064] The FAST feature detection algorithm is used to find the feature point positions of the initial reference image. The next step is to describe the feature points using the SIFT feature description algorithm to prepare for the next step of matching.

[0065] Input: The position coordinates of feature points detected by FAST.

[0066] Scale space construction: Gaussian blur is performed on the image to generate a multi-scale pyramid.

[0067] The SIFT feature description algorithm uses the feature points of the DoG pyramid obtained in the FAST feature detection algorithm to collect the Gaussian pyramid image where it is located. The gradient and directional feature distribution of pixels in the neighborhood window. Gradient modulus and direction The calculation expression is as follows: ; ; In the formula, ( ) represents the coordinates of the reference standard point in the initial reference image, Indicates the scale space value of the feature point, sampled according to the scale In principle, sampling is performed with a certain neighborhood radius as the sampling window. In this embodiment, the neighborhood radius is 16×16 as the sampling window. The gradient direction histogram is calculated in the feature point neighborhood (16×16 pixels) to determine the main direction to eliminate the influence of image rotation.

[0068] The gradient calculation process of feature points is as follows Figure 5 The characteristic points shown in the figure generate a histogram diagram. First, Figure 5 The feature point in (a) calculates the gradient direction and modulus of the surrounding pixels through the above formula, as follows: Figure 5 The gradient direction and modulus of the neighboring pixels are shown in (b), which are further divided into Figure 5 (c). The circle is divided into gradient histograms within 0~360°, with 45° as a column and 8 columns, where each histogram value represents the direction of the point, that is, the maximum value of the histogram is the direction of the key point, such as Figure 5 As shown in (d).

[0069] (2) Feature description of feature points: Through the step of direction matching of feature points, each feature point found by the FAST feature detection algorithm is given two important information: position (determined by FAST) and direction (given by SIFT).

[0070] Next, a description vector needs to be established for each feature point so that these feature points do not change with changes in lighting, viewing angle, etc. This SIFT descriptor includes not only this feature point, but also other pixels around the feature point that contribute to it, making the feature point description highly unique, so as to improve the correct matching rate of the feature point.

[0071] Descriptor set generation: Divide the neighborhood into × Sub-regions (specifically divided into 4×4 sub-regions in this embodiment), feature points are described in × Calculation within a (4×4) window Gradient learning in 8 directions (in this embodiment), each sub-region calculates The gradient histogram of the (8) directions is × × =S dimensions, that is, in this embodiment, a vector representation of 4×4×8=128 dimensions is combined into a 128-dimensional vector.

[0072] Output: A set of descriptors for each feature point and .

[0073] The gradient modulus and direction of the feature points are determined; finally, the FLANN algorithm is used for matching.

[0074] FLANN feature matching.

[0075] Input: A set of descriptors of feature points in the initial reference image and the deformed image.

[0076] Matching process: Index construction: Use the FLANN library to describe the initial reference image set Build a KD-tree index.

[0077] Approximate nearest neighbor search: For each set of descriptors in the deformed image , search for the most similar k neighbors (usually k=2) in the KD tree.

[0078] Filter matching pairs: Take a feature of the initial reference image and search for the two features closest to it in the deformed image. The distances are recorded as and , filter false matches by nearest neighbor distance ratio (NNDR): ; If the conditions are met, the matching feature points are retained and .

[0079] Speed ​​improvement: FAST detection is 3 to 5 times faster than SIFT's built-in DoG detection; FLANN matching is more than 10 times faster than Brute-Force matching.

[0080] Accuracy guarantee: The scale invariance of SIFT descriptor set compensates for FAST’s sensitivity to scale changes; FLANN’s approximate search ensures matching reliability through NNDR screening.

[0081] Experimental verification: Matching speed: In 1080p images, the single-frame feature matching time is ≤ 50ms (frame rate 20fps).

[0082] Error control: Compared with resistance strain gauges, strain measurement error is ≤0.5%.

[0083] The use of this matching method (FLANN algorithm for matching) greatly improves the speed of feature point detection and the accuracy of description, thereby improving the matching degree.

[0084] The process of calculating the transverse strain, longitudinal strain and Poisson's ratio of the test workpiece 10 according to the displacement of the deformation point relative to the reference standard point is as follows: Select the length D to be measured in the X direction of the surface of the test workpiece 10, The length L to be measured is selected in the direction. When the test workpiece 10 is loaded, the length D in the X direction and The length L in the direction is deformed, resulting in lateral displacement and longitudinal displacement , the transverse strain and longitudinal strain are shown as follows: Transverse strain: ; Longitudinal strain: ; Poisson's ratio: ; In the above formula, is the lateral strain of the test workpiece 10 when the load is applied, is the longitudinal strain of the test workpiece 10 when the load is applied, represents the Poisson's ratio of the material of the test workpiece 10, ( , ) is the coordinate of the reference standard point, ( , ) are the coordinates of the deformation point; The program interface of the video extensometer measurement system displays the Poisson's ratio of the test workpiece 10 in the deformation image under the corresponding number, as well as the displacement-time curve and the stress-time curve, exports the measurement results, and turns off the equipment.

[0085] In order to prevent errors caused by loose adhesion of the strain gauge and non-vertical direction, several groups of experiments were conducted. The value of the resistance strain gauge (resistance strain gauge) was recorded every five images, and the measured strain value was compared with the strain value recorded by the resistance strain gauge. The results are shown in the figure. Figure 6 .

[0086] The difference between the results of the resistance strain gauge and the video extensometer is 0.434%, which is sufficient to prove that the virtual video extensometer based on digital image correlation method proposed in this paper has high accuracy.

[0087] The novel human-computer interactive video extensometer measurement system and method proposed in the present invention can adapt to the displacement and strain measurement under various working conditions, and a stable lighting scheme is designed. The influence of ambient light can be ignored by the cooperation of an annular monochromatic light source and a bandpass filter. A CCD camera with a 2 / 3 target surface is used as an imaging camera, and two bilateral telecentric lenses with different fields of view are selected. It can not only adapt to the measurement of deformation and strain of objects of various sizes, but also reduce the error caused by the off-plane displacement of the object.

[0088] A human-computer interaction interface is designed to select virtual markers, output the displacement, strain and Poisson's ratio of the specimen, and generate corresponding displacement-time graphs and strain-time graphs. The combination of software and hardware enables accurate and efficient measurement of the video extensometer, breaking through the limitations of traditional extensometers.

[0089] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A novel human-computer interactive video extensometer measurement method, characterized in that: The following steps are involved: Equipment debugging preparation: install the test workpiece on the universal testing machine, install and debug the image acquisition equipment, paste the resistance strain gauge on the back of the test workpiece on the universal testing machine, and connect it to the resistance strain gauge, and the resistance strain gauge is connected to the video extensometer measurement system; Experimental test: After starting the image acquisition device to acquire the initial reference image of the test workpiece, start the universal testing machine to load the test workpiece, set the shooting frame rate of the image acquisition device, and continuously acquire the deformation image of the test workpiece during the loading process until the stretching rate of the test workpiece reaches the preset threshold of the resistance strain gauge; Calculation and result output: The collected initial reference image and deformation image are transmitted to the video extensometer measurement system, and the deformation image is numbered. The transverse strain, longitudinal strain and Poisson's ratio of the test workpiece are calculated according to the displacement of the deformation point in the deformation image relative to the reference standard point in the initial reference image, and displayed through the program interface of the video extensometer measurement system.

2. According to the novel human-computer interactive video extensometer measurement method of claim 1, it is characterized in that: Before the test workpiece is installed on the universal testing machine, white paint is first sprayed on the surface of the test workpiece as a base color, and then black paint dots are sprayed to form a random speckle pattern, and marking lines are engraved on the test workpiece; the test workpiece is installed on the universal testing machine, the fixture is first fixed to the upper and lower sides of the universal testing machine, and then the test workpiece is placed in the fixture and clamped.

3. According to the novel human-computer interactive video extensometer measurement method of claim 1, it is characterized in that: When the resistance strain gauge is pasted on the back of the test workpiece, prepare the same test workpiece and paste the resistance strain gauge at the same position without loading it, and connect it to the resistance strain gauge as a temperature compensation sheet; Installing and debugging image acquisition equipment includes: A telecentric lens is installed at the front end of the camera, a filter is arranged in front of the telecentric lens, a monochromatic annular light source is installed in front of the filter, a tripod is placed in front of the universal testing machine, a slide rail is installed on the upper part of the tripod, the slide rail and the workpiece to be tested are on the same vertical plane, the camera and the monochromatic annular light source are both slidably arranged on the slide rail, and the monochromatic annular light source is located between the camera and the universal testing machine; Install the tripod for fixing the camera, adjust the position of the camera and the monochromatic ring light source on the slide rail, and adjust the input power of the monochromatic ring light source to adjust the light intensity of the light source.

4. A novel human-computer interactive video extensometer measurement method according to claim 3, characterized in that: The camera uses a CCD camera with a target surface size of 2 / 3 inches; the shooting frame rate of the image acquisition device is set to take an image every 2 seconds, and the preset threshold of the resistance strain gauge is set to the resistance strain gauge elongation ≤5%.

5. The novel human-computer interactive video extensometer measurement method according to claim 1 is characterized in that: Select a marked point in the initial reference image as a reference standard point, select a deformation image with a corresponding number when the universal testing machine is in elastic deformation during the loading phase, and determine the deformation point in the deformation image corresponding to the reference standard point; The FAST feature detection algorithm is used to detect the reference standard points selected in the initial reference image, and the SIFT feature description algorithm is used to describe the feature points detected by the FAST algorithm. The process is as follows: (1) Direction matching of feature points: Input the initial reference image and the deformed image and perform grayscale processing; Select pixels in the initial reference image and the deformed image , compare its brightness difference with surrounding pixels; If the brightness difference of N pixels exceeds the threshold, the judgment is the feature point; Output the feature points in the initial reference image and the deformed image and ; The FAST feature detection algorithm is used to find the position of the reference standard point in the initial reference image, and the feature point is described by the feature description algorithm provided by the SIFT feature description algorithm: Input the feature point position coordinates; The SIFT feature description algorithm uses the feature points of the DoG pyramid obtained in the FAST feature detection algorithm to collect the Gaussian pyramid image where it is located. Gradient and directional feature distribution of pixels in the neighborhood window, gradient modulus and direction The calculation of is as follows: ； ； In the formula, ( ) represents the coordinates of the reference standard point in the initial reference image, Indicates the scale space value of the feature point, sampled according to the scale Principle: Sampling is performed with a certain neighborhood radius as the sampling window; (2) Feature description of feature points: For the feature points found by the FAST feature detection algorithm, two important information are given to the feature points: position and direction, and a description vector is established for each feature point; Descriptor set generation: Divide the neighborhood into × Sub-region, feature points are described in × Calculation within the window Gradient learning in each direction, obtaining ( × × ) vector representation of dimensions; Output the descriptor set for each feature point and ; Establish the gradient modulus and direction of the feature points and use the FLANN algorithm for matching: Input the descriptor set of feature points in the initial reference image and the deformed image, Use the FLANN library to generate a set of descriptors for feature points in the initial reference image Build a KD tree index, Approximate nearest neighbor search: a set of descriptors for each feature point in the deformed image , search for the most similar k neighbors in the KD tree; Filter the matching pairs, take a feature of the initial reference image, and search for the two features closest to it in the deformed image. The distances are recorded as and , filter out false matches through the nearest neighbor distance ratio NNDR: ； If the above conditions are met, the matching feature points are retained and .

6. A novel human-computer interactive video extensometer measurement method according to claim 5, characterized in that: The process of calculating the transverse strain, longitudinal strain and Poisson's ratio of the test workpiece based on the displacement of the deformation point relative to the reference standard point is as follows: Select the length D to be measured in the X direction of the test workpiece surface. Direction Select the length L to be measured. When the test workpiece is loaded, the length D in the X direction and The length L in the direction is deformed, resulting in lateral displacement and longitudinal displacement , the transverse strain and longitudinal strain are shown as follows: Transverse strain: ; Longitudinal strain: ; Poisson's ratio: ; In the above formula, is the lateral strain of the test workpiece when the load is applied, is the longitudinal strain of the test workpiece when the load is applied, represents the Poisson's ratio of the test workpiece material, ( , ) is the coordinate of the reference standard point, ( , ) are the coordinates of the deformation point; The program interface of the video extensometer measurement system displays the Poisson's ratio of the test workpiece in the deformation image under the corresponding number, as well as the displacement-time curve and stress-time curve.

7. A video extensometer measurement system applied to the novel human-computer interactive video extensometer measurement method according to any one of claims 1 to 6, characterized in that: The video extensometer measurement system includes: The camera is used to collect the image of the test workpiece. A telecentric lens is installed on the camera. A filter is arranged in front of the telecentric lens. A monochromatic annular light source is installed in front of the filter. The slide rail and the test workpiece to be tested are on the same vertical plane. The camera and the monochromatic annular light source are both slidably arranged on the slide rail. The monochromatic annular light source is located between the camera and the universal testing machine. The slide rail is installed on the upper part of the tripod. Computer, install and run the video extensometer measurement system program, establish data transmission connection with the camera and the universal testing machine, receive the camera captured images and the universal testing machine loading data, calculate the transverse strain, longitudinal strain and Poisson's ratio of the test workpiece according to the captured images, and display them in real time through the interface; The universal testing machine applies a loading force to the test workpiece to cause the test workpiece to deform and obtain the mechanical performance parameters of the test workpiece.

8. The video extensometer measurement system according to claim 7, characterized in that: The camera uses a CCD camera with a target size of 2 / 3 inches; the monochromatic annular light source uses a monochromatic LED light source with a wavelength of 465-470nm, and the irradiation angle of the monochromatic annular light source is 60° with the horizontal ground; the filter uses a filter with a central wavelength of 465nm, a half-width of 10nm, and a transmittance of 80%.

9. The video extensometer measurement system according to claim 7, characterized in that: The telecentric lens is a bilateral telecentric lens, which adopts a small-field bilateral telecentric lens or a large-field bilateral telecentric lens. The magnification of the small-field bilateral telecentric lens is 0.22, and the field of view range is 40×30mm; the magnification of the large-field bilateral telecentric lens is 0.0716, and the field of view range is 122.9×92.2mm.

10. The video extensometer measurement system according to claim 7, characterized in that: The program interface of the video extensometer measurement system installed and run on the computer includes a control area and a parameter area, wherein the control area is a function control area and the parameter area is a parameter display area.

Images