A monocular vision large-scale rotating body diameter dynamic measurement device and method

Abstract

The present invention discloses a monocular vision dynamic measurement device and method for the diameter of a large-scale rotating object. When not measuring, a first triangular reflector and a second triangular reflector are respectively close to a beam splitter and a third triangular reflector, the grating scale reading is 0, and a monocular line array camera obtains an initial single-frame line array image of two parallel light paths. During measurement, a mobile platform moves longitudinally along a ball screw and a slide rail, and the rotating object to be measured is placed within the area enclosed by the mobile platform, the ball screw, and the slide rail. The rotating object to be measured blocks a portion of the parallel light A and a portion of the parallel light B between the first and second triangular reflectors. The monocular line array camera obtains a single-frame line array image of the two parallel light paths after blocking. The diameter of the rotating object to be measured is calculated based on the initial single-frame line array image and the single-frame line array image after blocking. The present invention has the advantages of good installation robustness, simple structure, high measurement accuracy, good synchronization, dynamic detection, and no need for repeated calibration.

Description

Technical Field

[0001] The present invention belongs to the technical field of rotating body size measurement, and in particular relates to a monocular vision large-scale rotating body diameter dynamic measurement device and method. Background Art

[0002] In the automotive, bearing, and defense industries, online inspection of large rotating workpieces has always been a challenge that affects accuracy. Domestic and international scholars have conducted extensive research on measuring the diameter and related parameters of rotating parts.

[0003] Currently, diameter measurement of large rotating objects falls into two main categories: contact and non-contact. Contact measurement typically involves manual measurement of large outer diameters using large calipers and gauges after workpiece machining. This method is outdated, inconvenient, inefficient, and lacks precision. Continuous measurement of rotating objects is not possible, and human factors significantly influence the process. Non-contact detection primarily involves laser sensors and machine vision. Laser measurement utilizes a distributed dual-sensor system, requiring calibration. A solution employing a split-type laser sensor system employs two laser sensors. Split-type detection devices require calibration with a standard block of the same size as the rotating object before each use. Furthermore, the dual sensors suffer from poor synchronization, and the diameter of the rotating object is determined by comparing the data collected by the two sensors. This is complex, susceptible to interference, and exhibits poor accuracy. Vision measurement utilizes a diameter measurement system based on the principle of pinhole imaging. This system measures the diameter of cylindrical workpieces by calculating the backlit edge of the cylinder. However, measurement accuracy is affected by factors such as the position of the workpiece and the quality of the captured image. Furthermore, due to limited sensor resolution, measurement accuracy is relatively low.

[0004] Mechanical displacement is the change in position of a mechanical object in a specific direction. During mechanical motion, the axis of rotation deviates from its ideal position, and this deviation requires measurement and evaluation. Radial runout, a manifestation of displacement, is the most commonly used metric in mechanical measurement. In actual production, high-speed rotary shafts require monitoring of radial runout to ensure that the runout remains within a predetermined range. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the existing technology and provide a monocular vision large-scale rotating body diameter dynamic measurement device and method, which solves the problems of complex operation and low accuracy in the existing rotating body diameter measurement process.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a monocular vision large-scale rotating body diameter dynamic measurement device, comprising a parallel line light source, a spectroscope, a first triangular reflector, a second triangular reflector, a third triangular reflector, a monocular line array camera, a fixing device, a servo motor, a ball screw, a mobile platform, a slide rail and a grating ruler, the ball screw and the slide rail are respectively fixed on the left and right sides of the fixing device, the ball screw and the slide rail are arranged in parallel longitudinally, one end of the transversely arranged mobile platform is connected to the nut of the ball screw, and the other end is slidably connected to the slide rail, the output shaft of the servo motor is connected to the screw rod of the ball screw, the parallel line light source is fixed on one side of the servo motor, the spectroscope is fixed on the plane Behind the line light source, the first and second triangular reflectors are fixed on the left and right ends of the mobile platform respectively. The monocular line array camera is fixed on the front end of the slide rail. The third triangular reflector is set on the side of the monocular line array camera. The parallel line light source emits parallel light, which is divided into two beams of parallel light in orthogonal directions by the beam splitter. One beam of parallel light A is the 90° reflected light of the beam splitter, and the other beam of parallel light B is the transmitted light of the beam splitter. Parallel light A is directly incident on the monocular line array camera, and parallel light B is reflected by the first triangular reflector, the second triangular reflector and the third triangular reflector in sequence, and finally enters the monocular line array camera. The grating ruler is used to obtain the distance between the beam splitter and the first triangular reflector.

[0007] The grating ruler includes a matched scale grating and an indicator grating. The scale grating is fixed along the length direction of the ball screw, and the indicator grating is fixed on one end of the moving platform close to the scale grating.

[0008] Specifically, the present invention relates to a monocular vision dynamic measuring device for the diameter of a large-size rotating body, which also includes a first sliding rod, a second sliding rod, a first fixed block, a first slider, a second fixed block and a second slider. The first sliding rod is arranged above the ball screw, the second sliding rod is arranged above the slide rail, the first sliding rod and the second sliding rod are arranged in parallel, one end of the first sliding rod is fixed on the first fixed block, the other end of the first sliding rod passes through the through hole on the first slider and the first slider slides along the first sliding rod, the spectrometer is fixed on the first fixed block, the first triangular reflector is fixed on the first slider, one end of the second sliding rod is fixed on the second fixed block, the other end of the second sliding rod passes through the through hole on the second slider and the second slider slides along the second sliding rod, the second triangular reflector is fixed on the second slider, and the third triangular reflector is fixed on the second fixed block.

[0009] Specifically, the fixing device is a mechanical frame with a longitudinal length of H and a transverse width of L, where L>H, and the diameter measurement range of the measured rotating body is S~H.

[0010] Furthermore, the present invention relates to a monocular vision dynamic measurement method for the diameter of a large-size rotating body, which also includes a turntable and an encoder. The measured rotating body is fixed on the turntable, and the encoder is arranged on the turntable. The turntable drives the measured rotating body to rotate, and the encoder is used to measure the rotation angle of the rotating body. The outer contour of the measured rotating body can be determined based on the diameter and the rotation angle.

[0011] Specifically, based on the diameter, the radial end runout of the detected rotating body can be detected.

[0012] The present invention relates to a specific measurement method of a monocular vision large-scale rotating body diameter dynamic measurement device, which includes the following steps: (1) calibration: first, the rotating body to be measured is not placed, and the mobile platform is close to the spectroscope. At this time, the grating scale reading is 0, and the parallel light A is directly incident on the monocular line array camera. The parallel light B is reflected by the first triangular reflector, the second triangular reflector and the third triangular reflector in sequence, and finally enters the monocular line array camera. The two light rays form two light bands in the line array camera. The monocular line array camera obtains an initial single-frame line array image containing two parallel light bands, removes the background and noise of the initial single-frame line array image, and extracts the column coordinates P of the parallel light band edge. 0 1, P 0 2, P 0 3, P 0 4. Calculate the distance S between the two parallel light bands A and B, as well as the width W of the two parallel light bands A and B when the rotating object is not placed. S = P 0 3 – P 0 2. W = P 0 4-P 0 1-S.

[0013] Specifically, (101) a threshold method was used to remove the background and noise of the initial single-frame linear array image, retaining only the information of the parallel light bands.

[0014] (2)

[0015] Where, is the grayscale value of the initial single-frame linear image after removing background and noise, threshold is the threshold for background removal, is the grayscale value of the initial single-frame linear image.

[0016] (102) Using the local edge search algorithm, the four edge column coordinates P of the parallel light band of the initial single-frame linear array image after processing in step (101) are obtained. 0 1, P 0 2, P 0 3, P 0 4. Calculate the distance S between the two parallel light bands A and B, as well as the width W of the two parallel light bands A and B when the rotating object to be measured is not placed.

[0017] (2) Single diameter measurement: The mobile platform moves, and the measured rotating body is placed in the area surrounded by the mobile platform, ball screw and slide rail. The measured rotating body blocks part of the parallel light A and part of the parallel light B between the first triangular reflector and the second triangular reflector. The grating scale reading is G. The two blocked light rays also form two light bands in the linear array camera. The monocular linear array camera obtains a single-frame linear array image containing two parallel light bands after blocking. The background and noise of the single-frame linear array image after blocking are removed, and the bilateral centroid method is used to extract the column coordinates P of the parallel light band edge. 1 1, P 1 2, P 1 3, P 1 4. Calculate the width T1 of the parallel light A and the width T2 of the parallel light band B after the rotating object is blocked. T1 = P 1 2-P 1 1, T2 = P 1 4-P 1 3. T=T1+T2. Calculate the diameter D of the rotating object to be measured according to the following formula.

[0018]

[0019] Specifically, (201) a threshold method was used to remove the background and noise of a single-frame linear array image after occlusion, retaining only the information of parallel light bands.

[0020]

[0021] Where, is the grayscale value of a single frame linear array image after removing the background and noise, threshold is the threshold for background removal, is the grayscale value of a single frame linear array image after occlusion.

[0022] (202) Linear array cameras use a lensless imaging method that does not conform to the point diffusion model. The incident light is parallel light, so the black and white transition of edge pixels does not conform to the Gaussian distribution, and the Gaussian fitting method cannot be used to obtain sub-pixel coordinates. The present invention uses the bilateral centroid method to identify the transition center of the edge and uses the bilateral centroid method to extract the column coordinates P of the light band edge. 1 2, P 1 3, and P 1 1=P 0 1, P 1 4= P 0 4.

[0023]

[0024] Where Lc is the center coordinate of the dark area between the two parallel light bands, mis the column coordinate in each row, is the grayscale value of the mth column coordinate (pixel) after processing in step (201), P 1 1, P 1 2, P 1 3 and P 1 4 is the edge column coordinate of the parallel light band in the single-frame linear array image after occlusion.

[0025] (3) Measurement of diameter in one circle: The rotating body to be measured is fixed on a turntable, which drives the rotating body to be measured to rotate one circle. The monocular linear array camera collects the spatiotemporal image of the rotating body to be measured in real time. Each frame of the spatiotemporal image is a line, which is a single-frame linear array image. According to step (2), the diameter corresponding to each line of the single-frame linear array image is calculated, and the diameter of the rotating body to be measured in one circle is obtained.

[0026] Furthermore, based on the diameter obtained in step (2) or (3), the radial end runout of the rotating body under test can be detected.

[0027] Furthermore, based on the diameter of the rotating body obtained in step (3) and the rotation angle of the rotating body, the outer contour of the rotating body to be measured can be determined.

[0028] The present invention provides a monocular vision large-scale rotating body diameter measuring device capable of solving large-scale, high-precision, simple operation and improving the real-time performance of image processing algorithms.

[0029] The beneficial effects of the present invention are: (1) It overcomes the shortcomings of the traditional method, such as the need for multiple sensors and measurement accuracy. A monocular linear array camera is used to receive linear parallel light emitted by the same light source, with good synchronization. According to the width of the two parallel light bands obtained by the linear array camera and combined with the reading of the grating ruler, the sub-pixel level shaft diameter can be obtained, thereby improving the detection accuracy; (2) It only needs to use the standard diameter to initially calibrate the device at the factory, and there is no need for repeated calibration; (3) The linear array camera has a high frame rate, adopts optical and image processing methods, uses linear parallel light combined with an absolute grating as a reference, and can realize real-time dynamic measurement of the diameter, radial end runout and outer contour of a large rotating body; (4) The mechanical motion accuracy requirements of the servo motor are low. Since the diameter of the rotating body is superimposed by the pixels of the grating scale and the linear array camera, the parallel light band has a width. As long as the moving part can cover the rotating body, the repeated coverage area does not affect the calculation of the rotating body diameter; (5) During measurement, it is only necessary for the rotating body to block the parallel light, and there is no need to strictly control the position. The horizontal width L of the fixing device is greater than the vertical length H, which effectively prevents the parallel light between the second triangular reflector and the third triangular reflector from being blocked, and has high robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 The invention relates to an axonometric diagram of a monocular vision large-scale rotating body diameter dynamic measurement device.

[0031] Figure 2 The present invention relates to a monocular vision large-scale rotating body diameter dynamic measurement device.

[0032] Figure 3 The present invention relates to a flow chart of a method for dynamically measuring the diameter of a large-scale rotating object using monocular vision.

[0033] Figure 4 is a single-frame linear array image before and after threshold processing according to the present invention, wherein: Figure 4 In the middle, a is a single-frame linear image before processing. Figure 4 Middle b is a single frame linear image after processing.

[0034] Figure 5 The schematic diagram of the monocular vision large-scale rotating body diameter dynamic measurement method of the present invention is shown in FIG. Figure 5 Figure a is a schematic diagram of the principle of single diameter measurement in step (2). Figure 5 Where b is the initial single-frame linear array image obtained in step (1).

[0035] In the figure: 1 is a parallel line light source; 2 is a spectrometer; 3 is a first triangular reflector; 4 is a second triangular reflector; 5 is a third triangular reflector; 6 is a monocular line array camera; 7 is a fixing device; 8 is a servo motor; 9 is a ball screw; 10 is a moving platform; 11 is a slide rail; 12 is a grating ruler; 13 is a first slide bar; 14 is a second slide bar; 15 is a first fixed block; 16 is a first slider; 17 is a second fixed block; 18 is a second slider; 1201 is a scale grating; 1202 is an indicator grating; A is parallel light A; B is parallel light B. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described below with reference to the accompanying drawings and specific examples.

[0037] Example

[0038] like Figure 1 and 2As shown, the present embodiment involves a monocular vision large-scale rotating body diameter dynamic measurement device, comprising a parallel line light source 1, a spectroscope 2, a first triangular reflector 3, a second triangular reflector 4, a third triangular reflector 5, a monocular line array camera 6, a fixing device 7, a servo motor 8, a ball screw 9, a movable platform 10, a slide rail 11 and a grating scale 12, wherein the ball screw 9 and the slide rail 11 are respectively fixed on the left and right sides of the fixing device 7, the ball screw 9 and the slide rail 11 are arranged parallel to each other in the longitudinal direction, one end of the transverse movable platform 10 is connected to the nut of the ball screw 9, and the other end is slidably connected to the slide rail 11, the output shaft of the servo motor 8 is connected to the screw of the ball screw 9, the parallel line light source 1 is fixed on one side of the servo motor 8, the spectroscope 2 is fixed on the rear side of the parallel line light source 1, the first triangular reflector 3 and the second triangular reflector 4 are respectively fixed on the left and right ends of the movable platform 10, the monocular line array camera 6 Fixed at the front end of the slide rail 11, the third triangular reflector 5 is arranged on one side of the monocular line array camera 6. The parallel line light source 1 emits a line parallel light, which is divided into two beams of parallel light in orthogonal directions by the beam splitter 2. One beam of parallel light A is the 90° reflected light (horizontal) of the beam splitter 2, and the other beam of parallel light B is the transmitted light (longitudinal) of the beam splitter 2. The parallel light A is directly incident on the monocular line array camera, and the parallel light B is reflected by the first triangular reflector 3, the second triangular reflector 4 and the third triangular reflector 5 in sequence, and finally enters the monocular line array camera. The grating ruler 12 includes a matching scale grating (main grating) 1201 and an indicator grating (auxiliary grating) 1202. The scale grating 1201 is fixed along the length direction of the ball screw 9, and the indicator grating 1202 is fixed on the end of the mobile platform 10 close to the scale grating 1201. The grating ruler 12 is used to obtain the distance between the beam splitter 2 and the first triangular reflector 3.

[0039] This embodiment relates to a monocular vision dynamic measurement device for the diameter of a large-scale rotating object. When not measuring, the first triangular reflector 3 and the second triangular reflector 4 are respectively close to the beam splitter 2 and the third triangular reflector 5, the grating scale 12 reads 0, and the monocular line array camera 6 obtains an initial single-frame line array image of two parallel light beams. During measurement, driven by the servo motor 8, the mobile platform 10 moves longitudinally along the ball screw 9 and the slide rail 11. The measured rotating object is placed in the area surrounded by the mobile platform 10, the ball screw 9, and the slide rail 11. The measured rotating object blocks part of the parallel light beam A and the part of the parallel light beam B between the first triangular reflector 3 and the second triangular reflector 4. The monocular line array camera 6 obtains a single-frame line array image of the two parallel light beams after blocking. The diameter of the measured rotating object is calculated based on the initial single-frame line array image and the single-frame line array image after blocking.

[0040] The parallel line light source 1 , the beam splitter 2 , the first triangular reflector 3 , the second triangular reflector 4 , the third triangular reflector 5 and the monocular line array camera 6 involved in this embodiment constitute an optical module.

[0041] The fixing device 7 , servo motor 8 , ball screw 9 , moving platform 10 , slide rail 11 , slider and grating ruler 12 involved in this embodiment constitute a mechanical module.

[0042] The first triangular reflector 3 , the second triangular reflector 4 and the third triangular reflector 5 involved in this embodiment constitute a triangular reflector module.

[0043] Specifically, the present embodiment involves a monocular vision large-scale rotating body diameter dynamic measuring device, which also includes a first slide bar 13, a second slide bar 14, a first fixed block 15, a first slider 16, a second fixed block 17 and a second slider 18. The first slide bar 13 is arranged above the ball screw 9, and the second slide bar 14 is arranged above the slide rail 11. The first slide bar 13 and the second slide bar 14 are arranged in parallel. One end of the first slide bar 13 is fixed on the first fixed block 15, the other end of the first slide bar 13 passes through the through hole on the first slider 16 and the first slider 16 slides along the first slide bar 13, the spectrometer 2 is fixed on the first fixed block 15, the first triangular reflector 3 is fixed on the first slider 16, one end of the second slide bar 14 is fixed on the second fixed block 17, the other end of the second slide bar 14 passes through the through hole on the second slider 18 and the second slider 18 slides along the second slide bar 14, the second triangular reflector 4 is fixed on the second slider 18, and the third triangular reflector 5 is fixed on the second fixed block 17. This structure can improve the coaxiality among the beam splitter 2, the first triangular reflector 3, the second triangular reflector 4, and the third triangular reflector 5, reduce the influence of the movement of the mobile platform 10 on the optical path structure of the parallel light B, and improve the accuracy of detection.

[0044] Specifically, the spectroscope 2 is a BS spectroscope 2 .

[0045] Specifically, the grating scale 12 is an absolute grating scale.

[0046] Specifically, fixture 7 is a mechanical frame having a longitudinal length of H and a transverse width of L, where L>H. This prevents the parallel light between second triangular reflector 4 and third triangular reflector 5 from being blocked. The diameter measurement range of the rotating object under test is S to H. For example, if S is 10 mm and H can be up to 1000 mm, the measurement range from 10 to 1000 mm can be used to measure both small and large diameters while ensuring measurement accuracy.

[0047] Furthermore, the embodiment involves a monocular vision dynamic measurement device for the diameter of a large-size rotating body, which also includes a turntable and an encoder. The measured rotating body is fixed on the turntable, and the encoder is arranged on the turntable. The turntable drives the measured rotating body to rotate, and the encoder is used to measure the rotation angle of the rotating body. The outer contour of the measured rotating body can be determined based on the diameter and the rotation angle.

[0048] like Figure 3As shown, this embodiment involves a monocular vision large-scale rotating object diameter dynamic measurement method, which includes the following steps.

[0049] (1) Calibration: First, the rotating object to be measured is not placed, and the mobile platform 10 is close to the spectroscope 2. At this time, the grating scale 12 reads 0. The parallel light A is directly incident on the monocular line array camera 6. The parallel light B is reflected by the first triangular reflector 3, the second triangular reflector 4 and the third triangular reflector 5 in sequence, and finally enters the monocular line array camera 6. The two light beams (parallel light A and parallel light B) form two light bands in the line array camera, and the monocular line array camera 6 obtains an initial single-frame line array image containing two parallel light bands ( Figure 5 b middle light band), remove the background and noise of the initial single-frame linear array image, and extract the column coordinates P of the parallel light band edge 0 1, P 0 2, P 0 3, P 0 4, among which, P 0 1 and P 0 2 are the column coordinates of the left and right edges of the light band formed by the parallel light A in the initial single-frame linear array image, P 0 3 and P 0 4 are the column coordinates of the left and right edges of the light band formed by the parallel light B in the initial single-frame linear array image. Calculate the distance S between the two parallel light bands A and B, as well as the width W of the two parallel light bands A and B when the rotating body is not placed. S = P 0 3 – P 0 2. W = P 0 4-P 0 1-S, W = W1+ W2, W1 is the width of parallel light band A, W2 is the width of parallel light band B.

[0050] Specifically, (101) a threshold method was used to remove the background and noise of the initial single-frame linear array image, retaining only the information of the parallel light bands.

[0051] (2)

[0052] Where, is the gray value of the initial single-frame linear array image to remove background and noise. threshold is the threshold for background removal. For example, threshold = mean(gray)+5, where mean(gray) is the gray value average of the background area between parallel light bands. is the grayscale value of the initial single-frame linear image. Figure 4 is a single-frame linear array image before and after threshold processing according to the present invention, wherein: Figure 4In the middle, a is a single-frame linear image before processing. Figure 4 Figure b is a single-frame linear image after processing. It can be seen that two clear light bands can be obtained after threshold processing.

[0053] (102) Using the local edge search algorithm, the four edge column coordinates P of the parallel light band of the initial single-frame linear array image after processing in step (101) are obtained. 0 1, P 0 2, P 0 3, P 0 4. Calculate the distance S between the two parallel light bands A and B, as well as the width W of the two parallel light bands A and B when the rotating object to be measured is not placed.

[0054] During calibration, the edge may fluctuate by one pixel, so local edge search is used to find the edge coordinates. Figure 5 As shown in b, when the initial single-frame linear array image is placed horizontally, the local edge search algorithm is used. Specifically, the first step is to start from the left side of the initial single-frame linear array image captured by the monocular linear array camera, search to the right for the edge pixel of the parallel light band with the first grayscale value step, and search downward to determine whether it is the step point. The column coordinates of the pixel point are P 0 1 ; The second step is to search for the first grayscale value step parallel light band edge pixel point from the center of the initial single-frame linear array image to the left, and search downward to determine whether it is a step point. The column coordinates of the pixel point are P 0 2 ; The third step is to search for the edge pixel of the parallel light band with the first gray value step from the center of the initial single-frame linear array image to the right, and then search downward to determine whether it is the step point. The column coordinates of the pixel point are P 0 3 ; Starting from the right side of the initial single-frame linear image captured by the linear camera, search to the left for the first parallel light band edge pixel with a grayscale value step, and search downward to determine whether it is a step point. The column coordinates of the pixel point are P 0 4 .

[0055] (2) Single diameter measurement: The mobile platform 10 moves, and the rotating object to be measured is placed in the area surrounded by the mobile platform 10, the ball screw 9 and the slide rail 11, and the rotating object to be measured blocks part of the parallel light A ( Figure 5 Middle a X 1 part), and shielding part of the parallel light B between the first triangular reflector 3 and the second triangular reflector 4 ( Figure 5 Middle a X 2Part), the grating ruler 12 reads G, and the two blocked light rays (parallel light A and parallel light B) also form two light bands in the linear array camera. The monocular linear array camera 6 obtains a single-frame linear array image containing two parallel light bands after blocking ( Figure 5 The background and noise of the single-frame linear array image after occlusion are removed, and the edge column coordinates P of the parallel light band in the single-frame linear array image after occlusion are extracted using the bilateral centroid method. 1 1, P 1 2, P 1 3, P 1 4， in, P 1 1 and P 1 2 are the column coordinates of the left and right edges of the light band formed by the parallel light A in the single-frame linear array image after occlusion, P 1 3 and P 1 4 are the column coordinates of the left and right edges of the light band formed by the parallel light B in the single frame linear array image after occlusion. Calculate the width T1 of the parallel light A and the width T2 of the parallel light band B after occlusion when the measured rotating body is placed. T1= P 1 2-P 1 1, T2=P 1 4-P 1 3. T=T1+T2, calculate the diameter D of the rotating body to be measured according to the following formula:

[0056] Figure 5 The light band in b is not placed on the rotating body to be measured. When the grating ruler 12 reads 0, the monocular linear array camera 6 obtains an initial single-frame linear array image containing two parallel light bands. Figure 5 The lower light band in center a is the reading G on the grating ruler 12. When the rotating body is not inserted, the monocular line array camera 6 captures a single-frame linear image. Since both parallel light beams are unobstructed in both cases, the strip formed by parallel light A at the monocular line array camera 6 and the strip formed by parallel light B after reflection from the first triangular reflector 3, the second triangular reflector 4, and the third triangular reflector 5 are of the same width at the monocular line array camera 6. Furthermore, since the positions of the third triangular reflector 5 and the beam splitter 2 remain unchanged, the width of the parallel light beams A and B is the same when entering the monocular line array camera 6, i.e., the distance between the two strips is the same. Therefore, Figure 5 The single-frame linear array image in the lower middle part of a is the same as the initial single-frame linear array image, that is, the four edge column coordinates P 0 1 , P 0 2 , P 03 , P 0 4 , W1 and W2 are the same. Since the actual distance represented by the distance between the two parallel strips is the distance between the beam splitter 2 and the first triangular reflector 3, the distance is measured by the grating ruler 12, Figure 5 The actual distance represented by the distance between the two parallel strips in b is S (corresponding to the grating ruler 12 reading of 0). Figure 5 The actual distance represented by the distance between the two parallel strips at the bottom of a is S+G (the corresponding reading of the grating ruler 12 is G). Figure 5 The lower light band in middle a Figure 5 The structure of the light band b is the same and is not measured in the actual measurement process. It is only used here to explain the principle of the present invention.

[0057] Specifically, (201) the threshold method is used to remove the background and noise of the single-frame linear array image after occlusion, and only the information of the parallel light bands is retained:

[0058]

[0059] Where, is the grayscale value of a single frame linear array image after removing the background and noise, threshold is the threshold for background removal, is the grayscale value of a single frame linear array image after occlusion.

[0060] (202) Linear array cameras use a lensless imaging method that does not conform to the point diffusion model. The incident light is parallel light, so the black and white transition of edge pixels does not conform to the Gaussian distribution, and the Gaussian fitting method cannot be used to obtain sub-pixel coordinates. The present invention uses the bilateral centroid method to identify the transition center of the edge and uses the bilateral centroid method to extract the column coordinates P of the light band edge. 1 2, P 1 3, and P 1 1=P 0 1, P 1 4= P 0 4.

[0061]

[0062] Where Lc is the center column coordinate of the dark area between the two parallel light bands, m is the column coordinate in each row, is the grayscale value of the coordinate (pixel) in the mth column after processing in step (201).

[0063] (3) Measurement of diameter in one circle: The rotating body to be measured is fixed on a turntable, which drives the rotating body to be measured to rotate one circle. The monocular linear array camera collects the spatiotemporal image of the rotating body to be measured in real time. Each frame (row) of the spatiotemporal image is a single-frame linear array image. According to step (2), the diameter corresponding to each frame (row) of the single-frame linear array image can be calculated, and then the diameter of the rotating body to be measured in one circle can be obtained.

[0064] Specifically, the measured rotating body rotates, and the single-frame linear array images obtained on the monocular linear array camera 6 are arranged frame by frame in the time acquisition order to form a spatiotemporal image. Each frame image in the spatiotemporal image is a row, which is a single-frame linear array image. The horizontal and vertical coordinates of the spatiotemporal image represent space and time respectively. The single-frame linear array image is a one-dimensional image. The horizontal resolution is the spatial resolution, and the vertical resolution depends on the speed of the measured rotating body and the frame rate of the linear array camera, which is the temporal resolution. It is mainly based on the angular velocity change of the measured rotating body. The monocular linear array camera collects the spatiotemporal image of the measured rotating body rotating one circle in real time, and processes each single-frame linear array image of the spatiotemporal image according to step (2) to obtain the diameter of the measured rotating body one circle.

[0065] Step (3) first uses the threshold method to remove the background and noise of each line of single-frame linear array images in the spatiotemporal image, and then uses the bilateral centroid method to extract the column coordinates of the light band edge of each line of single-frame linear array images after removing the background and noise. P i 2 , P i 3 ,and P i 1 = P 0 1 , P i 4 = P 0 4 .

[0066]

[0067] Where Lc is the center column coordinate of the dark area between the two parallel light bands, m is the column coordinate in each row, is the grayscale value of the coordinate (pixel) in the i-th row and m-th column of the spatiotemporal image after removing the background and noise, where P i 1 and P i 2 are the column coordinates of the left and right edges of the light band formed by the parallel light A in the i-th row of the single-frame linear array image in the space-time image, Pi 3 and P i 4 are the column coordinates of the left and right edges of the light band formed by the parallel light B in the single-frame linear array image of the i-th row in the space-time image, P i 1 = P 0 1 , P i 4 = P 0 4 .

[0068] Furthermore, based on the diameter obtained in step (2) or (3), the radial end runout of the rotating body under test can be detected.

[0069] Furthermore, based on the diameter of the rotating body obtained in step (3) and the rotation angle of the rotating body, the outer contour of the rotating body can be determined.

[0070] Of course, the above contents are only preferred embodiments of the present invention and should not be considered as limiting the scope of the embodiments of the present invention. The present invention is not limited to the above examples. Equivalent changes and improvements made by ordinary technicians in this technical field within the essential scope of the present invention should all fall within the scope of the patent of the present invention.

Claims

1. A monocular vision large-scale rotating body diameter dynamic measurement method, characterized in that: The measurement is performed using a monocular vision large-scale rotating body diameter dynamic measurement device, which includes the following steps: (1) Calibration: First, the rotating object to be measured is not placed, and the mobile platform is close to the spectrometer. At this time, the grating scale reading is 0. The parallel light A is directly incident on the monocular line array camera. The parallel light B is reflected by the first triangular reflector, the second triangular reflector, and the third triangular reflector in sequence, and finally enters the monocular line array camera. The two light rays form two light bands in the line array camera. The monocular line array camera obtains an initial single-frame line array image containing two parallel light bands. The background and noise of the initial single-frame line array image are removed, and the column coordinates P of the parallel light band edge are extracted. 0 1, P 0 2, P 0 3, P 0 4, among which P 0 1 and P 0 2 are the column coordinates of the left and right edges of the light band formed by the parallel light A in the initial single-frame linear array image, P 0 3 and P 0 4 are the column coordinates of the left and right edges of the light band formed by the parallel light B in the initial single-frame linear array image. Calculate the distance S between the two parallel light bands A and B, as well as the width W of the two parallel light bands A and B when the rotating body is not placed. S = P 0 3 – P 0 2. W = P 0 4-P 0 1-S; (2) Single diameter measurement: The mobile platform moves, and the measured rotating body is placed in the area surrounded by the mobile platform, ball screw and slide rail. The measured rotating body blocks part of the parallel light A and part of the parallel light B between the first triangular reflector and the second triangular reflector. The grating scale reading is G. The two blocked light rays also form two light bands in the linear array camera. The monocular linear array camera obtains a single-frame linear array image containing two parallel light bands after blocking. The background and noise of the single-frame linear array image after blocking are removed, and the bilateral centroid method is used to extract the column coordinates P of the parallel light band edge. 1 1, P 1 2, P 1 3, P 1 4, among which P 1 1 and P 1 2 are the column coordinates of the left and right edges of the light band formed by the parallel light A in the single frame linear array image after occlusion, P 1 3 and P 1 4 are the column coordinates of the left and right edges of the light band formed by the parallel light B in the single frame linear array image after occlusion. Calculate the width T1 of the parallel light A and the width T2 of the parallel light band B after occlusion when the measured rotating body is placed. T1= P 1 2-P 1 1, T2 = P 1 4-P 1 3. T=T1+T2. Calculate the diameter D of the rotating body to be measured according to the following formula: D = S + G + WT; (3) Diameter measurement in one cycle: The rotating body to be measured is fixed on a turntable, which drives the rotating body to be measured to rotate one cycle. The monocular linear array camera collects the spatiotemporal image of the rotating body to be measured in real time. Each frame of the spatiotemporal image is a line, which is a single-frame linear array image. According to step (2), the diameter corresponding to each line of the single-frame linear array image is calculated, and the diameter of the rotating body to be measured in one cycle is obtained. A monocular vision large-scale rotating body diameter dynamic measuring device is used for measurement, including a parallel line light source, a spectrometer, a first triangular reflector, a second triangular reflector, a third triangular reflector, a monocular line array camera, a fixing device, a servo motor, a ball screw, a mobile platform, a slide rail and a grating ruler. The ball screw and the slide rail are respectively fixed on the left and right sides of the fixing device. The ball screw and the slide rail are arranged in parallel longitudinally. One end of the transversely arranged mobile platform is connected to the nut of the ball screw, and the other end is slidably connected to the slide rail. The output shaft of the servo motor is connected to the lead screw of the ball screw. The parallel line light source is fixed on one side of the servo motor, the spectrometer is fixed on the back side of the parallel line light source, and the first The triangular reflector and the second triangular reflector are fixed on the left and right ends of the mobile platform respectively. The monocular line array camera is fixed on the front end of the slide rail. The third triangular reflector is set on one side of the monocular line array camera. The parallel line light source emits parallel light, which is divided into two beams of parallel light in orthogonal directions by the beam splitter. One beam of parallel light A is the 90° reflected light of the beam splitter, and the other beam of parallel light B is the transmitted light of the beam splitter. The parallel light A is directly incident on the monocular line array camera, and the parallel light B is reflected by the first triangular reflector, the second triangular reflector and the third triangular reflector in sequence, and finally enters the monocular line array camera. The grating ruler is used to obtain the distance between the beam splitter and the first triangular reflector.

2. The monocular vision large-scale rotating object diameter dynamic measurement method according to claim 1 is characterized in that: Step (1) is as follows: (101) The threshold method is used to remove the background and noise of the initial single-frame linear array image, and only the information of the parallel light bands is retained: (2) Where, is the grayscale value of the initial single-frame linear image after removing background and noise, threshold is the threshold for background removal, is the grayscale value of the initial single-frame linear array image; (102) Using the local edge search algorithm, the four edge column coordinates P of the parallel light band of the initial single-frame linear array image after processing in step (101) are obtained. 0 1, P 0 2, P 0 3, P 0 4. Calculate the distance S between the two parallel light bands A and B, as well as the width W of the two parallel light bands A and B when the rotating object to be measured is not placed.

3. The monocular vision large-scale rotating body diameter dynamic measurement method according to claim 1 is characterized in that: Step (2) is as follows: (201) The threshold method is used to remove the background and noise of the single-frame linear array image after occlusion, and only the information of the parallel light bands is retained: ； Where, is the grayscale value of a single frame linear array image after removing the background and noise, threshold is the threshold for background removal, is the grayscale value of a single frame linear array image after occlusion; (202) The bilateral centroid method is used to identify the transition center of the edge and the bilateral centroid method is used to extract the column coordinates P of the light band edge. 1 2, P 1 3, and P 1 1 = P 0 1, P 1 4= P 0 4; Where Lc is the center column coordinate of the dark area between the two parallel light bands, m is the column coordinate in each row, is the grayscale value of the mth column coordinate after processing in step (201).

4. The monocular vision large-scale rotating object diameter dynamic measurement method according to claim 1 is characterized in that: Based on the diameter obtained in step (2) or (3), the radial end runout of the rotating body under test can be detected.

5. The monocular vision large-scale rotating object diameter dynamic measurement method according to claim 1 is characterized in that: Based on the diameter of the rotating body obtained in step (3) and the rotation angle of the rotating body, the outer contour of the rotating body to be measured can be determined.

6. The monocular vision large-scale rotating object diameter dynamic measurement method according to claim 1 is characterized in that: The grating ruler includes a matched scale grating and an indicator grating. The scale grating is fixed along the length direction of the ball screw, and the indicator grating is fixed on one end of the moving platform close to the scale grating.

7. The monocular vision large-scale rotating object diameter dynamic measurement method according to claim 1 is characterized in that: It also includes a first sliding rod, a second sliding rod, a first fixed block, a first slider, a second fixed block and a second slider, the first sliding rod is arranged above the ball screw, the second sliding rod is arranged above the slide rail, the first sliding rod and the second sliding rod are arranged in parallel, one end of the first sliding rod is fixed on the first fixed block, the other end of the first sliding rod passes through the through hole on the first slider and the first slider slides along the first sliding rod, the spectrometer is fixed on the first fixed block, the first triangular reflector is fixed on the first slider, one end of the second sliding rod is fixed on the second fixed block, the other end of the second sliding rod passes through the through hole on the second slider and the second slider slides along the second sliding rod, the second triangular reflector is fixed on the second slider, and the third triangular reflector is fixed on the second fixed block.

8. The monocular vision large-scale rotating object diameter dynamic measurement method according to claim 1 is characterized in that: The fixing device is a mechanical frame, the longitudinal length of which is H and the transverse width is L, L>H, and the diameter measurement range of the measured rotating body is S~H.

9. The monocular vision large-scale rotating object diameter dynamic measurement method according to claim 1 is characterized in that: It also includes a turntable and an encoder. The measured rotating body is fixed on the turntable. The encoder is set on the turntable. The turntable drives the measured rotating body to rotate. The encoder is used to measure the rotation angle of the rotating body. The outer contour of the measured rotating body can be determined based on the diameter and rotation angle.

Images