CN115471536A - Flexible rubber workpiece 3D wall thickness extraction method based on SAD (sum of absolute differences) stereo matching and high-precision equipment cooperation - Google Patents
Flexible rubber workpiece 3D wall thickness extraction method based on SAD (sum of absolute differences) stereo matching and high-precision equipment cooperation Download PDFInfo
- Publication number
- CN115471536A CN115471536A CN202110650600.2A CN202110650600A CN115471536A CN 115471536 A CN115471536 A CN 115471536A CN 202110650600 A CN202110650600 A CN 202110650600A CN 115471536 A CN115471536 A CN 115471536A
- Authority
- CN
- China
- Prior art keywords
- wall thickness
- flexible rubber
- sad
- precision
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 15
- 238000000034 method Methods 0.000 claims abstract description 47
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000012937 correction Methods 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000003708 edge detection Methods 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 14
- 238000005259 measurement Methods 0.000 abstract description 14
- 239000000463 material Substances 0.000 abstract description 4
- 238000000691 measurement method Methods 0.000 abstract description 4
- 238000012545 processing Methods 0.000 abstract description 3
- 238000011160 research Methods 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000009434 installation Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 239000004033 plastic Substances 0.000 description 3
- 201000005625 Neuroleptic malignant syndrome Diseases 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005266 casting Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
- G06T2207/10012—Stereo images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Image Processing (AREA)
Abstract
A3D wall thickness extraction method for a flexible rubber workpiece based on SAD stereo matching and high-precision equipment cooperation. The invention provides a 3D wall thickness extraction method of a flexible rubber workpiece based on SAD (sum of absolute differences) stereo matching and high-precision equipment cooperation, aiming at solving the problem of measuring the wall thickness of a flexible rubber product with high precision. The method comprises the steps of 1, calibrating an industrial camera; 2. obtaining an image to be detected; 3. preprocessing an image; 4. extracting edges; 5. SAD stereo matching; 6. based on image characteristics, collaboratively splicing with high-precision equipment; 7. 3D, solving the characteristic wall thickness; 8. the robustness and the numerical value consistency of the test measurement method are analyzed. The invention is applied to the field of flexible material vision measurement by combining high-precision equipment control and an image processing algorithm, realizes high-precision shooting and splicing detection of workpieces of various types, overcomes the problems of large error, poor robustness of a common measurement method facing the workpieces of various types and the like of the conventional manual measurement, and has high accuracy and high speed.
Description
Technical Field
The invention relates to a flexible rubber workpiece 3D wall thickness extraction method based on SAD stereo matching and high-precision equipment cooperation.
Background
With the continuous development of social economy, the competition of the automobile market is intensified, and in the production of modern automobile parts, the number of parts is large, and the quality of each part of the automobile indirectly determines the quality credit of the automobile. The plastic dust-proof cover is an important part of the automobile parts, the quality of the plastic dust-proof cover also influences the output order of the parts, and the quality detection process is an important process in the production of the workpiece. The rubber dustproof sleeve is used as a corrugated pipe-shaped workpiece cast by a flexible material, and the quality detection process of the product is difficult to realize automation due to the special product shape and material. The traditional measuring method is that after the casting of a workpiece is finished, the workpiece is cut along a bus, and a vernier caliper is used for manual contact measurement on a tangent plane, so that the process is complicated and time-consuming, and manual measurement easily causes the problems that a measuring point is inaccurate and the measuring number is not right. Therefore, the demand for the automatic measurement of the plastic dust-proof sleeve is increasing.
The development of foreign computer vision technology is very rapid, and since the mid-sixties of the last century, the research on computer vision technology has been increasing, especially in the field of binocular stereo vision research, and various novel methods for stereo vision research have been proposed and used. The conventional two-dimensional image research is popularized to a three-dimensional scene for the first time by Robert of the American Massachusetts institute of technology, and the three-dimensional world analysis and research on pictures are carried out by a computer for the first time, which is the first river of stereoscopic vision research. Since then, more and more scholars are engaged in the research on the field of stereoscopic vision, which promotes the development of stereoscopic vision technology to some extent. Compared with foreign countries, although the research of the computer vision industry is relatively late in China, especially the research of binocular vision technology, in recent years, a plurality of scientific researchers are put into the fields and relatively advanced results are obtained.
In recent years, binocular vision systems are applied more and more widely to the field of measurement, but the application of binocular vision to a flexible target object with a short distance and a small size is less at present. Based on the method, special research is carried out on measuring the wall thickness of the flexible rubber workpiece by using a binocular stereo vision technology.
Disclosure of Invention
The invention combines high-precision equipment control and an image processing algorithm to be applied to the field of flexible material vision measurement, overcomes the problems of large error, poor robustness faced to common measurement methods of various types of workpieces and the like of the conventional manual measurement, has high accuracy and high speed, greatly improves the measurement precision and the measurement efficiency, and provides the 3D wall thickness extraction method of the flexible rubber workpiece based on SAD stereo matching and high-precision equipment cooperation.
The above-mentioned invention purpose is realized through the following technical scheme:
firstly, a dot calibration plate is adopted to calibrate and correct distortion of a plurality of images, epipolar line correction is carried out, and internal and external parameters and relative poses of a binocular camera are obtained;
automatically calculating the advancing track of the high-precision sliding table according to the length of the workpiece, and shooting the measured workpiece in a segmented manner by using a binocular camera to obtain an image;
step three, preprocessing the image to obtain an ROI;
step four, utilizing an edge extraction algorithm to extract edges of the obtained ROI;
finding out corresponding points of the characteristics of the left camera and the right camera by using an SAD algorithm;
splicing based on image characteristics, and verifying by combining motion trail analysis of a high-precision sliding table;
step seven, solving the 3D wall thickness of the characteristic position by utilizing the edge profile;
and step eight, checking the robustness of the measuring method and checking whether the measuring result meets the requirement of consistency.
The invention has the following effects:
the invention uses SAD stereo matching and feature registration to carry out binocular vision detection. The difficulty of the invention mainly comprises the cooperative splicing of high-precision equipment and the measurement adaptability of various flexible rubber workpieces, and has the following advantages:
1. the invention does not rely on accurate image coordinates: the detection method takes the self characteristics of the workpiece as a reference, extracts the edge, matches the characteristics of SAD stereo, and carries out splicing and verification based on the cooperation of image characteristics and high-precision equipment. The algorithm stability is guaranteed, and even if a certain error exists in the installation of a mechanical structure of the detection system and a certain deviation occurs in the acquired image, the spliced image can be obtained stably and the three-dimensional thickness of the workpiece can be calculated.
2. The accuracy is high: due to the fact that the non-standard flexible rubber original piece is prone to deformation, the extreme value is obtained by adopting a traversal algorithm when the 3D wall thickness of the feature point is calculated, and the actual wall thickness value is obtained. Meanwhile, the invention combines SAD stereo matching and high-precision equipment collaborative splicing, and ensures the accuracy of detection, as shown in Table 1.
3. The algorithm has strong robustness: by detecting the characteristic positions of the flexible rubber workpieces, automatically positioning the splicing positions and controlling the high-precision sliding table to check, the automatic splicing and calculation of the 3D wall thickness of all the characteristic positions of the workpieces of different models and images are realized. The influence caused by errors in installation and operation of the mechanical mechanism is completely eliminated. The flexibility of the method ensures the convenience of updating and maintaining the visual detection system.
Drawings
FIG. 1 is a flow chart of a 3D wall thickness extraction method of a flexible rubber workpiece based on SAD stereo matching and high-precision equipment cooperation;
fig. 2 is a schematic structural diagram of an apparatus according to a first embodiment, in which fig. 1 shows a high-precision slide table; 2. a servo motor; 3. a support; 4. a cylinder; 5. a clamp; 6. a binocular camera system;
FIG. 3 is a partial image of a workpiece taken by a camera according to one embodiment;
FIG. 4 is a partial image of a workpiece taken by a camera in different regions according to one embodiment;
FIG. 5 is a partial image of a workpiece taken by a camera according to one embodiment;
FIG. 6 is a partial ROI acquired using image pre-processing as set forth in one embodiment;
FIG. 7 is a diagram of XLD extracted using the Canny operator, according to an embodiment;
FIG. 8 is a complete workpiece image based on image feature stitching as set forth in one embodiment;
FIG. 9 is a partial image of a workpiece captured by a camera in different regions according to one embodiment;
FIG. 10 is a partial image of a workpiece taken by a camera according to one embodiment;
FIG. 11 is a partial image of a workpiece captured by a camera according to one embodiment;
FIG. 12 is a partial image of a workpiece taken by a camera according to one embodiment;
FIG. 13 is a partial ROI acquired using image pre-processing as set forth in one embodiment;
FIG. 14 is a diagram of XLD extracted using the Canny operator, according to an embodiment;
FIG. 15 is a diagram illustrating a complete workpiece image based on image feature stitching according to an exemplary embodiment;
Detailed Description
The first specific implementation way is as follows: the method for extracting the 3D wall thickness of the flexible rubber workpiece based on SAD stereo matching and high-precision equipment cooperation is specifically prepared according to the following steps:
firstly, calibrating and correcting distortion of a plurality of images by adopting a dot calibration plate to obtain internal and external parameters and relative poses of a binocular camera;
step two, automatically calculating the advancing track of the high-precision sliding table according to the length of the workpiece, and shooting the measured workpiece by the binocular camera in a segmented manner to obtain images as shown in fig. 3, 4 and 5;
step three, acquiring the ROI by utilizing preprocessing modes such as binaryzation, smooth filtering, image enhancement, opening and closing operation and the like, as shown in figure 6;
step four, utilizing a Canny algorithm to extract edges of the obtained ROI, as shown in figure 7;
finding out corresponding points of the characteristics of the left camera and the right camera by utilizing an SAD algorithm;
splicing based on image characteristics, and verifying by combining motion trail analysis of a high-precision sliding table, as shown in fig. 8;
step seven, solving the 3D wall thickness of the characteristic position by utilizing the edge profile;
step eight, checking the robustness of the measuring method and checking whether the measuring result meets the requirement of consistency;
the embodiment has the following effects:
the implementation method utilizes SAD stereo matching and high-precision equipment to cooperatively carry out binocular vision detection. The difficulty of the implementation method mainly comprises the cooperation splicing of high-precision equipment and the measurement adaptability of various flexible rubber workpieces, and the implementation method has the following advantages:
1. the implementation method does not depend on accurate image coordinates: the detection method takes the self characteristics of the workpiece as a reference, extracts the edge, matches the characteristics of SAD stereo, and carries out splicing and verification based on the cooperation of image characteristics and high-precision equipment. The stability of the algorithm is guaranteed, even if a certain error exists in the installation of a mechanical structure of the detection system, the acquired images can deviate to a certain degree, and the spliced images can be obtained stably and the three-dimensional thickness of the workpiece can be calculated.
2. The accuracy rate is high: due to the fact that the non-standard flexible rubber original piece is easy to deform, an extreme value is obtained by adopting a traversal algorithm when the 3D wall thickness of the feature point is calculated, and an actual wall thickness value is obtained. Meanwhile, the invention combines SAD stereo matching and image feature registration splicing, and ensures the accuracy of detection by calculating the 3D wall thickness of the feature position, as shown in Table 1.
3. The algorithm has strong robustness: by detecting the characteristic positions of the flexible rubber workpieces, automatically positioning the splicing positions and combining PLC (programmable logic controller) to control the high-precision sliding table to check, the automatic image splicing of the workpieces of different types and the calculation of the 3D wall thickness of all the characteristic positions are realized. The influence caused by errors in installation and operation of the mechanical mechanism is completely eliminated. The flexibility of the method ensures the convenience of updating and maintaining the visual detection system.
The second embodiment is as follows: the first difference between the present embodiment and the specific embodiment is: and step two, automatically calculating the advancing track of the high-precision sliding table according to the length of the workpiece, and shooting the measured workpiece in a sectional manner by using a binocular camera to obtain an image. Specifically, the high-pixel binocular camera is small in repeated vision, and in order to cope with flexible rubber workpieces with complicated length and characteristics, the automatic high-precision sliding table is controlled to move in a track and trigger the outside of the binocular camera, the flexible rubber workpieces are shot in a block division mode, and images required by splicing are obtained. Meanwhile, the track length of the high-precision sliding table is recorded, and a verifiable number is provided for the subsequent splicing based on the image characteristics. Other steps and parameters are the same as those in the first embodiment.
The third concrete implementation mode: the present embodiment differs from the first and second embodiments in that: and fifthly, finding out the characteristic corresponding points of the left camera and the right camera by using an SAD algorithm and an SAD algorithm. Summing the absolute values of the corresponding pixel differences for the corresponding pixel blocks of the aligned left and right view images:
wherein d ∈ [ d ] min ,d max ]In the formula (d) min And d max The length of the parallax search area is defined as 1= d, which can be determined from the maximum and minimum values of the two images where the same point in space appears in the two images max -d min +1. When the similarity is calculated for a point in the entire disparity search region, the disparity at the minimum or maximum metric value is used as the matching point for the point. Other steps and parameters are the same as in one of the first and second embodiments.
The fourth concrete implementation mode: the present embodiment differs from the first to third embodiments in that: splicing is carried out based on image characteristics in the sixth step, and verification is carried out by controlling the motion trail analysis of the high-precision sliding table, wherein the specific process is as follows:
(1) Carrying out extreme value extraction on XLD extracted by a Canny algorithm, and extracting the position information of the feature points in each group of contours;
(2) And matching the coordinate information of the same characteristic point in the two images according to the extracted coordinate information of the characteristic point and the relation of the motion distance of the high-precision sliding table accurately controlled by the PLC. By the extracted feature point information, a spatial conversion algorithm is used, redundant parts are cut off, the same feature points in the two images are combined, and the whole image splicing work is completed;
(3) And accurately controlling the moving distance of the high-precision sliding table according to the PLC, converting the distance into a pixel value in a pixel coordinate system, and comparing the pixel value with the intercepted redundant part, wherein the numerical value is the same, namely the splicing is successful. Other steps and parameters are the same as those in one of the first to third embodiments.
The fifth concrete implementation mode is as follows: the first to fourth differences of this embodiment from the first to fourth embodiments are: and seventhly, solving the 3D wall thickness of the characteristic position by using the edge profile. Because flexible rubber work piece, easy emergence is out of shape, adopts the method of traversing when calculating the wall thickness of feature point 3D, and the specific process is as follows:
(1) And (4) combining the coordinate information of the feature points at the same position in the binocular camera with the internal reference relative pose of the binocular camera to obtain the actual space X, Y and Z coordinates of any point on the graph.
(2) The characteristic point of the right edge of the flexible rubber workpiece is taken as a reference, and the coordinate of the characteristic point is (x) 1 ,y 1 ,z 1 ) Find and x at the left edge 1 And extracting the xyz coordinates of the upper point and the lower point of each point of the same point, wherein the total coordinate of the points is 2N +1.
(3) By usingTraversing the 2N +1 point and sequentially bringing the coordinates into a formula (x) n ,y n ,z n ) To yield Thick min I.e. the 3D wall thickness of the feature point. Other steps and parameters are the same as in one of the first to fourth embodiments.
The sixth specific implementation mode: the first to fifth differences of this embodiment from the first to fifth embodiments are: step eight, detecting the robustness of the measuring method and detecting whether the measuring result meets the requirement of consistency, and the method comprises the following specific steps:
(1) The flexible rubber workpieces of different types are operated according to the steps, and whether the flexible rubber workpieces can be completely spliced or not is checked;
(2) And continuously detecting the same workpiece for 8-10 times, and checking whether the group numerical value of the same characteristic point is extremely poor and the consistency reaches +/-0.03, so that the method is effective. Other steps and parameters are the same as those in one of the first to fifth embodiments. The following examples were employed to demonstrate the beneficial effects of the present invention:
the first embodiment is as follows:
the method for extracting the 3D wall thickness of the flexible rubber workpiece based on SAD stereo matching and high-precision equipment cooperation is specifically prepared according to the following steps:
step one, adopting a dot calibration plate to calibrate and correct distortion of a plurality of images, and performing epipolar line correction to obtain internal and external parameters and relative poses of the binocular camera:
(1) Shooting a calibration plate through a binocular camera to obtain left and right eye images, completing calibration of the binocular camera by using a calibration program, and obtaining the relation between internal parameters of the camera and the relative positions of the two cameras;
(2) And according to the calibration relation of the ideal projection point and the distortion projection point of the standard template picture, the distortion coefficient of the camera distortion model can be obtained. Suppose (x, y) is the actual position of the distortion point on the imaging plane, based on
k 1 、k 2 、k 3 Respectively representing the first, second and third order of radial distortion, P 1 And P 2 Is the projected point of a target point in space on two camera coordinate systems. And correcting the camera coordinate of the image through the distortion coefficient, converting the camera coordinate system into an image pixel coordinate system through the internal reference matrix after correction, and assigning a new image coordinate according to the pixel value of the source image coordinate.
(3) Converting an image pixel coordinate system subjected to distortion correction into a camera coordinate system (more zooming and Z-axis than an image physical coordinate system) through an internal reference matrix, performing parallel polar line correction through rotating matrixes R1 and R2, correcting the camera coordinate of an image through a distortion coefficient, converting the camera coordinate system into the image pixel coordinate system through the internal reference matrix after correction, and assigning a new image coordinate according to the pixel value of a source image coordinate;
and step two, automatically calculating the traveling track of the high-precision sliding table according to the length of the workpiece, and shooting the measured workpiece in a segmented manner by using a binocular camera to obtain images, as shown in fig. 9, 10, 11 and 12. Specifically, a high-pixel binocular camera has a small repetitive field of view. In order to deal with the flexible rubber workpiece with complex length and characteristics, the PLC is used for controlling the high-precision sliding table and the binocular camera, and the flexible rubber workpiece block is automatically shot in a region dividing mode according to the length and the characteristics of the workpiece to obtain images required by splicing. Meanwhile, recording the track of the high-precision sliding table which passes through each shooting under the control of the PLC, and providing verifiable data for the subsequent splicing based on image feature registration;
step three, acquiring the ROI by utilizing preprocessing modes such as binaryzation, smooth filtering, image enhancement, opening and closing operation and the like, as shown in figure 13;
step four, performing edge extraction on the obtained ROI by using a Canny algorithm, as shown in FIG. 14:
(1) Suppressing the noise following normal distribution by Gaussian smoothing filtering, and using I σ =I*G σ Wherein denotes a convolution operation; g σ Is a two-dimensional gaussian kernel with standard deviation σ, defined as:
(2) Convolution of the input image with the table horizontal and vertical operators computes dx, dy:
then d is x =f(x,y)*Sobel x (x,y)d y =f(x,y)*Sobel y (x, y). The magnitude of the image gradient can further be found: m (x, y) = | d x (x,y)|+|d y (x, y) |, angle: theta.theta. M =arctan(d y /d x )。
(3) And performing NMS treatment according to the gradient direction acquired in the last step. And comparing the neighborhood pixels of the image pixels along the gradient direction or the reverse direction, if the neighborhood pixels are the maximum value, retaining, and if not, inhibiting, namely setting the pixels to be 0. Namely:
N[i,j]=NMS(M[i,j],ζ[i,j])
(4) And a high threshold and a low threshold are set, and the two thresholds can filter noise in the image and improve the image quality. The processing principle of the weak edge is that strong edge points exist in the weak edge points of the real edge;
and fifthly, finding out the characteristic corresponding points of the left camera and the right camera by using an SAD algorithm and an SAD algorithm. Summing the absolute values of the corresponding pixel differences for the corresponding pixel blocks of the aligned left and right view images:
wherein d ∈ [ d ] min ,d max ]In the formula (d) min And d max The length of the specified parallax search region is 1= d, which can be obtained from the maximum and minimum values of the two images where the same point in space appears in the two images max -d min +1. After calculating the similarity in the whole parallax searching area for one point, using the parallax with the minimum or maximum metric value as the matching point of the point;
step six, splicing is carried out based on image characteristics, and verification is carried out by combining the motion trail analysis of the high-precision sliding table, as shown in fig. 15:
(1) Carrying out extreme value extraction on contour lines extracted by a Canny algorithm, and extracting the position information of feature points in each group of contour lines;
(2) And matching the coordinate information of the same characteristic point in the two images according to the extracted coordinate information of the characteristic point and the relation of the motion distance of the high-precision sliding table accurately controlled by the PLC. Extracting feature point information, using a spatial conversion algorithm to cut out redundant parts, and combining the same feature points in the two images to complete the whole image splicing work;
(3) Accurately controlling the moving distance of the high-precision sliding table according to the PLC, converting the distance into a pixel value in a pixel coordinate system, comparing the pixel value with the intercepted redundant part, and judging that the splicing is successful if the numerical value is the same;
and seventhly, solving the 3D wall thickness of the characteristic position by using the edge profile. Because flexible rubber work piece, easy emergence is out of shape, adopts the method of traversing when calculating the wall thickness of feature point 3D, and the specific process is as follows:
(1) And (4) combining the coordinate information of the feature points at the same position in the binocular camera with the internal reference relative pose of the binocular camera to obtain the actual space X, Y and Z coordinates of any point on the graph.
(2) The characteristic point of the right edge of the flexible rubber workpiece is taken as a reference, and the coordinate of the characteristic point is (x) 1 ,y 1 ,z 1 ) Find and x at the left edge 1 The xyz coordinates of the top and bottom 50 points are extracted for the same point, and the total 101 point coordinates are extracted.
(3) By usingTraversing the 101 points, and sequentially substituting the coordinates into the formula (x) n ,y n ,z n ) To yield Thick min I.e. the 3D wall thickness of the feature point.
Step eight, checking the robustness of the measuring method and checking whether the measuring result meets the consistency requirement:
(1) The flexible rubber workpieces of different types are operated according to the steps, and whether the flexible rubber workpieces can be completely spliced or not is checked;
(2) Continuously detecting the same workpiece for 8-10 times, and checking whether the group numerical value range of the same characteristic point is plus or minus 0.03 (range is 0.06) or not, wherein the method is effective and is shown in table 1; the characteristic points are selected from the following table according to the actual requirements of users:
TABLE 1 product detection data example table containing 5 wave crests and 5 wave troughs
SAD stereo matching is global matching based on pixel gray value and calibration parameter correction, and effectively matches all characteristic corresponding points of an image; the method is a method with high efficiency and high accuracy used according to actual requirements of customers and the self characteristics of the rubber dustproof workpieces combined with measurement, and the flexible rubber workpieces with different lengths and different numbers of characteristic points are successfully spliced by using the two points, so that the method has high robustness; the method is characterized in that the PLC controls high-precision equipment to cooperatively splice and measure, and splice and check are carried out, so that the method is one of the innovation points of the method; the wall thickness is obtained by utilizing the spatial three-dimension, the measurement consistency is judged, the accuracy of the measurement method is verified, and the problem of extracting the 3D wall thickness of the flexible rubber workpiece is successfully solved.
Claims (6)
1. A3D wall thickness extraction method for a flexible rubber workpiece based on SAD stereo matching and high-precision equipment cooperation is characterized by comprising the following steps: A3D wall thickness extraction method of a flexible rubber workpiece based on SAD stereo matching and high-precision equipment cooperation is specifically carried out according to the following steps:
firstly, a dot calibration plate is adopted to calibrate and correct distortion of a plurality of images, epipolar line correction is carried out, and internal and external parameters and relative poses of a binocular camera are obtained;
automatically calculating the traveling track of the high-precision sliding table according to the length of the workpiece, and shooting the measured workpiece in sections by using a binocular camera to obtain an image;
step three, preprocessing the image to obtain an ROI;
step four, utilizing an edge extraction algorithm to extract edges of the obtained ROI;
finding out corresponding points of the characteristics of the left camera and the right camera by utilizing an SAD algorithm;
splicing based on image characteristics, and verifying by combining with the motion trail analysis of the high-precision sliding table;
step seven, solving the 3D wall thickness of the characteristic position by utilizing the edge contour;
and step eight, checking the robustness of the measuring method and checking whether the measuring result meets the requirement of consistency.
2. The method for extracting the 3D wall thickness of the flexible rubber workpiece based on the cooperation of SAD stereo matching and high-precision equipment as claimed in claim 1, is characterized in that: and step two, automatically calculating the advancing track of the high-precision sliding table according to the length of the workpiece, and shooting the measured workpiece in a sectional manner by using a binocular camera to obtain an image. Specifically, the high-pixel binocular camera is small in repeated vision, and in order to cope with flexible rubber workpieces with complicated length and characteristics, the automatic high-precision sliding table is controlled to move in a track and trigger the outside of the binocular camera, the flexible rubber workpieces are shot in a block division mode, and images required by splicing are obtained. Meanwhile, the track length of the high-precision sliding table is recorded, and verifiable data is provided for the subsequent splicing based on image characteristics.
3. The method for extracting the 3D wall thickness of the flexible rubber workpiece based on the cooperation of SAD stereo matching and high-precision equipment as claimed in claim 1, is characterized in that: and fifthly, finding out the characteristic corresponding points of the left camera and the right camera by using an SAD algorithm and an SAD algorithm. Summing the absolute values of the corresponding pixel differences for the corresponding pixel blocks of the aligned left and right view images:
wherein d ∈ [ d ] min ,d max ]D in the formula min And d max The length of the parallax search area is defined as 1= d, which can be determined from the maximum and minimum values of the two images where the same point in space appears in the two images max -d min +1. When the similarity is calculated for a point in the entire disparity search region, the disparity at the minimum or maximum metric value is used as the matching point for the point.
4. The method for extracting the 3D wall thickness of the flexible rubber workpiece based on the cooperation of SAD stereo matching and high-precision equipment as claimed in claim 1, is characterized in that: splicing is carried out based on image characteristics in the sixth step, and verification is carried out by controlling the motion trail analysis of the high-precision sliding table, wherein the specific process is as follows:
(1) Performing extreme value extraction on XLD extracted by an edge detection algorithm, and extracting the position information of the feature points in each group of profiles;
(2) And matching the coordinate information of the same characteristic point in the two images according to the extracted coordinate information of the characteristic point and the relationship of the motion distance of the high-precision sliding table. By the extracted feature point information, a spatial conversion algorithm is used, redundant parts are cut off, the same feature points in the two images are combined, and the whole image splicing work is completed;
(3) And (4) according to the movement distance of the precisely controlled high-precision sliding table, converting the distance into a pixel value in a pixel coordinate system, and comparing the pixel value with the intercepted redundant part, wherein the numerical value is the same, namely the splicing is successful.
5. The method for extracting the 3D wall thickness of the flexible rubber workpiece based on the cooperation of SAD stereo matching and high-precision equipment as claimed in claim 1, is characterized in that: and seventhly, solving the 3D wall thickness of the characteristic position by using the edge profile. Because flexible rubber work piece, easy emergence is out of shape, adopts the method of traversing when calculating the wall thickness of feature point 3D, and the concrete process is as follows:
(1) And (4) combining the coordinate information of the feature points at the same position in the binocular camera with the internal reference and relative pose of the binocular camera to obtain the actual space X, Y and Z coordinates of any point on the graph.
(2) The characteristic point of the right edge of the flexible rubber workpiece is taken as a reference, and the coordinate of the characteristic point is (x) 1 ,y 1 ,z 1 ) Find and x at the left edge 1 And extracting xyz coordinates of upper and lower N points at the same point, wherein the total coordinate of 2N +1 points is obtained.
6. The method for extracting the 3D wall thickness of the flexible rubber workpiece based on the cooperation of SAD stereo matching and high-precision equipment as claimed in claim 1, wherein the method comprises the following steps: step eight, detecting the robustness of the measuring method and detecting whether the measuring result meets the requirement of consistency, wherein the method comprises the following specific steps:
(1) The flexible rubber workpieces of different types are operated according to the steps, and whether the flexible rubber workpieces can be completely spliced or not is checked;
(2) And continuously detecting the same workpiece for 8-10 times, and checking whether the group numerical value of the same characteristic point has extreme difference and the consistency reaches +/-0.03, so that the method is effective.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110650600.2A CN115471536A (en) | 2021-06-11 | 2021-06-11 | Flexible rubber workpiece 3D wall thickness extraction method based on SAD (sum of absolute differences) stereo matching and high-precision equipment cooperation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110650600.2A CN115471536A (en) | 2021-06-11 | 2021-06-11 | Flexible rubber workpiece 3D wall thickness extraction method based on SAD (sum of absolute differences) stereo matching and high-precision equipment cooperation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115471536A true CN115471536A (en) | 2022-12-13 |
Family
ID=84363381
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110650600.2A Pending CN115471536A (en) | 2021-06-11 | 2021-06-11 | Flexible rubber workpiece 3D wall thickness extraction method based on SAD (sum of absolute differences) stereo matching and high-precision equipment cooperation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115471536A (en) |
-
2021
- 2021-06-11 CN CN202110650600.2A patent/CN115471536A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109308693B (en) | Single-binocular vision system for target detection and pose measurement constructed by one PTZ camera | |
CN109598762B (en) | High-precision binocular camera calibration method | |
CN111721259B (en) | Underwater robot recovery positioning method based on binocular vision | |
US11488322B2 (en) | System and method for training a model in a plurality of non-perspective cameras and determining 3D pose of an object at runtime with the same | |
CN109470149B (en) | Method and device for measuring position and posture of pipeline | |
CN103292695A (en) | Monocular stereoscopic vision measuring method | |
CN112964186B (en) | Device and method for measuring clearance in automatic shaft hole assembly process | |
CN111062940B (en) | Screw positioning and identifying method based on machine vision | |
CN102157004A (en) | Automatic image mosaicking method for high-accuracy image measuring apparatus of super-view field part | |
CN106996748A (en) | A kind of wheel footpath measuring method based on binocular vision | |
CN111531407B (en) | Workpiece attitude rapid measurement method based on image processing | |
Wang et al. | Error analysis and improved calibration algorithm for LED chip localization system based on visual feedback | |
CN113119129A (en) | Monocular distance measurement positioning method based on standard ball | |
CN113516716A (en) | Monocular vision pose measuring and adjusting method and system | |
CN113538399A (en) | Method for obtaining accurate contour of workpiece, machine tool and storage medium | |
CN113103226A (en) | Visual guide robot system for ceramic biscuit processing and manufacturing | |
CN112109374A (en) | Method for positioning and controlling assembling and disassembling of bending die based on computer vision system | |
CN115471536A (en) | Flexible rubber workpiece 3D wall thickness extraction method based on SAD (sum of absolute differences) stereo matching and high-precision equipment cooperation | |
CN116596987A (en) | Workpiece three-dimensional size high-precision measurement method based on binocular vision | |
CN115222893A (en) | Three-dimensional reconstruction splicing method for large-size components based on structured light measurement | |
JP2006145419A (en) | Image processing method | |
CN112950527A (en) | Stereo matching morphology measurement method based on limited geometric association constraint | |
CN205120108U (en) | Precision compensating system based on camera device | |
CN113916149B (en) | Cylindrical straight gear appearance vision measuring device based on multi-eye CCD industrial camera and image processing method | |
CN109544509A (en) | One kind is based on the matched workpiece localization method of secondary template, equipment and storage equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |