A fast measurement method for the deviation of the reference hole of beam parts

Abstract

The present invention provides a rapid measurement method for the deviation of reference holes of beam parts, which belongs to structured light three-dimensional scanning measurement and analysis. The present invention is based on the secondary developed three-dimensional structured light scanning equipment and a special hole measurement adapter. By analyzing the principles and characteristics of the three-dimensional structured light scanning equipment, the texture image is extracted, and the ceramic ball proportional template based on shape matching is used to realize automatic and rapid positioning of the ceramic ball; random sampling and fitting of ceramic balls are used to realize the precise positioning of ceramic balls in three-dimensional space; finally, rapid and accurate detection of part hole spacing and hole axis deviation is realized. Compared with the special fixture detection method, the present invention saves the need for special fixtures, and the measurement results are quantifiable. In addition, compared with the structured light scanning measurement method, the present invention significantly improves the speed of post-measurement data processing and analysis.

Description

Technical Field

[0001] The invention belongs to the field of structured light three-dimensional scanning measurement and analysis, and relates to a rapid measurement method for reference hole deviations of beam parts, and specifically to a rapid measurement and analysis method for reference hole spacing and hole axis deviations of beam parts. Background Art

[0002] The main load-bearing components of aircraft parts can be divided into beams, frames, ribs, wall panels, long girders, joints and other types. With the vigorous development of the domestic aviation industry, the development trend of aircraft parts is gradually developing towards high precision, large size and complexity. Beam parts are mostly double-sided cavities with thin walls. When CNC machining beam parts, two holes and one side are usually used as the reference. However, beam parts are prone to deformation during heat treatment, and the deformation is mainly warping and lateral bending, supplemented by twisting. Deformation not only easily makes the shape and position tolerances of beam parts exceed the requirements of the drawings, but also causes certain difficulties in ensuring the thickness of the flange and web during internal processing. Whether the hole spacing and hole axis deviation of the reference holes of beam parts after heat treatment exceed the tolerance range is particularly important for the subsequent CNC machining of beam parts. At present, the inspection fixture method is usually used for inspection of beam parts after heat treatment. The two positioning holes of the parts can completely pass through the positioning pins of the inspection fixture to determine whether the deformation of the reference hole meets the requirements. There are problems such as inaccurate measurement, inability to quantify, and the need for special inspection fixtures. The conventional structured light scanning measurement method can quickly obtain the three-dimensional point cloud of the measured part. However, obtaining the hole position and hole axis data from the point cloud requires many manual steps. Inaccurate data selection can easily lead to large deviations in the results, and the measurement results cannot be obtained quickly and automatically. Summary of the invention

[0003] In order to solve the problems that the inspection of the reference hole spacing and hole axis deviation of existing beam parts after heat treatment requires a special inspection fixture, the fixture needs to be calibrated regularly and wastes resources, and the subsequent point cloud processing operation using a digital measurement method is cumbersome, the present invention provides a rapid measurement method for the reference hole deviation of beam parts, and uses a secondary developed three-dimensional structured light scanning device and a special hole position measurement double ball adapter to perform three-dimensional scanning measurement on beam parts. The measurement results of the secondary developed three-dimensional structured light scanning device include: the texture image, depth image, transformation coefficient of the depth image to the three-dimensional point cloud and the splicing point set for each measurement, and the texture image and the depth image pixels correspond one to one. The double balls of the hole position measurement double ball adapter use matte white ceramic balls, and the support uses matte black. By establishing a ceramic ball matching template based on the proportional shape in the texture image space, the ceramic ball area can be quickly located in the texture image. The ceramic ball area is mapped and transformed into a three-dimensional point cloud, and the points inside the sphere are extracted by a random sampling consistency method, so as to realize the rapid extraction of the ceramic ball point cloud and the evaluation of its spherical center coordinates and radius. By presetting a black background on the measuring platform and the support of the double-ball adapter for hole position measurement, and performing threshold extraction in the texture image, the background and the double-ball adapter area for hole position measurement can be quickly removed to obtain the measured part area, and then the part point cloud can be obtained. The part point cloud is spliced ​​through the identification points in each measurement result, and the ceramic ball point cloud and the ball center coordinates are transformed to the global coordinate system after splicing. The part point cloud and the part digital model are best fitted, and the part digital model and the digital model of the double-ball adapter for hole position measurement are combined into the point cloud coordinate system. By comparing the distance between the centers of the ceramic balls, the point cloud of the same ball is merged, and the ceramic balls are randomly sampled and fitted; the two balls of the double-ball adapter for hole position measurement are identified; the ceramic balls of the double-ball adapter for hole position measurement in the point cloud and the digital model are paired one by one. The least squares method is used to fit the corresponding distance between the ceramic balls in the digital model and the point cloud, and then the theoretical pose and measured pose of the double-ball adapter for hole position measurement in the digital model are obtained. Finally, the position of the hole center and the hole axis is obtained by the correspondence between the hole position measurement double ball adapter and the hole center and the hole axis, and then the hole spacing and axis deviation angle are calculated. Compared with the special fixture detection method, this beam positioning hole spacing and axial deviation detection method saves the need for special fixtures, and the measurement results can be quantified; compared with the structured light scanning measurement method, it significantly improves the speed of post-measurement data processing and analysis. A fast measurement method for the deviation of the reference hole of beam parts, first, the beam parts are scanned and measured in three dimensions using the secondary developed three-dimensional structured light scanning equipment and the special hole position measurement double ball adapter.

[0004] The technical solution adopted by the present invention is:

[0005] A method for quickly measuring the deviation of a reference hole of a beam part, the specific steps of the method are as follows:

[0006] Step 1: Use the secondary developed three-dimensional structured light scanning device to scan the matte ceramic ball and establish a ceramic ball shape template in the texture image space.

[0007] Step 2: Place the part 5 to be measured on the measuring platform.

[0008] Step 3: insert the hole position measurement double-ball adapter 4 into the hole to be measured of the measured part 5.

[0009] Step 4, pasting the measurement joint points at appropriate positions on the measured part 5 or the measurement platform.

[0010] Step 5: Use the secondary developed 3D structured light scanning device to measure and save the measurement results.

[0011] Step 6: Analyze and process each measurement result.

[0012] Step 7, traverse each measurement result and execute step 6.

[0013] Step 8: According to the splicing point set in each measurement result in step 5, the part point cloud in each measurement result is transformed into the coordinate system of the three-dimensional structured light scanning device to form a part point cloud set.

[0014] Step 9: Optimally fit the part digital model and part point cloud.

[0015] Step 10: merge the point clouds of the same ceramic ball.

[0016] Step 11, identifying the ceramic ball point cloud on the double-ball adapter 4 for measuring the same hole position.

[0017] Step 12, calculating the correspondence between the double-ball adapter 4 for hole position measurement in the ceramic ball point cloud and the double-ball adapter 4 for hole position measurement in the part digital model.

[0018] Step 13, using the least squares method to fit the part digital model and the corresponding ceramic ball in the ceramic ball point cloud, to obtain the optimal posture of the hole position measurement double-ball adapter 4 in the part digital model.

[0019] Step 14, convert and calculate the hole center distance deviation and hole axis angle deviation.

[0020] Furthermore, the hole position measurement double-ball adapter 4 includes a top matte ceramic ball 1, a bottom matte ceramic ball 2 and a support 3; the support 3 includes a cylinder and a cone, the bottom diameter of the cone is larger than the diameter of the cylinder, wherein the cylinder is inserted into the positioning hole to be measured of the measured part 5; one end of the bottom matte ceramic ball 2 is fixedly connected to the top of the conical end of the support 3, and the other end is connected to the top matte ceramic ball 1 through an axis, wherein the double balls are matte white ceramic balls, and the support 3 is matte black.

[0021] Furthermore, the secondary developed three-dimensional structured light scanning device software upgrade has measurement results including: texture image, depth image, transformation coefficients of depth image to three-dimensional point cloud and set of splicing points for each measurement, and one-to-one correspondence between texture image and depth image pixels.

[0022] Furthermore, the above specific steps are as follows:

[0023] The step 1 is to establish a matching template SMT based on the proportion shape of the matte ceramic ball in the texture image space. ball The matte ceramic ball is scanned by a secondary developed three-dimensional structured light scanning device to obtain a texture image of the ceramic ball; the texture image of the ceramic ball is Gaussian filtered to obtain a filtered image G ; in image G In the image, a square area is manually created with the center of the ceramic ball as the center, so that the distance from the edge of the square area to the edge of the ceramic ball is about 12 pixels; G The image of the square area is cut out as the template input image, and the matching template SMT based on the proportional shape is established ball .

[0024] In step 2, the measured part 5 is placed in a free state on a measuring platform with a matte black background.

[0025] In the step 3, the hole position measuring double ball adapter 4 is respectively inserted into a plurality of positioning holes of the measured part 5 whose positioning hole deviations are to be measured.

[0026] In step 4, the measuring joints are pasted at appropriate positions on the measured part 5 or the measuring platform.

[0027] The step 5 uses the secondary developed three-dimensional structured light scanning device to measure the measured part 5 multiple times, and each measurement obtains the measurement result MR from the secondary developed three-dimensional structured light scanning device. i , where i = 1, 2, 3, .... represents the i-th measurement result, MR i Contains the texture image imageT measured at the i-th time i , depth image imageD i , the scale factor xScale of the depth image to 3D point cloud transformation i ,yScale i 、zScale i , splicing mark point set PtSet i , where the texture image and the depth image pixels correspond one to one; all measurement results form a measurement result set MRs = {MR i , i = 1, 2, 3...}; depth image imageD i Pixels in theij , the gray value is grayVal ij , where i and j are pixel points ij The y and x coordinates in the image coordinate system are mapped to three-dimensional space points according to the following formula.

[0028] pt x =xScale×j

[0029] pt y =yScale×i

[0030] pt z =zScale×grayVal ij

[0031] In the formula, pt x ,pt y and pt z Pixel ij Corresponding to the x, y, z coordinates of the point in the 3D point cloud.

[0032] Step 6 extracts MR from the measurement result set MRs i , for a single measurement result MR i Data analysis and processing, the specific steps are as follows:

[0033] Step 6.1, in MR i Texture image imageT i In the example, the template matching method is used to obtain the matching result of each ceramic ball matchedB. j ; Use proportional shape-based matching template SMT ball Match texture image imageT i , match the texture image imageT i The ceramic ball on the double-ball adapter 4 is measured at the center hole position, and the matching result is matchedB. j , where j=1, 2, 3... represents the jth ceramic ball data that is successfully matched.

[0034] Step 6.2, in matchedB j Get the position coordinates of the ceramic ball (x j ,y j ) and the radius of the ceramic ball (r j ), respectively (x j ,y j ) is the origin, r j Create a circle with the radius as the jth ceramic ball region BRegion. J .

[0035] Step 6.3, calculate and extract the ceramic ball point cloud BCloudj and store it. i Extract the ceramic ball region BRegion J The coordinates of the inner pixel points are transformed into the ceramic ball point cloud BCloudj according to the mapping transformation relationship between the image coordinates and the three-dimensional point cloud coordinates in step 5. The random sampling method (RSAC) is used to fit the sphere to BCloudj, and the points in the ceramic ball point cloud whose distance from the spherical surface is greater than the threshold 0.02mm are deleted.

[0036] Step 6.4, identify each ceramic ball region BRegion according to the above steps 6.1, 6.2 and 6.3 J The ceramic ball point cloud is stored in the i-th measurement result MR i middle.

[0037] Step 6.5, in the texture image imageT i The effective measurement area of ​​the part is obtained in the texture image imageT. i The remaining area after removing the measurement platform area and the hole position measurement double ball adapter 4 area; in the texture image imageT i The gray threshold method is used to extract the measurement platform area PRegion i , hole position measurement double ball adapter 4 area ARegion i , from the image overall area texture image imageT i Subtract PRegion i and ARegion i Get the measured part area partRegion i .

[0038] Step 6.6, identify the part point cloud and store it. According to the mapping transformation relationship between the image coordinates and the three-dimensional point cloud coordinates in step 5, the part region to be measured partRegion i , mapping transformation to part point cloud partCloud i , and store it in the i-th measurement result MR i middle.

[0039] Step 7 traverses each measurement result MR in MRs i , proceed to step 6.

[0040] The step 8 is based on each measurement result MR in step 5. i The splicing mark point set PtSet i, transform the part point cloud in each measurement result to the global coordinate system. Extract the splicing points in each measurement result, use the IPC analysis method to calculate the transformation matrix TGlobal from the part point cloud of each measurement result to the global coordinate system after splicing i , i represents the transformation matrix from the part point cloud of the i-th measurement result to the global coordinate system, and the ceramic ball point cloud BCloudj and the part point cloud partCloud in each measurement result are transformed into i Multiply them with the coordinate transformation matrix respectively to get the ceramic ball point cloud set BCloudSet in the global coordinate system G and part point cloud set PCloudSet G .

[0041] Step 9 optimally fits the part digital model and the part point cloud. G Merge into the overall part point cloud partCloud; use the optimal fitting method to align the part digital model with the part point cloud to obtain the coordinate transformation matrix T from the part digital model to the part point cloud PC , using the coordinate transformation matrix T PC The coordinates of the center point of the ceramic ball of the double-ball adapter 4 for hole position measurement in the part digital model are transformed into the global coordinate system of the point cloud.

[0042] Step 10 described above combines and merges the same ceramic ball point cloud. Calculate the ceramic ball point cloud set BCloudSet G The distance between the centers of any two ceramic ball point clouds is fitted. When the distance between the two ball centers is less than 0.2 mm, the two ceramic ball point clouds are merged to form a new BCloudSet G.

[0043] The step 11 identifies the ceramic ball point cloud on the double ball adapter 4 for measuring the same hole position. Calculate BCloudSet G The distance dis between the ball centers of any two ceramic ball point clouds is fitted. When |dis-disBall|<0.5mm, the two balls are taken as the balls on the double-ball adapter 4 measured at the same hole position, where disBall refers to the distance between the ball centers of the double-ball adapter 4 measured at the same hole position in the part digital model.

[0044] The step 12 calculates the correspondence between the hole position measurement double ball adapter 4 in the ceramic ball point cloud and the hole position measurement double ball adapter 4 in the part digital model. The sum of the distances from the center point of each hole position measurement double ball adapter 4 in the part digital model to the center point of the hole position measurement double ball adapter 4 in the ceramic ball point cloud is calculated, and the two hole position measurement double ball adapters 4 corresponding to the minimum distance and the minimum distance correspond, and the center points correspond; according to the above method, the correspondence between each hole position measurement double ball adapter 4 in the ceramic ball point cloud and the corresponding hole position measurement double ball adapter 4 and ceramic ball in the part digital model is found.

[0045] The step 13 adopts the least square method to fit the part digital model and the corresponding ceramic ball in the ceramic ball point cloud, and then obtains the theoretical pose and measured pose of the hole position measurement double-ball adapter 4 in the part digital model.

[0046] The step 14 converts and calculates the hole center distance deviation and the hole axis angle deviation. According to the corresponding relationship between the hole position measurement double ball adapter 4 and the hole position and hole axis, the hole position and axis of each hole are calculated, and then the position deviation and axis angle between holes are calculated.

[0047] The beneficial effects of the present invention are as follows: compared with the special fixture detection method, the present invention saves the need for a special fixture, and the measurement result can be quantified; in addition, compared with the structured light scanning measurement method, the present invention significantly improves the speed of post-measurement data processing and analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a simplified diagram of the double-ball adapter for hole position measurement.

[0049] Figure 2 The following is a simplified diagram of the part being tested.

[0050] In the figure: 1 top matte ceramic ball; 2 bottom matte ceramic ball; 3 support; 4 hole measurement double ball adapter; 5 measured part. DETAILED DESCRIPTION

[0051] The present invention is based on the secondary developed three-dimensional structured light scanning equipment and the special hole position measurement adapter 4. By analyzing the principles and characteristics of the three-dimensional structured light scanning equipment, the texture image is extracted, and the ceramic ball proportion template based on shape matching is used to realize automatic and rapid positioning of the ceramic ball; random sampling and fitting of the ceramic ball is used to realize the precise positioning of the ceramic ball in three-dimensional space; and finally a method for rapid and precise detection of the hole spacing and hole axis deviation of the part is realized.

[0052] A method for quickly measuring the deviation of a reference hole of a beam part, the specific steps of the method are as follows:

[0053] Step 1: Scan the matte ceramic ball with the secondary developed 3D structured light scanning device, and establish the ceramic ball shape template in the texture image space, wherein the secondary developed 3D structured light scanning device software is upgraded, and its measurement results include: texture image, depth image, transformation coefficients from depth image to 3D point cloud and splicing point set for each measurement, and the texture image and depth image pixels correspond one to one. Specifically:

[0054] The matte ceramic ball is scanned by a secondary developed three-dimensional structured light scanning device to obtain a texture image of the ceramic ball; the texture image of the ceramic ball is Gaussian filtered to obtain a filtered image G ; in imageG In the image, a square area is manually created with the center of the ceramic ball as the center, so that the distance from the edge of the square area to the edge of the ceramic ball is about 12 pixels; G The image of the square area is cut out as the template input image, and the matching template SMT based on the proportional shape is established ball .

[0055] Step 2: Place the part 5 to be tested in a free state on a measurement platform with a matte black background.

[0056] Step 3, inserting hole position measurement double-ball adapters 4 into multiple positioning holes of the positioning hole deviation to be measured on the measured part 5 respectively, wherein the hole position measurement double-ball adapter 4 includes a top matte ceramic ball 1, a bottom matte ceramic ball 2 and a support 3; the support 3 includes a cylinder and a cone, the bottom diameter of the cone is larger than the diameter of the cylinder, wherein the cylinder is inserted into the positioning hole to be measured of the measured part 5; one end of the bottom matte ceramic ball 2 is fixedly connected to the top of the conical end of the support 3, and the other end is connected to the top matte ceramic ball 1 through an axis, wherein the double balls are matte white ceramic balls, and the support 3 is matte black, such as Figure 1 and Figure 2 shown.

[0057] Step 4, pasting the measurement joint points at appropriate positions on the measured part 5 or the measurement platform.

[0058] Step 5: Use the secondary developed 3D structured light scanning device to measure and save the measurement results.

[0059] The second-developed three-dimensional structured light scanning device is used to measure the measured part 5 multiple times, and the measurement result MR is obtained from the second-developed three-dimensional structured light scanning device each time. i , where i = 1, 2, 3, .... represents the i-th measurement result, MR i Contains the texture image imageT measured at the i-th time i , depth image imageD i , the scale factor xScale of the depth image to 3D point cloud transformation i ,yScale i 、zScale i , splicing mark point set PtSet i , where the texture image and the depth image pixels correspond one to one; all measurement results form a measurement result set MRs = {MR i , i = 1, 2, 3...}; depth image imageD i Pixels in the ij , the gray value is grayVal ij, where i and j are pixels ij The y and x coordinates in the image coordinate system are mapped to three-dimensional space points according to the following formula.

[0060] pt x =xScale×j

[0061] pt y =yScale×i

[0062] pt z =zScale×grayVal ij

[0063] In the formula, pt x ,pt y and pt z Pixel ij Corresponding to the x, y, z coordinates of the point in the 3D point cloud.

[0064] Step 6: Extract MR from the measurement result set MRs i , for a single measurement result MR i Data analysis and processing, the specific steps are as follows:

[0065] Step 6.1, in MR i Texture image imageT i In the example, the template matching method is used to obtain the matching result of each ceramic ball matchedB. j ; Use proportional shape-based matching template SMT ball Match texture image imageT i , match the texture image imageT i The ceramic ball on the double-ball adapter 4 is measured at the center hole position, and the matching result is matchedB. j , where j=1, 2, 3... represents the jth ceramic ball data that is successfully matched.

[0066] Step 6.2, in matchedB j Get the position coordinates of the ceramic ball (x j ,y j ) and the radius of the ceramic ball (r j ), respectively (x j ,y j ) is the origin, r j Create a circle with the radius as the jth ceramic ball region BRegion. J .

[0067] Step 6.3, calculate and extract the ceramic ball point cloud BCloudj and store it. iExtract the ceramic ball region BRegion J The coordinates of the inner pixel points are transformed into the ceramic ball point cloud BCloudj according to the mapping transformation relationship between the image coordinates and the three-dimensional point cloud coordinates in step 5. The random sampling method (RSAC) is used to fit the sphere to BCloudj, and the points in the ceramic ball point cloud whose distance from the spherical surface is greater than the threshold 0.02mm are deleted.

[0068] Step 6.4, identify each ceramic ball region BRegion according to the above steps 6.1, 6.2 and 6.3 J The ceramic ball point cloud is stored in the i-th measurement result MR i middle.

[0069] Step 6.5, in the texture image imageT i The effective measurement area of ​​the part is obtained in the texture image imageT. i The remaining area after removing the measurement platform area and the hole position measurement double ball adapter 4 area; in the texture image imageT i The gray threshold method is used to extract the measurement platform area PRegion i , hole position measurement double ball adapter 4 area ARegion i , from the image overall area texture image imageT i Subtract PRegion i and ARegion i Get the measured part area partRegion i .

[0070] Step 6.6, identify the part point cloud and store it. According to the mapping transformation relationship between the image coordinates and the three-dimensional point cloud coordinates in step 5, the part region to be measured partRegion i , mapping transformation to part point cloud partCloud i , and store it in the i-th measurement result MR i middle.

[0071] Step 7, traverse each measurement result MR in MRs i , proceed to step 6.

[0072] Step 8: According to each measurement result MR i The splicing mark point set PtSet i , transform the part point cloud in each measurement result to the global coordinate system. Extract the splicing points in each measurement result, use the IPC analysis method to calculate the transformation matrix TGlobal from the part point cloud of each measurement result to the global coordinate system after splicing i, i represents the transformation matrix from the part point cloud of the i-th measurement result to the global coordinate system, and the ceramic ball point cloud BCloudj and the part point cloud partCloud in each measurement result are transformed into i Multiply them with the coordinate transformation matrix respectively to get the ceramic ball point cloud set BCloudSet in the global coordinate system G and part point cloud set PCloudSet G .

[0073] Step 9: Optimally fit the part model and the part point cloud. Set the part point cloud set PCloudSet G Merge into the overall part point cloud partCloud; use the optimal fitting method to align the part digital model with the part point cloud to obtain the coordinate transformation matrix T from the part digital model to the part point cloud PC , using the coordinate transformation matrix T PC The coordinates of the center point of the ceramic ball of the double-ball adapter 4 for hole position measurement in the part digital model are transformed into the global coordinate system of the point cloud.

[0074] Step 10: Merge the same ceramic ball point cloud. Calculate the ceramic ball point cloud set BCloudSet G The distance between the centers of any two ceramic ball point clouds is fitted. When the distance between the two ball centers is less than 0.2 mm, the two ceramic ball point clouds are merged to form a new BCloudSet G .

[0075] Step 11, identify the ceramic ball point cloud on the double ball adapter 4 for measuring the same hole position. Calculate BCloudSet G The distance dis between the ball centers of any two ceramic ball point clouds is fitted. When |dis-disBall|<0.5mm, the two balls are taken as the balls on the double-ball adapter 4 measured at the same hole position, where disBall refers to the distance between the ball centers of the double-ball adapter 4 measured at the same hole position in the part digital model.

[0076] Step 12, calculate the correspondence between the hole position measurement double ball adapter 4 in the ceramic ball point cloud and the hole position measurement double ball adapter 4 in the part digital model. Calculate the sum of the distances from the center point of each hole position measurement double ball adapter 4 in the part digital model to the center point of the hole position measurement double ball adapter 4 in the ceramic ball point cloud, and the two hole position measurement double ball adapters 4 corresponding to the minimum distance and the minimum distance correspond, and the center points correspond; find out the correspondence between each hole position measurement double ball adapter 4 in the ceramic ball point cloud and the corresponding hole position measurement double ball adapter 4 and ceramic ball in the part digital model according to the above method.

[0077] Step 13, using the least squares method to fit the part digital model and the corresponding ceramic ball in the ceramic ball point cloud, and then obtaining the theoretical pose and measured pose of the hole position measurement double-ball adapter 4 in the part digital model.

[0078] Step 14, convert and calculate the hole center distance deviation and hole axis angle deviation. According to the correspondence between the hole position measurement double ball adapter 4 and the hole position and hole axis, the hole position and axis of each hole are calculated, and then the position deviation and axis angle between holes are calculated.

[0079] The above-described embodiments merely express the implementation methods of the present invention, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A rapid measurement method for the deviation of the reference hole of beam parts. It is characterized in that The method comprises the following steps: Step 1: Use the secondary developed 3D structured light scanning device to scan the matte ceramic ball in the texture image space. Create a ceramic ball shape template; Step 2, placing the part to be measured (5) on the measuring platform; Step 3, inserting the hole position measurement double ball adapter (4) into the position hole to be measured of the measured part (5); Step 4, pasting the measuring joint points at appropriate positions on the measured part (5) or the measuring platform; Step 5, using the secondary developed 3D structured light scanning device to measure and save the measurement results; Step 6, analyzing and processing each measurement result; Step 7, traverse each measurement result and execute step 6; Step 8: According to the splicing point set in each measurement result of step 5, transform the part point cloud in each measurement result To the coordinate system of the 3D structured light scanning device to form a part point cloud set; Step 9, optimally fitting the part digital model and the part point cloud; Step 10, stitching and merging the point cloud of the same ceramic ball; Step 11, identifying the ceramic ball point cloud on the double-ball adapter (4) for measuring the same hole position; Step 12, calculate the hole position measurement double ball adapter (4) in the ceramic ball point cloud and the hole position measurement double ball adapter (4) in the part digital model. The corresponding relationship of the ball adapter (4); Step 13, using the least square method to fit the part digital model and the corresponding ceramic ball in the ceramic ball point cloud, to obtain the optimal posture of the hole position measurement double ball adapter (4) in the part digital model; Step 14, convert and calculate the hole center distance deviation and hole axis angle deviation.

2. A rapid measurement method for reference hole deviation of beam parts according to claim 1, It is characterized in that The hole position measurement double-ball adapter (4) comprises a top matte ceramic ball (1), a bottom matte ceramic ball (2) and a support (3); the support (3) comprises a cylinder and a cone, the bottom diameter of the cone is larger than the diameter of the cylinder, wherein the cylinder is inserted into the position hole to be measured of the measured part (5); one end of the bottom matte ceramic ball (2) is fixedly connected to the top of the conical end of the support (3), and the other end is connected to the top matte ceramic ball (1) through an axis, wherein the double balls are matte white ceramic balls, and the support (3) is matte black.

3. A rapid measurement method for reference hole deviation of beam parts according to claim 1, It is characterized in that The measurement results of the secondary developed three-dimensional structured light scanning device software upgrade include: the texture image, depth image, transformation coefficients of the depth image to the three-dimensional point cloud and the set of splicing points for each measurement, and the texture image and depth image pixels correspond one to one.

4. A method for quickly measuring the deviation of a reference hole of a beam part according to any one of claims 1 to 3, It is characterized in that The specific steps of the method are as follows: The step 1 is to establish a matching template SMT based on the proportion shape of the matte ceramic ball in the texture image space. ball ; Use the secondary developed three-dimensional structured light scanning device to scan the matte ceramic ball to obtain the texture image of the ceramic ball; Gaussian filter the texture image of the ceramic ball to obtain the filtered image G ; in image G In the image, a square area is manually created with the center of the ceramic ball as the center, so that the distance from the edge of the square area to the edge of the ceramic ball is 12 pixels; in the image G The image of the square area is cut out as the template input image, and the matching template SMT based on the proportional shape is established ball ; In step 2, the part to be tested (5) is placed in a free state on a measuring platform with a matte black background; In step 3, the double-ball adapters (4) for hole position measurement are respectively inserted into a plurality of positioning holes of the measured part (5) whose positioning hole deviations are to be measured; Step 4 includes pasting the measuring joint points at appropriate positions on the measured part (5) or the measuring platform; The step 5 uses the secondary developed three-dimensional structured light scanning device to measure the measured part (5) multiple times, and obtains the measurement result MR from the secondary developed three-dimensional structured light scanning device for each measurement. i , where i = 1, 2, 3, .... represents the i-th measurement result, MR i Contains the texture image imageT measured at the i-th time i , depth image imageD i , the scale factor xScale of the depth image to 3D point cloud transformation i ,yScale i 、zScale i , splicing mark point set PtSet i , where the texture image and the depth image pixels correspond one to one; all measurement results form a measurement result set MRs = {MR i , i = 1, 2, 3...}; depth image imageD i Pixels in the ij , the gray value is grayVal ij , where i and j are pixels ij The y and x coordinates in the image coordinate system are mapped to three-dimensional space points according to the following formula; pt x =xScale×j pt y = yScale × i pt z =zScale×grayVal ij In the formula, pt x ,pt y and pt z Pixel ij The x, y, and z coordinates of the corresponding points in the 3D point cloud; Step 6 extracts MR from the measurement result set MRs i , for a single measurement result MR i Data analysis and processing, the specific steps are as follows: Step 6.1, in the texture image imageT of MR i , the matching result matchedB of each ceramic ball is obtained by using the template matching method i ; The matching template SMT based on the proportional shape is used to match the texture image imageT j , and the ceramic balls on the hole position measuring double-ball adapter (4) in the texture image imageT are matched to obtain the matching result matchedB ball , where j = 1, 2, 3... represents the data of the jth ceramic ball with successful matching; i The texture image imageT i is matched, and the ceramic balls on the hole position measuring double-ball adapter (4) in the texture image imageT are obtained, and the matching result matchedB j is obtained, where j = 1, 2, 3... represents the data of the jth ceramic ball with successful matching; Step 6.2, in matchedB j Get the position coordinates of the ceramic ball (x j ,y j ) and the radius r of the ceramic ball j , respectively (x j ,y j ) is the origin, r j Create a circle with the radius as the jth ceramic ball region BRegion. J ; Step 6.3, calculate and extract the ceramic ball point cloud BCloudj and store it in the texture image imageT i Extract the ceramic ball region BRegion J The coordinates of the inner pixel points are transformed into the ceramic ball point cloud BCloudj according to the mapping transformation relationship between the image coordinates and the three-dimensional point cloud coordinates in step 5. The random sampling method RSAC is used to fit the sphere to BCloudj, and the points in the ceramic ball point cloud whose distance from the sphere surface is greater than the threshold value of 0.02mm are deleted; Step 6.4, identify each ceramic ball region BRegion according to the above steps 6.1, 6.2 and 6.3 J The ceramic ball point cloud is stored in the i-th measurement result MR i middle; Step 6.5, in the texture image imageT i The effective measurement area of ​​the part is obtained from the texture image imageT i The remaining area after removing the measurement platform area and the hole position measurement double ball adapter (4) area; in the texture image imageT i The gray threshold method is used to extract the measurement platform area PRegion i , hole position measurement double ball adapter (4) area ARegion i , from the image overall area texture image imageT i Subtract PRegion i and ARegion i Get the measured part area partRegion i ; Step 6.6, identify the part point cloud and store it; according to the mapping transformation relationship between the image coordinates and the three-dimensional point cloud coordinates in step 5, the part area to be tested partRegion i , mapping transformation to part point cloud partCloud i , and store it in the i-th measurement result MR i middle; Step 7 traverses each measurement result MR in MRs i , proceed to step 6; The step 8 is based on each measurement result MR in step 5. i The splicing mark point set PtSet i , transform the part point cloud in each measurement result to the global coordinate system; extract the splicing points in each measurement result, and use the IPC analysis method to calculate the transformation matrix TGlobal from the part point cloud of each measurement result to the global coordinate system after splicing i , i represents the transformation matrix from the part point cloud of the i-th measurement result to the global coordinate system, and the ceramic ball point cloud BCloudj and the part point cloud partCloud in each measurement result are transformed into i Multiply them with the coordinate transformation matrix respectively to get the ceramic ball point cloud set BCloudSet in the global coordinate system G and part point cloud set PCloudSet G ; Step 9 above sets the part point cloud set PCloudSet G Merge into the overall part point cloud partCloud; use the optimal fitting method to align the part digital model with the part point cloud to obtain the coordinate transformation matrix T from the part digital model to the part point cloud PC , using the coordinate transformation matrix T PC Transforming the coordinates of the center point of the ceramic ball of the double-ball adapter (4) for hole position measurement in the digital model of the part into the global coordinate system of the point cloud; Step 10 calculates the ceramic ball point cloud set BCloudSet G The distance between the centers of any two ceramic ball point clouds is fitted. When the distance between the two ball centers is less than 0.2 mm, the two ceramic ball point clouds are merged to form a new BCloudSet G ; Step 11 above calculates BCloudSet G The distance dis between the ball centers of any two ceramic ball point clouds is fitted. When |dis-disBall|<0.5mm, the two balls are taken as the balls on the double ball adapter (4) measured at the same hole position, where disBall refers to the distance between the ball centers of the double ball adapter (4) measured at the same hole position in the part digital model; The step 12 calculates the sum of the distances from the center point of each hole position measurement double ball adapter (4) in the part digital model to the center point of the hole position measurement double ball adapter (4) in the ceramic ball point cloud, and the two hole position measurement double ball adapters (4) corresponding to the minimum distance and the minimum distance correspond, and the center points of the balls correspond; according to the above method, find out the corresponding relationship between the hole position measurement double ball adapter (4) and the ceramic ball corresponding to each hole position measurement double ball adapter (4) in the ceramic ball point cloud in the part digital model; The step 13 adopts the least square method to fit the part digital model and the corresponding ceramic balls in the ceramic ball point cloud, thereby obtaining the theoretical position and measured position of the hole position measurement double ball adapter (4) in the part digital model; The step 14 converts and calculates the hole center distance deviation and the hole axis angle deviation; based on the corresponding relationship between the hole position measurement double ball adapter (4) and the hole position and hole axis, the hole position and axis of each hole are calculated, and then the position deviation and axis angle between the holes are calculated.

Images