WO2023109960A1 - 三维扫描的处理方法、装置和三维扫描设备 - Google Patents

三维扫描的处理方法、装置和三维扫描设备 Download PDF

Info

Publication number
WO2023109960A1
WO2023109960A1 PCT/CN2022/139716 CN2022139716W WO2023109960A1 WO 2023109960 A1 WO2023109960 A1 WO 2023109960A1 CN 2022139716 W CN2022139716 W CN 2022139716W WO 2023109960 A1 WO2023109960 A1 WO 2023109960A1
Authority
WO
WIPO (PCT)
Prior art keywords
matching
dimensional
point
frame
light plane
Prior art date
Application number
PCT/CN2022/139716
Other languages
English (en)
French (fr)
Inventor
刘增艺
Original Assignee
先临三维科技股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 先临三维科技股份有限公司 filed Critical 先临三维科技股份有限公司
Publication of WO2023109960A1 publication Critical patent/WO2023109960A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • 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
    • G06T2200/00—Indexing scheme for image data processing or generation, in general
    • G06T2200/08—Indexing scheme for image data processing or generation, in general involving all processing steps from image acquisition to 3D model generation

Definitions

  • the present disclosure relates to the technical field of three-dimensional scanning, and in particular, to a processing method and device for three-dimensional scanning, and three-dimensional scanning equipment.
  • 3D scanning technology is more and more widely used in industry.
  • some industrial objects to be measured are often bulky industrial parts, which have high requirements for scanning efficiency.
  • the most common scanning solution at present is binocular laser scanning, and the number of laser lines of binocular laser scanning is less than 20 lines. This scanning efficiency cannot meet some scenarios requiring high scanning efficiency.
  • the number of scanning lines needs to be increased, but the increase in the number of scanning lines in the binocular stereo vision system will lead to a decrease in the accuracy of matching.
  • the method to improve scanning efficiency is to increase the number of scanning lines, but increasing the number of scanning lines in the binocular scanning system will lead to matching
  • the problem of sharp drop in accuracy has not yet been proposed an effective solution.
  • the main purpose of the present disclosure is to provide a three-dimensional scanning processing method, device, and three-dimensional scanning equipment to solve the scene in the related art that the scanning efficiency of binocular laser scanning cannot meet some high scanning efficiency requirements.
  • the method of improving scanning efficiency is through Increasing the number of scanning lines, but increasing the number of scanning lines in a binocular scanning system will lead to a sharp drop in matching accuracy.
  • a processing method for three-dimensional scanning includes: projecting multiple lines to the surface of the measured object through a pattern projector; collecting two-dimensional images of the surface of the measured object through three cameras, correspondingly obtaining three frames of two-dimensional images; determining two frames of the three frames of two-dimensional images Matching point pairs between the two correspondingly obtain three sets of matching point pairs; verify the matching consistency between the matching point pairs; perform three-dimensional reconstruction on the matching point pairs with matching consistency to obtain the surface of the measured object 3D point.
  • determining matching point pairs between two pairs of the three frames of two-dimensional images to obtain three sets of matching point pairs includes: acquiring a plurality of lines in the three frames of two-dimensional images, wherein the lines are formed by Composed of multiple pixel points, a pixel point on a line in a frame of two-dimensional image is used as a selected point, and multiple candidate matching points matching the selected point in another frame of two-dimensional image are determined; based on the principle of triangulation, the performing three-dimensional reconstruction on the selected point and the plurality of candidate matching points to obtain a plurality of first candidate three-dimensional points; determining the first candidate three-dimensional points satisfying preset conditions as second candidate three-dimensional points, wherein the first There are multiple two candidate three-dimensional points; the candidate matching point corresponding to the second candidate three-dimensional point and the selected point form the matching point pair.
  • the method further includes: acquiring a plurality of light planes corresponding to the second candidate three-dimensional points, wherein the light planes are the A light plane corresponding to the selected point; determining a target light plane corresponding to the selected point from the plurality of light planes.
  • determining the target light plane corresponding to the selected point from the multiple light planes includes: obtaining multiple light planes corresponding to multiple pixel points on the same line as the selected point; calculating each The number of occurrences of each light plane, and the light plane with the most occurrences is taken as the target light plane.
  • verifying the matching consistency between the three sets of matching point pairs includes: using the matching point corresponding to the target light plane as a target matching point; combining the selected point and the target matching point to form a target matching point pairs to obtain three groups of target matching point pairs; verify the consistency among the three groups of target matching point pairs.
  • the three frames of two-dimensional images are respectively the first frame of two-dimensional images, the second frame of two-dimensional images and the third frame of two-dimensional images, and verifying the matching consistency between the three groups of target matching point pairs includes: Obtaining a selected point in the first frame of two-dimensional image, and a target light plane 1 determined by matching the selected point through the first frame of two-dimensional image and the second frame of two-dimensional image; obtaining the A selected point in the first frame of two-dimensional image, and the target light plane 2 determined by matching the selected point through the first frame of two-dimensional image and the third frame of two-dimensional image; acquiring the first frame The target matching point of the selected point in the two-dimensional image in the second frame of two-dimensional image; acquiring the target matching point in the second frame of two-dimensional image passing through the second frame of two-dimensional image and the The third frame of the two-dimensional image matches the determined target light plane three; judging whether the target light plane one, the target light plane two and the target light plane three are the same light plane; when the target light
  • determining a plurality of candidate matching points matching the selected point in another frame of two-dimensional image includes: obtaining the selected point Corresponding to the epipolar line equation of the other frame of two-dimensional image; taking intersections of the epipolar line equation and multiple lines in the other frame of two-dimensional image as the plurality of candidate matching points.
  • a three-dimensional scanning device includes three cameras, and the three cameras are combined in pairs to obtain three binocular systems; wherein, the three binocular systems are used to collect two-dimensional images of the surface of the measured object, and correspondingly obtain three frames A two-dimensional image, wherein, determine the matching point pairs between two two-dimensional images of the three frames, correspondingly obtain three sets of matching point pairs; verify the matching consistency between the matching point pairs; Three-dimensional reconstruction is performed on the matched point pairs to obtain three-dimensional points on the surface of the measured object.
  • a processing device for three-dimensional scanning includes: a projection unit configured to project multiple lines onto the surface of the object to be measured through a pattern projector; a collection unit configured to collect two-dimensional images of the surface of the object to be measured through three cameras, corresponding to three frames and two three-dimensional image; the first determination unit is configured to determine the matching point pairs between two pairs of the three frames of two-dimensional images, and correspondingly obtains three sets of matching point pairs; the verification unit is configured to verify the matching point pairs between matching consistency; the reconstruction unit is configured to perform three-dimensional reconstruction on the matching point pairs with matching consistency to obtain the three-dimensional points on the surface of the measured object.
  • the determination unit includes: a first acquisition subunit configured to acquire multiple lines in the three frames of two-dimensional images, wherein the lines are composed of multiple pixel points, the first determination subunit, It is configured to use a pixel point on a line in a frame of two-dimensional image as a selected point, and determine a plurality of candidate matching points matching the selected point in another frame of two-dimensional image;
  • the reconstruction subunit is configured to be based on The principle of triangulation is to carry out three-dimensional reconstruction of the selected point and the plurality of candidate matching points to obtain a plurality of first candidate three-dimensional points;
  • the second determination subunit is configured to use the first candidate three-dimensional points that meet the preset conditions Determined as the second candidate three-dimensional point, wherein there are multiple second candidate three-dimensional points;
  • the first component subunit is configured to combine the candidate matching point corresponding to the second candidate three-dimensional point with the selected points make up the matching point pairs.
  • the device further includes: an acquisition unit configured to acquire a plurality of light planes corresponding to the second candidate three-dimensional point before verifying the matching consistency among the three sets of matching point pairs, wherein the The light plane is the light plane corresponding to the selected point; the second determining unit is configured to determine a target light plane corresponding to the selected point from the plurality of light planes.
  • the second determination unit includes: a second acquisition subunit configured to acquire multiple light planes corresponding to multiple pixel points on the same line of the selected point; a calculation subunit configured to Count the number of occurrences of each light plane, and use the light plane with the most occurrences as the target light plane.
  • the verification unit includes: using the matching point corresponding to the target light plane as a target matching point; a second composition subunit configured to form a target matching point pair between the selected point and the target matching point , to obtain three groups of target matching point pairs; the verification subunit is configured to verify the consistency among the three groups of target matching point pairs.
  • the three frames of two-dimensional images are respectively the first frame of two-dimensional images, the second frame of two-dimensional images and the third frame of two-dimensional images
  • the verification subunit includes: a first acquisition module configured to acquire the The selected point in the first frame of two-dimensional image, and the target light plane one determined by the selected point through the matching of the first frame of two-dimensional image and the second frame of two-dimensional image; the second acquisition module, configured to acquire a selected point in the first frame of two-dimensional image, and a target light plane 2 determined by matching the selected point through the first frame of two-dimensional image and the third frame of two-dimensional image;
  • the third acquisition module is configured to acquire the target matching point of the selected point in the first frame of two-dimensional image in the second frame of two-dimensional image;
  • the fourth acquisition module is configured to acquire the second frame of two-dimensional image
  • the target matching point in the frame of two-dimensional image matches the target light plane three determined by the second frame of two-dimensional image and the third frame of two-dimensional image;
  • the judging module
  • the first determination subunit includes: a fifth acquisition unit configured to acquire an epipolar line equation corresponding to the selected point to the other frame of two-dimensional image; combine the epipolar line equation with the Intersection points of multiple lines in another two-dimensional image are used as the multiple candidate matching points.
  • a computer-readable storage medium includes a stored program, wherein the program executes the processing method for three-dimensional scanning described in any one of the above .
  • a processor is provided, and the processor is used for running a program, wherein, when the program is running, the processing method for three-dimensional scanning described in any one of the above is executed.
  • the following steps are adopted: projecting multiple lines to the surface of the measured object through a pattern projector; collecting two-dimensional images of the surface of the measured object through three cameras, and correspondingly obtaining three frames of two-dimensional images; determining the three frames Matching point pairs between two two-dimensional images correspond to three sets of matching point pairs; verify the matching consistency between the matching point pairs; perform three-dimensional reconstruction on the matching point pairs with matching consistency to obtain the matched point pairs Measure the three-dimensional points on the surface of the object, and solve the scene where the scanning efficiency of binocular laser scanning cannot meet the requirements of some high scanning efficiency in the related technology.
  • the method to improve the scanning efficiency is to increase the number of scanning lines, but increase the scan The number of lines will lead to a sharp drop in matching accuracy.
  • the two-dimensional images of the surface of the measured object are collected by three cameras to obtain three frames of two-dimensional images, and the three frames of two-dimensional images are matched in pairs to obtain three sets of matching point pairs, and three-dimensional matching point pairs with matching consistency are obtained. Reconstruct to obtain the three-dimensional points on the surface of the object to be measured, thereby achieving the effect of improving the accuracy of matching.
  • FIG. 1 is a flow chart of a processing method for three-dimensional scanning provided according to an embodiment of the present disclosure
  • Fig. 2 is a schematic diagram of an optional three-frame two-dimensional image provided according to an embodiment of the present disclosure
  • Fig. 3 is a schematic diagram of an optional three-dimensional scanning device provided according to an embodiment of the present disclosure
  • Fig. 4 is a schematic diagram of a three-dimensional scanning processing device provided according to an embodiment of the present disclosure.
  • FIG. 1 is a flow chart of a processing method for three-dimensional scanning provided according to an embodiment of the present disclosure. As shown in FIG. 1 , the method includes the following steps:
  • Step S101 projecting multiple lines onto the surface of the object to be measured by a pattern projector.
  • a pattern projector of a 3D scanning device projects a plurality of scan lines (for example, laser scan lines may be used) onto the surface of an object to be constructed into a 3D model.
  • step S102 three cameras are used to collect two-dimensional images of the surface of the measured object, and correspondingly three frames of two-dimensional images are obtained.
  • the three cameras of the three-dimensional scanning device collect two-dimensional images of the surface of the object to be constructed in a three-dimensional model, and correspondingly obtain three frames of two-dimensional images (for example, Fig. 2 three 2D images shown).
  • the three frames of two-dimensional images are respectively a first frame of two-dimensional images, a second frame of two-dimensional images and a third frame of two-dimensional images.
  • the three-dimensional scanning device is a handheld three-dimensional scanner, which can move and scan relative to the measured object.
  • the three cameras collect synchronously, that is, the three cameras collect synchronously at the first time and at the second time until the scanning is completed, so as to ensure the correspondence of the images obtained by the three cameras.
  • the three cameras can also be asynchronous, but it is necessary to ensure that the acquisition time interval is extremely small, and the position of the 3D scanning device relative to the measured object is almost constant. If the three-dimensional scanning device is used fixedly every time the measured object is collected, the three-dimensional scanning device is fixed at the first position to collect one part of the measured object, and then fixed to the second position to collect another part of the measured object until it is detected If the object to be measured is scanned, there is no limit to whether the acquisitions of the three cameras are synchronized.
  • Step S103 determining matching point pairs between two pairs of the three frames of two-dimensional images, and correspondingly obtaining three sets of matching point pairs.
  • Two-by-two matching is performed on the three frames of two-dimensional images to obtain three corresponding sets of matching point pairs.
  • a point A(1,2) on the first frame of the two-dimensional image, A(1,2) is matched in the second frame of the two-dimensional image and the matching point coordinates are B 1 (2,3);
  • the matching point pairs of the first two-dimensional image and the second two-dimensional image are A(1,2) and B 1 (2,3);
  • A(1,2) is obtained by matching the third two-dimensional image
  • the matching point coordinates are C(1,3); then the matching point pairs of the first two-dimensional image and the third two-dimensional image are A(1,2) and C(1,3);
  • the matching point obtained by matching B 1 (2, 3) on the third two-dimensional image in the third frame is C(1, 3); then the matching point pair between the second two-dimensional image and the third two-dimensional image is B 1 (2,3) and C(1,3).
  • Step S104 verifying the matching consistency between the matching point pairs.
  • Step S105 performing three-dimensional reconstruction on the matching point pairs with matching consistency, to obtain three-dimensional points on the surface of the measured object.
  • three-dimensional reconstruction is performed on the point to obtain the three-dimensional point of the object that needs to be constructed into a three-dimensional model, and the three-dimensional model of the object is constructed based on the three-dimensional point.
  • the two-dimensional images of the object are collected by three cameras to obtain three frames of two-dimensional images, pairwise matching is performed on the three frames of two-dimensional images to obtain matching point pairs, and three-dimensional reconstruction is performed on the matching point pairs with matching consistency , so as to obtain the 3D model of the object, which improves the accuracy of 3D reconstruction matching.
  • determining matching point pairs between two two-dimensional images of three frames, correspondingly obtaining three sets of matching point pairs includes: acquiring three frames of two-dimensional images Multiple lines, wherein the line is composed of multiple pixels, a pixel on the line in one frame of two-dimensional image is used as a selected point, and multiple candidate matches matching the selected point in another frame of two-dimensional image are determined point; based on the principle of triangulation, three-dimensional reconstruction is performed on the selected point and multiple candidate matching points to obtain multiple first candidate three-dimensional points; the first candidate three-dimensional point meeting the preset condition is determined as the second candidate three-dimensional point, wherein, There are multiple second candidate three-dimensional points; the candidate matching point corresponding to the second candidate three-dimensional point and the selected point form a matching point pair.
  • three frames of two-dimensional images contain multiple lines (i.e., multiple scanning lines emitted by the pattern projector), and the lines are composed of multiple pixel points.
  • the corresponding image of Cam L in Figure 2 is The first frame of two-dimensional image
  • Cam M corresponds to the second frame of two-dimensional image
  • Cam H corresponds to the third frame of two-dimensional image.
  • the candidate matching points in the second frame of the 2D image include B1, B2, B3, B4, B5 and B6.
  • the screening method is to calculate the distance Distance(k) from the first candidate 3D point to the light plane Plane(k) corresponding to all scan lines.
  • the distance value of a first candidate three-dimensional point is within the set distance threshold, that is, Distance(k) ⁇ dist TH, it means that the first candidate three-dimensional point is in the light plane k (th projection line), That is, the first candidate three-dimensional point satisfies the preset requirement.
  • the second candidate three-dimensional point has O 1 , O 2 and O 3
  • the candidate matching points B 1 , B 2 and B 3 corresponding to O 1 , O 2 and O 3 form a matching point pair with point A.
  • the matching point of each pixel point on the first frame of two-dimensional image on the second frame of two-dimensional image is obtained through the above method. Also use the same method to obtain the matching point pairs of the first frame of 2D image and the third frame of 2D image, and the matching point pairs of the second frame of 2D image and the third frame of 2D image.
  • the method before verifying the matching consistency among the three sets of matching point pairs, the method further includes: acquiring multiple light planes corresponding to the second candidate three-dimensional point , where the light plane is the light plane corresponding to the selected point; the target light plane corresponding to the selected point is determined from multiple light planes.
  • the second candidate 3D point is determined by calculating the distance Distance(k) between the point and the light plane Plane(k) corresponding to all scan lines.
  • the distance value of a second candidate three-dimensional point is within the set distance threshold, that is, Distance(k) ⁇ dist TH, it means that the point is in the k-th light plane (the k-th projection line). That is to say, one second candidate three-dimensional point corresponds to one light plane.
  • the light planes corresponding to the second candidate three-dimensional points O 1 , O 2 and O 3 are K 1 , K 2 and K 3 respectively.
  • the best light plane (that is, the above-mentioned target light plane) is selected from these light planes.
  • the reconstructed three-dimensional points on the surface of the object can be obtained more accurately.
  • determining the target light plane corresponding to the selected point from multiple light planes includes: acquiring multiple pixel points on the same line as the selected point Corresponding multiple light planes; count the number of occurrences of each light plane, and use the light plane with the most occurrences as the target light plane.
  • the method for selecting the best light plane from the light planes K 1 , K 2 and K 3 is: firstly obtain multiple light planes corresponding to multiple pixel points of the selected point on the same line. As shown in FIG. 2 , there are multiple pixel points A 1 , A 2 , A 3 and A 4 on the same line as point A.
  • the multiple light planes corresponding to A 1 are K 1 and K 2 ; the multiple light planes corresponding to A 2 are K 1 and K 4 ; the multiple light planes corresponding to A 3 are K 1 , K 2 and K 3 ; A 4
  • the corresponding multiple light planes are K 1 and K 2 ; the number of occurrences of each light plane is calculated, and the light plane with the most occurrences is taken as the best light plane, that is, the best light plane corresponding to point A is K 1 .
  • the lines in the three-frame two-dimensional images may be discontinuous lines, then by obtaining the selected point on the line and searching for multiple light planes corresponding to multiple pixels in the neighborhood to determine the corresponding to the selected point Optimal light plane.
  • the optimal light plane corresponding to the selected point is determined through multiple light planes corresponding to multiple pixel points of the selected point on the same line, which improves the accuracy of the light plane corresponding to the pixel point.
  • verifying the matching consistency among the three sets of matching point pairs includes: using the matching point corresponding to the target light plane as the target matching point; The target matching point pairs are formed with the target matching points, and three sets of target matching point pairs are obtained; the consistency among the three sets of target matching point pairs is verified.
  • the matching point corresponding to the best light plane is the best matching point.
  • the best matching point pair of the first two-dimensional image and the second two-dimensional image is A(1,2) and B1 (2,3), and the corresponding best light plane is K1 ;
  • the first frame The best matching point pairs between the 2D image and the third frame of 2D image are A(1, 2) and C(1, 3), and the corresponding best light plane is K 1 ;
  • the best matching point pair of the frame two-dimensional image is B 1 (2, 3) and C (1, 3), and the corresponding best light plane is K 1 ;
  • Only the 3D points reconstructed by the best matching point pairs may be the 3D points on the surface of the object that needs to build a 3D model, so it is necessary to verify the consistency between the best matching point pairs.
  • the three frames of two-dimensional images are respectively the first frame of the two-dimensional image, the second frame of the two-dimensional image and the third frame of the two-dimensional image, and the verified
  • the matching consistency between the three groups of target matching point pairs includes: obtaining a selected point in the first frame of two-dimensional image, and the selected point passing through the first frame of two-dimensional image and the second Frame 2D image matching determined target light plane 1; Acquiring a selected point in the first frame 2D image, and the selected point passing through the first frame 2D image and the third frame 2D
  • the target light plane 2 determined by image matching; obtaining the target matching point of the selected point in the two-dimensional image of the first frame in the two-dimensional image of the second frame; obtaining the two-dimensional image of the second frame
  • the target matching point matches the target light plane 3 determined by matching the second two-dimensional image with the third frame two-dimensional image; judging the target light plane one, the target light plane two and the target light plane three Whether
  • a point A(1,2) on the first two-dimensional image, A(1,2) is matched in the second two-dimensional image and the best matching point coordinate is B 1 (2,3); then the first The best matching point pair of the frame 2D image and the second frame of 2D image is A(1, 2) and B 1 (2, 3), and the corresponding best light plane is K 1 .
  • the coordinates of the best matching point obtained by A(1, 2) in the third frame of 2D image matching are C(1, 3); then the best matching point pair of the first frame of 2D image and the third frame of 2D image is A(1, 2) and C(1, 3), the corresponding optimal light plane is K 1 .
  • the best matching point obtained by matching B 1 (2, 3) on the second frame of two-dimensional image in the third frame of two-dimensional image is C (1, 3); then the second frame of two-dimensional image and the third frame of two-dimensional
  • the best matching point pair of the image is B 1 (2, 3) and C (1, 3), and the corresponding best light plane is K 1 .
  • the best light plane obtained by pairwise matching is the same, that is, K 1 , indicating that the three groups of best matching point pairs have matching consistency.
  • a pixel point on a line in a frame of two-dimensional image is used as a selected point, and multiple pixels matching the selected point in another frame of two-dimensional image are determined.
  • candidate matching points including: obtaining the epipolar line equation corresponding to another frame of two-dimensional image from the selected point; taking the intersection points of the epipolar line equation and multiple lines in another frame of two-dimensional image as multiple candidate matching points.
  • the polar line is calculated based on the relative positions of Cam L and Cam M.
  • Line equations such as the schematic epipolar equations labeled in Figure 2.
  • the intersection points of the epipolar line equation and multiple lines in another frame of two-dimensional image are used as multiple candidate matching points.
  • the three-dimensional scanning processing method uses a pattern projector to project multiple lines onto the surface of the measured object; collects two-dimensional images of the surface of the measured object through three cameras, and correspondingly obtains three frames of two-dimensional images; determine Matching point pairs between two pairs of the three frames of two-dimensional images correspond to three sets of matching point pairs; verify the matching consistency between the matching point pairs; perform three-dimensional reconstruction on the matching point pairs with matching consistency, Obtaining the three-dimensional points on the surface of the measured object solves the problem that the scanning efficiency of the binocular laser scanning in the related art cannot meet some high scanning efficiency requirements.
  • the method to improve the scanning efficiency is to increase the number of scanning lines, but the Increasing the number of scanning lines in the system will lead to a sharp drop in matching accuracy.
  • the two-dimensional images of the surface of the measured object are collected by three cameras to obtain three frames of two-dimensional images, and the three frames of two-dimensional images are matched in pairs to obtain three sets of matching point pairs, and three-dimensional matching point pairs with matching consistency are obtained. Reconstruct to obtain the three-dimensional points on the surface of the object to be measured, thereby achieving the effect of improving the accuracy of matching.
  • An embodiment of the present disclosure also provides a three-dimensional scanning device.
  • the three-dimensional scanning device includes three cameras, and the three cameras are combined in pairs to obtain three binocular systems; wherein, the three binocular systems are used to collect objects to be measured.
  • the two-dimensional image of the surface corresponds to three frames of two-dimensional images, wherein the matching point pairs between two two-dimensional images of the three frames are determined, and three sets of matching point pairs are correspondingly obtained; the matching consistency between the matching point pairs is verified; Perform 3D reconstruction on the matched point pairs with matching consistency to obtain 3D points on the surface of the measured object.
  • FIG. 3 it is a schematic diagram of an optional three-dimensional scanning device provided according to an embodiment of the present disclosure, wherein Cam L, Cam M and Cam R are three cameras, and two of the three cameras form three binocular systems.
  • Projector is a pattern projector.
  • the embodiment of the present disclosure also provides a three-dimensional scanning processing device. It should be noted that the three-dimensional scanning processing device of the present disclosure embodiment can be used to execute the processing method for three-dimensional scanning provided by the present disclosure embodiment.
  • the processing apparatus for three-dimensional scanning provided by the embodiments of the present disclosure will be introduced below.
  • FIG. 4 is a schematic diagram of a three-dimensional scanning processing device according to an embodiment of the disclosure. As shown in FIG. 4 , the device includes: a projection unit 401 , an acquisition unit 402 , a first determination unit 403 , a verification unit 404 and a reconstruction unit 405 .
  • the projection unit 401 is configured to project multiple lines to the surface of the measured object through the pattern projector.
  • the acquisition unit 402 is configured to acquire two-dimensional images of the surface of the measured object through three cameras, and correspondingly obtain three frames of two-dimensional images.
  • the first determining unit 403 is configured to determine matching point pairs between two pairs of three frames of two-dimensional images, and correspondingly obtain three groups of matching point pairs.
  • the verification unit 404 is configured to verify matching consistency between matching point pairs.
  • the reconstruction unit 405 is configured to perform three-dimensional reconstruction on the matched point pairs with matching consistency to obtain three-dimensional points on the surface of the measured object.
  • the three-dimensional scanning processing device uses the projection unit 401 to project multiple lines to the surface of the measured object through the pattern projector; the acquisition unit 402 collects two-dimensional images of the surface of the measured object through three cameras, and correspondingly obtains three frames Two-dimensional image; the first determination unit 403 determines the matching point pairs between two pairs of three frames of two-dimensional images, correspondingly obtaining three groups of matching point pairs; the verification unit 404 verifies the matching consistency between the matching point pairs; the reconstruction unit 405 pairs Three-dimensional reconstruction is performed on matching point pairs with matching consistency to obtain three-dimensional points on the surface of the measured object, which solves the scene where the scanning efficiency of binocular laser scanning in the related technology cannot meet some high scanning efficiency requirements.
  • the method to improve scanning efficiency is through Increase the number of scanning lines, but increasing the number of scanning lines in the binocular scanning system will lead to a sharp drop in matching accuracy.
  • Three cameras are used to collect two-dimensional images of the surface of the measured object to obtain three frames of two-dimensional images. Match the three frames of two-dimensional images two by two to obtain three sets of matching point pairs, perform three-dimensional reconstruction on the matching point pairs with matching consistency, and obtain the three-dimensional points on the surface of the object to be measured, thereby achieving the effect of improving the accuracy of matching .
  • the determining unit includes: a first acquiring subunit configured to determine matching point pairs between two two-dimensional images of three frames, correspondingly obtain three Before group matching point pairs, multiple lines in three frames of two-dimensional images are obtained, wherein the lines are composed of multiple pixels, and the first determination subunit is configured to use one pixel point on the lines in one frame of two-dimensional images as A selected point is used to determine a plurality of candidate matching points matching the selected point in another two-dimensional image; the reconstruction subunit is configured to perform three-dimensional reconstruction on the selected point and the plurality of candidate matching points based on the principle of triangulation, to obtain a plurality of first candidate three-dimensional points; the second determination subunit is configured to determine the first candidate three-dimensional points satisfying preset conditions as second candidate three-dimensional points, wherein there are multiple second candidate three-dimensional points; the first composition The subunit is configured to form a matching point pair between the candidate matching point corresponding to the second candidate three-dimensional point and the
  • the device further includes: an acquisition unit configured to acquire the second candidate three-dimensional point before verifying the matching consistency among the three sets of matching point pairs A plurality of corresponding light planes, wherein the light plane is a light plane corresponding to the selected point; the second determining unit is configured to determine a target light plane corresponding to the selected point from the plurality of light planes.
  • the second determination unit includes: a second acquisition subunit configured to acquire multiple pixel points corresponding to multiple pixel points on the same line as the selected point. light planes; the calculating subunit is configured to calculate the number of occurrences of each light plane, and use the light plane with the largest number of occurrences as the target light plane.
  • the verification unit includes: using the matching point corresponding to the target light plane as the target matching point; the second component subunit is configured to match the selected point with the target The matching points form target matching point pairs to obtain three sets of target matching point pairs; the verification subunit is configured to verify the consistency among the three sets of target matching point pairs.
  • the three frames of two-dimensional images are respectively the first frame of two-dimensional images, the second frame of two-dimensional images and the third frame of two-dimensional images
  • the verification subunit includes :
  • the first acquisition module is configured to acquire the selected point in the first frame of two-dimensional image, and the target light plane 1 determined by the selected point through the matching of the first frame of two-dimensional image and the second frame of two-dimensional image;
  • the second The acquisition module is configured to acquire the selected point in the first frame of two-dimensional image, and the target light plane 2 determined by the selected point after matching the first frame of two-dimensional image with the third frame of two-dimensional image;
  • the third acquisition module It is configured to obtain the target matching point of the selected point in the first frame of two-dimensional image in the second frame of two-dimensional image;
  • the fourth acquisition module is configured to obtain the target matching point in the second frame of two-dimensional image passing through the second The target light plane three determined by matching the frame two-dimensional image with the third frame two-
  • the first determination subunit includes: a fifth acquisition unit configured to acquire an epipolar equation corresponding to another frame of two-dimensional image from a selected point; The intersection points of the epipolar line equation and multiple lines in another frame of two-dimensional image are used as multiple candidate matching points.
  • the processing device for the three-dimensional scanning includes a processor and a memory, the projection unit, the acquisition unit, the first determination unit, the verification unit and the reconstruction unit are all stored in the memory as program units, and the above-mentioned programs stored in the memory are executed by the processor. Program unit to realize the corresponding function.
  • the processor includes a kernel, and the kernel fetches corresponding program units from the memory.
  • One or more kernels can be set, and the reconstruction of the three-dimensional points of the measured object can be realized by adjusting the kernel parameters.
  • Memory may include non-permanent memory in computer-readable media, in the form of random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM), memory including at least one memory chip.
  • RAM random access memory
  • ROM read-only memory
  • flash RAM flash memory
  • An embodiment of the present invention provides a computer-readable storage medium on which a program is stored, and when the program is executed by a processor, a processing method for three-dimensional scanning is realized.
  • An embodiment of the present invention provides a processor, and the processor is used to run a program, wherein a processing method for three-dimensional scanning is executed when the program is running.
  • An embodiment of the present invention provides a device.
  • the device includes a processor, a memory, and a program stored in the memory and operable on the processor.
  • the processor executes the program, the following steps are implemented: Projecting multiple lines to the measured object through the pattern projector Object surface; collect two-dimensional images of the surface of the measured object through three cameras, correspondingly obtain three frames of two-dimensional images; determine the matching point pairs between two two frames of three two-dimensional images, correspondingly obtain three sets of matching point pairs; verify the matching Matching consistency between point pairs; 3D reconstruction is performed on matching point pairs with matching consistency to obtain 3D points on the surface of the measured object.
  • determining matching point pairs between two pairs of three frames of two-dimensional images includes: obtaining multiple lines in three frames of two-dimensional images, wherein the lines are composed of multiple pixel points , using a pixel point on the line in one frame of two-dimensional image as the selected point, and determining multiple candidate matching points matching the selected point in another frame of two-dimensional image; based on the principle of triangulation, the selected point and multiple candidate points Matching points for 3D reconstruction to obtain a plurality of first candidate 3D points; determining the first candidate 3D points satisfying preset conditions as second candidate 3D points, wherein there are multiple second candidate 3D points; The candidate matching points corresponding to the points form matching point pairs with the selected points.
  • the method before verifying the matching consistency between the three sets of matching point pairs, the method further includes: acquiring a plurality of light planes corresponding to the second candidate three-dimensional point, wherein the light plane is a light plane corresponding to the selected point; Determine the target light plane corresponding to the selected point from a plurality of light planes.
  • determining the target light plane corresponding to the selected point from multiple light planes includes: obtaining multiple light planes corresponding to multiple pixel points on the same line as the selected point; times, the light plane with the most occurrences is taken as the target light plane.
  • verify the matching consistency between the three sets of matching point pairs including: using the matching point corresponding to the target light plane as the target matching point; combining the selected point and the target matching point to form a target matching point pair to obtain three sets of target Match point pairs; verify the consistency between the three sets of target match point pairs.
  • the three frames of two-dimensional images are respectively the first frame of two-dimensional images, the second frame of two-dimensional images and the third frame of two-dimensional images, and verifying the matching consistency between the three groups of target matching point pairs includes: obtaining the first A selected point in the two-dimensional image frame, and a target light plane 1 determined by matching the selected point with the two-dimensional image in the first frame and the two-dimensional image in the second frame; acquiring the selected point in the two-dimensional image in the first frame, and Target light plane 2 determined by matching the selected point through the first frame of two-dimensional image and the third frame of two-dimensional image; obtaining the target matching point of the selected point in the first frame of two-dimensional image in the second frame of two-dimensional image; Obtain the target light plane 3 determined by matching the target matching points in the second two-dimensional image with the second frame two-dimensional image and the third frame two-dimensional image; determine whether the target light plane one, target light plane two and target light plane three are The same light plane; when the target light plane 1, target light plane 2 and target light
  • determining a plurality of candidate matching points matching the selected point in another frame of two-dimensional image includes: obtaining the selected point corresponding to another An epipolar line equation of a two-dimensional image; the intersections of the epipolar line equation and multiple lines in another two-dimensional image are used as multiple candidate matching points.
  • the devices in this article can be servers, PCs, PADs, mobile phones, etc.
  • the present disclosure also provides a computer program product, which, when executed on a data processing device, is adapted to execute a program initialized with the following method steps: projecting multiple lines onto the surface of a measured object through a pattern projector; Measure the two-dimensional image of the surface of the object, and obtain three frames of two-dimensional images correspondingly; determine the matching point pairs between two two-dimensional images of the three frames, and obtain three sets of matching point pairs correspondingly; verify the matching consistency between the matching point pairs; Perform 3D reconstruction on the matched point pairs with matching consistency to obtain 3D points on the surface of the measured object.
  • determining matching point pairs between two pairs of three frames of two-dimensional images includes: obtaining multiple lines in three frames of two-dimensional images, wherein the lines are composed of multiple pixel points , using a pixel point on the line in one frame of two-dimensional image as the selected point, and determining multiple candidate matching points matching the selected point in another frame of two-dimensional image; based on the principle of triangulation, the selected point and multiple candidate points Matching points for 3D reconstruction to obtain a plurality of first candidate 3D points; determining the first candidate 3D points satisfying preset conditions as second candidate 3D points, wherein there are multiple second candidate 3D points; The candidate matching points corresponding to the points form matching point pairs with the selected points.
  • the method before verifying the matching consistency between the three sets of matching point pairs, the method further includes: acquiring a plurality of light planes corresponding to the second candidate three-dimensional point, wherein the light plane is a light plane corresponding to the selected point; Determine the target light plane corresponding to the selected point from a plurality of light planes.
  • determining the target light plane corresponding to the selected point from multiple light planes includes: obtaining multiple light planes corresponding to multiple pixel points on the same line as the selected point; times, the light plane with the most occurrences is taken as the target light plane.
  • verify the matching consistency between the three sets of matching point pairs including: using the matching point corresponding to the target light plane as the target matching point; combining the selected point and the target matching point to form a target matching point pair to obtain three sets of target Match point pairs; verify the consistency between the three sets of target match point pairs.
  • the three frames of two-dimensional images are respectively the first frame of two-dimensional images, the second frame of two-dimensional images and the third frame of two-dimensional images, and verifying the matching consistency between the three groups of target matching point pairs includes: obtaining the first A selected point in the two-dimensional image frame, and a target light plane 1 determined by matching the selected point with the two-dimensional image in the first frame and the two-dimensional image in the second frame; acquiring the selected point in the two-dimensional image in the first frame, and Target light plane 2 determined by matching the selected point through the first frame of two-dimensional image and the third frame of two-dimensional image; obtaining the target matching point of the selected point in the first frame of two-dimensional image in the second frame of two-dimensional image; Obtain the target light plane 3 determined by matching the target matching points in the second two-dimensional image with the second frame two-dimensional image and the third frame two-dimensional image; determine whether the target light plane one, target light plane two and target light plane three are The same light plane; when the target light plane 1, target light plane 2 and target light
  • determining a plurality of candidate matching points matching the selected point in another frame of two-dimensional image includes: obtaining the selected point corresponding to another An epipolar line equation of a two-dimensional image; the intersections of the epipolar line equation and multiple lines in another two-dimensional image are used as multiple candidate matching points.
  • the embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions
  • the device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.
  • a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
  • processors CPUs
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • Memory may include non-permanent storage in computer readable media, in the form of random access memory (RAM) and/or nonvolatile memory such as read only memory (ROM) or flash RAM.
  • RAM random access memory
  • ROM read only memory
  • flash RAM flash random access memory
  • Computer-readable media including both permanent and non-permanent, removable and non-removable media, can be implemented by any method or technology for storage of information.
  • Information may be computer readable instructions, data structures, modules of a program, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Flash memory or other memory technology, Compact Disc Read-Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, Magnetic tape cartridge, tape magnetic disk storage or other magnetic storage device or any other non-transmission medium that can be used to store information that can be accessed by a computing device.
  • computer-readable media excludes transitory computer-readable media, such as modulated data signals and carrier waves.
  • the embodiments of the present application may be provided as methods, systems or computer program products. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
  • a computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • the three-dimensional scanning processing method, device and three-dimensional scanning equipment provided by the embodiments of the present disclosure project multiple lines to the surface of the measured object through the pattern projector; collect two-dimensional images of the surface of the measured object through three cameras, and obtain correspondingly Three frames of two-dimensional images; determining matching point pairs between two pairs of the three frames of two-dimensional images, and correspondingly obtaining three sets of matching point pairs; verifying the matching consistency between the matching point pairs; pairing the matching point pairs with matching consistency Matching point pairs for three-dimensional reconstruction to obtain three-dimensional points on the surface of the measured object solves the problem that the scanning efficiency of binocular laser scanning in the related art cannot meet some high scanning efficiency requirements.
  • the method to improve scanning efficiency is to increase the scanning line
  • increasing the number of scanning lines in a binocular scanning system will lead to a sharp drop in matching accuracy.
  • the two-dimensional images of the surface of the measured object are collected by three cameras to obtain three frames of two-dimensional images, and the three frames of two-dimensional images are matched in pairs to obtain three sets of matching point pairs, and three-dimensional matching point pairs with matching consistency are obtained. Reconstruct to obtain the three-dimensional points on the surface of the object to be measured, thereby achieving the effect of improving the accuracy of matching.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Graphics (AREA)
  • Geometry (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Length Measuring Devices By Optical Means (AREA)

Abstract

本公开公开了一种三维扫描的处理方法、装置和三维扫描设备。该方法包括：通过图案投射器投射多线至被测物体表面；通过三个相机采集所述被测物体表面的二维图像，对应得到三帧二维图像；确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证所述匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点。通过本公开，解决了相关技术中双目激光扫描的扫描效率无法满足一些高扫描效率要求的场景，提高扫描效率的方法是通过增加扫描线数，但是在双目扫描系统中增加扫描线数，会导致出现匹配的准确性急剧下降的问题。

Description

三维扫描的处理方法、装置和三维扫描设备
本公开要求于2021年12月17日提交中国专利局、申请号为202111556738.2、申请名称“三维扫描的处理方法、装置和三维扫描设备”的中国专利申请的优先权，其全部内容通过引用结合在本公开中。
技术领域
本公开涉及三维扫描技术领域，具体而言，涉及一种三维扫描的处理方法、装置和三维扫描设备。
背景技术
随着手持式三维扫描技术的日益成熟，三维扫描技术在工业上的应用越来越广泛。在实际扫描场景中，一些被测的工业物体往往是体积庞大的工业件，对扫描效率有较高的要求。现有最常见的扫描方案为双目激光扫描，而双目激光扫描的激光线数在20线以下，该扫描效率无法满足一些高扫描效率要求的场景。而想要提高扫描效率则需要增加扫描线数，但在双目立体视觉系统中扫描线数的增加会导致匹配的准确性的降低。
针对相关技术中双目激光扫描的扫描效率无法满足一些高扫描效率要求的场景，提高扫描效率的方法是通过增加扫描线数，但是在双目扫描系统中增加扫描线数，会导致出现匹配的准确性急剧下降的问题，目前尚未提出有效的解决方案。
发明内容
本公开的主要目的在于提供一种三维扫描的处理方法、装置和三维扫描设备，以解决相关技术中双目激光扫描的扫描效率无法满足一些高扫描效率要求的场景，提高扫描效率的方法是通过增加扫描线数，但是在双目扫描系统中增加扫描线数，会导致出现匹配的准确性急剧下降的问题。
为了实现上述目的，根据本公开的一个方面，提供了一种三维扫描的处理方法。该方法包括：通过图案投射器投射多线至被测物体表面；通过三个相机采集所述被测物体表面的二维图像，对应得到三帧二维图像；确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证所述匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点。
进一步地，确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对， 包括：获取所述三帧二维图像中的多个线条，其中，所述线条由多个像素点组成，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与所述选定点匹配的多个候选匹配点；基于三角法原理将所述选定点与所述多个候选匹配点进行三维重建，得到多个第一候选三维点；将满足预设条件的第一候选三维点确定为第二候选三维点，其中，所述第二候选三维点有多个；将所述第二候选三维点对应的所述候选匹配点与所述选定点组成所述匹配点对。
进一步地，在验证所述三组匹配点对之间的匹配一致性之前，所述方法还包括：获取所述第二候选三维点对应的多个光平面，其中，所述光平面为所述选定点对应的光平面；从所述多个光平面中确定所述选定点对应的目标光平面。
进一步地，从所述多个光平面中确定所述选定点对应的目标光平面，包括：获取与所述选定点在同一线条上的多个像素点对应的多个光平面；计算每个光平面出现的次数，将出现次数最多的光平面作为目标光平面。
进一步地，验证所述三组匹配点对之间的匹配一致性，包括：将所述目标光平面对应的匹配点作为目标匹配点；将所述选定点与所述目标匹配点组成目标匹配点对，得到三组目标匹配点对；验证所述三组目标匹配点对之间的一致性。
进一步地，所述三帧二维图像分别为第一帧二维图像，第二帧二维图像和第三帧二维图像，验证所述三组目标匹配点对之间的匹配一致性包括:获取所述第一帧二维图像中的选定点，以及所述选定点经过所述第一帧二维图像与所述第二帧二维图像匹配确定的目标光平面一；获取所述第一帧二维图像中的选定点，以及所述选定点经过所述第一帧二维图像与所述第三帧二维图像匹配确定的目标光平面二；获取所述第一帧二维图像中的选定点在所述第二帧二维图像中的目标匹配点；获取所述第二帧二维图像中所述目标匹配点经过所述第二帧二维图像与所述第三帧二维图像匹配确定的目标光平面三；判断所述目标光平面一，所述目标光平面二和所述目标光平面三是否为同一光平面；当所述目标光平面一，所述目标光平面二和所述目标光平面三为同一光平面时，则所述三组目标匹配点对具有匹配一致性。
进一步地，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与所述选定点匹配的多个候选匹配点，包括：获取所述选定点对应到所述另一帧二维图像的极线方程；将所述极线方程与所述另一帧二维图像中的多个线条的交点作为所述多个候选匹配点。
为了实现上述目的，根据本公开的另一方面，提供了一种三维扫描设备。该设备中包括三个相机，将所述三个相机两两组合，得到三个双目系统；其中，所述三个双目系统用于采集被测物体表面的二维图像，对应得到三帧二维图像，其中，确定所述 三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证所述匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点。
为了实现上述目的，根据本公开的另一方面，提供了一种三维扫描的处理装置。该装置包括：投射单元，被配置为通过图案投射器投射多线至被测物体表面；采集单元，被配置为通过三个相机采集所述被测物体表面的二维图像，对应得到三帧二维图像；第一确定单元，被配置为确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证单元，被配置为验证所述匹配点对之间的匹配一致性；重建单元，被配置为对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点。
进一步地，所述确定单元包括：第一获取子单元，被配置为获取所述三帧二维图像中的多个线条，其中，所述线条由多个像素点组成，第一确定子单元，被配置为将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与所述选定点匹配的多个候选匹配点；重建子单元，被配置为基于三角法原理将所述选定点与所述多个候选匹配点进行三维重建，得到多个第一候选三维点；第二确定子单元，被配置为将满足预设条件的第一候选三维点确定为第二候选三维点，其中，所述第二候选三维点有多个；第一组成子单元，被配置为将所述第二候选三维点对应的所述候选匹配点与所述选定点组成所述匹配点对。
进一步地，所述装置还包括：获取单元，被配置为在验证所述三组匹配点对之间的匹配一致性之前，获取所述第二候选三维点对应的多个光平面，其中，所述光平面为所述选定点对应的光平面；第二确定单元，被配置为从所述多个光平面中确定所述选定点对应的目标光平面。
进一步地，所述第二确定单元包括：第二获取子单元，被配置为获取与所述选定点在同一线条上的多个像素点对应的多个光平面；计算子单元，被配置为计算每个光平面出现的次数，将出现次数最多的光平面作为目标光平面。
进一步地，所述验证单元包括：将所述目标光平面对应的匹配点作为目标匹配点；第二组成子单元，被配置为将所述选定点与所述目标匹配点组成目标匹配点对，得到三组目标匹配点对；验证子单元，被配置为验证所述三组目标匹配点对之间的一致性。
进一步地，所述三帧二维图像分别为第一帧二维图像，第二帧二维图像和第三帧二维图像，所述验证子单元包括：第一获取模块，被配置为获取所述第一帧二维图像中的选定点，以及所述选定点经过所述第一帧二维图像与所述第二帧二维图像匹配确定的目标光平面一；第二获取模块，被配置为获取所述第一帧二维图像中的选定点， 以及所述选定点经过所述第一帧二维图像与所述第三帧二维图像匹配确定的目标光平面二；第三获取模块，被配置为获取所述第一帧二维图像中的选定点在所述第二帧二维图像中的目标匹配点；第四获取模块，被配置为获取所述第二帧二维图像中所述目标匹配点经过所述第二帧二维图像与所述第三帧二维图像匹配确定的目标光平面三；判断模块，被配置为判断所述目标光平面一，所述目标光平面二和所述目标光平面三是否为同一光平面；当所述目标光平面一，所述目标光平面二和所述目标光平面三为同一光平面时，则所述三组目标匹配点对具有匹配一致性。
进一步地，所述第一确定子单元包括：第五获取单元，被配置为获取所述选定点对应到所述另一帧二维图像的极线方程；将所述极线方程与所述另一帧二维图像中的多个线条的交点作为所述多个候选匹配点。
为了实现上述目的，根据本公开的另一方面，提供了一种计算机可读存储介质，所述存储介质包括存储的程序，其中，所述程序执行上述任意一项所述的三维扫描的处理方法。
为了实现上述目的，根据本公开的另一方面，提供了一种处理器，所述处理器用于运行程序，其中，所述程序运行时执行上述任意一项所述的三维扫描的处理方法。
通过本公开，采用以下步骤：通过图案投射器投射多线至被测物体表面；通过三个相机采集所述被测物体表面的二维图像，对应得到三帧二维图像；确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证所述匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点，解决了相关技术中双目激光扫描的扫描效率无法满足一些高扫描效率要求的场景，提高扫描效率的方法是通过增加扫描线数，但是在双目扫描系统中增加扫描线数，会导致出现匹配的准确性急剧下降的问题。通过三个相机采集被测物体表面的二维图像，得到三帧二维图像，将三帧二维图像两两进行匹配，得到三组匹配点对，对具有匹配一致性的匹配点对进行三维重建，得到待测物体表面的三维点，进而达到了提高匹配的准确性的效果。
附图说明
构成本公开的一部分的附图用来提供对本公开的进一步理解，本公开的示意性实施例及其说明用于解释本公开，并不构成对本公开的不当限定。在附图中：
图1是根据本公开实施例提供的三维扫描的处理方法的流程图；
图2是根据本公开实施例提供的可选的三帧二维图像的示意图；
图3是根据本公开实施例提供的可选的三维扫描设备的示意图；
图4是根据本公开实施例提供的三维扫描的处理装置的示意图。
具体实施方式
需要说明的是，在不冲突的情况下，本公开中的实施例及实施例中的特征可以相互组合。下面将参考附图并结合实施例来详细说明本公开。
为了使本技术领域的人员更好地理解本公开方案，下面将结合本公开实施例中的附图，对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本公开一部分的实施例，而不是全部的实施例。基于本公开中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都应当属于本公开保护的范围。
需要说明的是，本公开的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本公开的实施例。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
下面结合优选的实施步骤对本发明进行说明，图1是根据本公开实施例提供的三维扫描的处理方法的流程图，如图1所示，该方法包括如下步骤：
步骤S101，通过图案投射器投射多线至被测物体表面。
例如，通过三维扫描设备的图案投射器投射多条扫描线(例如，可以使用激光扫描线)到需要构建三维模型的物体的表面。
步骤S102，通过三个相机采集被测物体表面的二维图像，对应得到三帧二维图像。
通过三维扫描设备的三个相机(例如，三个相机分别为Cam L，Cam M和Cam R)采集要构建三维模型的物体表面的二维图像，对应得到三帧二维图像(例如，图2所示的三帧二维图像)。三帧二维图像分别为第一帧二维图像，第二帧二维图像和第三帧二维图像。在本公开的实施例中，三维扫描设备为手持式三维扫描仪，可相对被测物体移动扫描。优选地，三个相机同步采集，即三个相机在第一时间同步采集、在第二时间同步采集，直至扫描完成，确保三个相机获取的图像的对应性。三个相机也可不 同步，但需确保采集时间间隔极小，三维扫描设备相对被测物体的位置几乎不变。如果将三维扫描设备在每次采集被测物体时固定使用，即将三维扫描设备固定在第一位置采集被测物体的一个部位、再固定至第二位置采集被测物体的另一个部位，直至被测物体被扫描完成，则可不限制三个相机的采集是否同步。
步骤S103，确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对。
对三帧二维图像进行两两匹配，得到对应的三组匹配点对。例如，图2所示，第一帧二维图像上一点A(1，2)，A(1，2)在第二帧二维图像匹配得到的匹配点坐标为B 1(2，3)；那么第一帧二维图像与第二帧二维图像的匹配点对为A(1，2)和B 1(2，3)；A(1，2)在第三帧二维图像匹配得到的匹配点坐标为C(1，3)；那么第一帧二维图像与第三帧二维图像的匹配点对为A(1，2)和C(1，3)；第二帧二维图像上的B 1(2，3)在第三帧二维图像匹配得到的匹配点为C(1，3)；那么第二帧二维图像与第三帧二维图像的匹配点对为B 1(2，3)和C(1，3)。
步骤S104，验证匹配点对之间的匹配一致性。
对得到的三组匹配点对验证是否具有匹配一致性。
步骤S105，对具有匹配一致性的匹配点对进行三维重建，得到被测物体表面的三维点。
当三组匹配点对具有匹配一致性时，则对该点进行三维重建，得到需要构建三维模型的物体的三维点，以此三维点构建物体的三维模型。
通过上述步骤，通过三个相机采集物体的二维图像，得到三帧二维图像，对三帧二维图像进行两两匹配，得到匹配点对，对于具有匹配一致性的匹配点对进行三维重建，以此得到物体的三维模型，提高了三维重建匹配的准确性。
可选地，在本公开实施例提供的三维扫描的处理方法中，确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对，包括：获取三帧二维图像中的多个线条，其中，线条由多个像素点组成，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与选定点匹配的多个候选匹配点；基于三角法原理将选定点与多个候选匹配点进行三维重建，得到多个第一候选三维点；将满足预设条件的第一候选三维点确定为第二候选三维点，其中，第二候选三维点有多个；将第二候选三维点对应的候选匹配点与选定点组成匹配点对。
例如，如图2所示，三帧二维图像中包含多个线条(即图案投射器发射的多个扫面线)，线条是由多个像素点组成，图2中Cam L对应的图像为第一帧二维图像，Cam M对应的为第二帧二维图像，Cam H对应的为第三帧二维图像。将第一帧二维图像中 的A点作为选定点，在第二帧二维图像中找到与A匹配的多个候选匹配点。例如，在第二帧二维图像中的候选匹配点有B1，B2，B3，B4，B5和B6。通过三角法将A点与候选匹配点B 1，B 2，B 3，B 4，B 5和B 6进行三维重建，得到多个第一候选三维点O 1，O 2，O 3，O 4，O 5和O 6。对第一候选三维点进行初步筛选，将满足预设条件的作为第二候选三维点。筛选方法为计算第一候选三维点到所有扫描线对应的光平面Plane(k)的距离Distance(k)。如果某一第一候选三维点的距离值在设定距离阈值之内，即Distance(k)<dist TH，那么说明该第一候选三维点在k号光平面内(第k条投射线)，也就是该第一候选三维点满足预设要求。假设第二候选三维点有O 1，O 2和O 3，那么O 1，O 2和O 3对应的候选匹配点B 1，B 2和B 3与A点组成匹配点对。通过上述方法得到第一帧二维图像上每一个像素点在第二帧二维图像上的匹配点。同样采用同样的方法得到第一帧二维图像与第三帧二维图像的匹配点对，以及第二帧二维图像与第三帧二维图像的匹配点对。
通过获取三帧二维图像中的多个线条，便于确定三帧二维图像中各个像素点的坐标。通过三帧二维图像进行两两匹配，得到三组匹配点对，可以有效提高匹配的准确性。
可选地，在本公开实施例提供的三维扫描的处理方法中，在验证三组匹配点对之间的匹配一致性之前，该方法还包括：获取第二候选三维点对应的多个光平面，其中，光平面为选定点对应的光平面；从多个光平面中确定选定点对应的目标光平面。
在对三组匹配点对进行验证之前，需要确定最佳匹配点对，理论上，在两帧二维图像进行匹配，得到的多个匹配点对中只有一对是正确的，所以需要确定最佳的匹配点对。首先第二候选三维点是通过计算该点与所有扫描线对应的光平面Plane(k)的距离Distance(k)来确定的。当某一第二候选三维点的距离值在设定距离阈值之内，即Distance(k)<dist TH，说明该点在k号光平面内(第k条投射线)。也就是说一个第二候选三维点会对应一个光平面。例如，第二候选三维点O 1，O 2和O 3对应的光平面分别为K 1，K 2和K 3。在这些光平面中选出最佳光平面(也就是上述的目标光平面)。
对得到的三组匹配点对进行进一步的筛选，能够更加准确的得到物体表面的重建三维点。
可选地，在本公开实施例提供的三维扫描的处理方法中，从多个光平面中确定选定点对应的目标光平面，包括：获取与选定点在同一线条上的多个像素点对应的多个光平面；计算每个光平面出现的次数，将出现次数最多的光平面作为目标光平面。
例如，从光平面K 1，K 2和K 3中选取最佳光平面的方法为：首先得到选定点在同一线条上的多个像素点对应的多个光平面。如图2所示，与A点在同一线条的多个像 素点A 1，A 2，A 3和A 4。A 1对应的多个光平面为K 1和K 2；A 2对应的多个光平面为K 1和K 4；A 3对应的多个光平面为K 1，K 2和K 3；A 4对应的多个光平面为K 1和K 2；计算每个光平面出现的次数，将出现次数最多的光平面作为最佳光平面，也就是A点对应的最佳光平面为K 1。在实际使用过程中，三帧二维图像中的线条可能是间断的线条，那么通过获取选定点在线条上搜索邻域内的多个像素点对应的多个光平面来确定选定点对应的最佳光平面。
通过与选定点在同一线条上的多个像素点对应的多个光平面来确定选定点对应的最佳光平面，提高了像素点对应的光平面的准确性。
可选地，在本公开实施例提供的三维扫描的处理方法中，验证三组匹配点对之间的匹配一致性，包括：将目标光平面对应的匹配点作为目标匹配点；将选定点与目标匹配点组成目标匹配点对，得到三组目标匹配点对；验证三组目标匹配点对之间的一致性。
验证三组匹配点对之间的匹配一致性，只需要验证最佳匹配点对(即上述目标匹配点对)之间的一致性。最佳光平面对应的匹配点即为最佳匹配点。例如，第一帧二维图像与第二帧二维图像的最佳匹配点对为A(1，2)和B 1(2，3)，对应的最佳光平面为K 1；第一帧二维图像与第三帧二维图像的最佳匹配点对为A(1，2)和C(1，3)，对应的最佳光平面为K 1；第二帧二维图像与第三帧二维图像的最佳匹配点对为B 1(2，3)和C(1，3)，对应的最佳光平面为K 1；
只有通过最佳匹配点对重建的三维点才有可能是需要构建三维模型的物体表面的三维点，所以需要验证最佳匹配点对之间的一致性。
可选地，在本公开实施例提供的三维扫描的处理方法中，所述三帧二维图像分别为第一帧二维图像，第二帧二维图像和第三帧二维图像，验证所述三组目标匹配点对之间的匹配一致性包括:获取所述第一帧二维图像中的选定点，以及所述选定点经过所述第一帧二维图像与所述第二帧二维图像匹配确定的目标光平面一；获取所述第一帧二维图像中的选定点，以及所述选定点经过所述第一帧二维图像与所述第三帧二维图像匹配确定的目标光平面二；获取所述第一帧二维图像中的选定点在所述第二帧二维图像中的目标匹配点；获取所述第二帧二维图像中所述目标匹配点经过所述第二帧二维图像与所述第三帧二维图像匹配确定的目标光平面三；判断所述目标光平面一，所述目标光平面二和所述目标光平面三是否为同一光平面；当所述目标光平面一，所述目标光平面二和所述目标光平面三为同一光平面时，则所述三组目标匹配点对具有匹配一致性。
例如，第一帧二维图像上一点A(1，2)，A(1，2)在第二帧二维图像匹配得到 的最佳匹配点坐标为B 1(2，3)；那么第一帧二维图像与第二帧二维图像的最佳匹配点对为A(1，2)和B 1(2，3)，对应的最佳光平面为K 1。A(1，2)在第三帧二维图像匹配得到的最佳匹配点坐标为C(1，3)；那么第一帧二维图像与第三帧二维图像的最佳匹配点对为A(1，2)和C(1，3)，对应的最佳光平面为K 1。第二帧二维图像上的B 1(2，3)在第三帧二维图像匹配得到的最佳匹配点为C(1，3)；那么第二帧二维图像与第三帧二维图像的最佳匹配点对为B 1(2，3)和C(1，3)，对应的最佳光平面为K 1。依据上述例子，两两匹配得到的最佳光平面为同一个，即K 1，说明三组最佳匹配点对具有匹配一致性。
通过验证匹配点对所对应的光平面来确定是否具有匹配一致性，这是由于只有光平面为同一个才能够说明匹配点对在三维中是同一个三维点，所以上述步骤进一步提高了匹配的准确性。
可选地，在本公开实施例提供的三维扫描的处理方法中，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与选定点匹配的多个候选匹配点，包括：获取选定点对应到另一帧二维图像的极线方程；将极线方程与另一帧二维图像中的多个线条的交点作为多个候选匹配点。
例如，在第二帧二维图像中确定第一帧二维图像上的A点的多个候选匹配点时，需要通过极线方程来实现，首先依据Cam L和Cam M的相对位置计算得到极线方程，例如，图2中所标注的示意性极线方程。将极线方程与另一帧二维图像中的多个线条的交点作为多个候选匹配点。
本公开实施例提供的三维扫描的处理方法，通过图案投射器投射多线至被测物体表面；通过三个相机采集所述被测物体表面的二维图像，对应得到三帧二维图像；确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证所述匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点，解决了相关技术中双目激光扫描的扫描效率无法满足一些高扫描效率要求的场景，提高扫描效率的方法是通过增加扫描线数，但是在双目扫描系统中增加扫描线数，会导致出现匹配的准确性急剧下降的问题。通过三个相机采集被测物体表面的二维图像，得到三帧二维图像，将三帧二维图像两两进行匹配，得到三组匹配点对，对具有匹配一致性的匹配点对进行三维重建，得到待测物体表面的三维点，进而达到了提高匹配的准确性的效果。
本公开实施例还提供了一种三维扫描设备，三维扫描设备中包括三个相机，将三个相机两两组合，得到三个双目系统；其中，三个双目系统用于采集被测物体表面的二维图像，对应得到三帧二维图像，其中，确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证匹配点对之间的匹配一致性；对具有匹配一致性的匹配 点对进行三维重建，得到被测物体表面的三维点。
如图3所示，为根据本公开实施例提供的可选的三维扫描设备的示意图，其中，Cam L，Cam M和Cam R为三个相机，三个相机两个组成三个双目系统。Projector为图案投射器。
需要说明的是，在附图的流程图示出的步骤可以在诸如一组计算机可执行指令的计算机系统中执行，并且，虽然在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤。
本公开实施例还提供了一种三维扫描的处理装置，需要说明的是，本公开实施例的三维扫描的处理装置可以用于执行本公开实施例所提供的用于三维扫描的处理方法。以下对本公开实施例提供的三维扫描的处理装置进行介绍。
图4是根据本公开实施例的三维扫描的处理装置的示意图。如图4所示，该装置包括：投射单元401，采集单元402，第一确定单元403，验证单元404和重建单元405。
投射单元401，被配置为通过图案投射器投射多线至被测物体表面。
采集单元402，被配置为通过三个相机采集被测物体表面的二维图像，对应得到三帧二维图像。
第一确定单元403，被配置为确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对。
验证单元404，被配置为验证匹配点对之间的匹配一致性。
重建单元405，被配置为对具有匹配一致性的匹配点对进行三维重建，得到被测物体表面的三维点。
本公开实施例提供的三维扫描的处理装置，通过投射单元401通过图案投射器投射多线至被测物体表面；采集单元402通过三个相机采集被测物体表面的二维图像，对应得到三帧二维图像；第一确定单元403确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证单元404验证匹配点对之间的匹配一致性；重建单元405对具有匹配一致性的匹配点对进行三维重建，得到被测物体表面的三维点，解决了相关技术中双目激光扫描的扫描效率无法满足一些高扫描效率要求的场景，提高扫描效率的方法是通过增加扫描线数，但是在双目扫描系统中增加扫描线数，会导致出现匹配的准确性急剧下降的问题，通过三个相机采集被测物体表面的二维图像，得到三帧二维图像，将三帧二维图像两两进行匹配，得到三组匹配点对，对具有匹配一致性的匹配点对进行三维重建，得到待测物体表面的三维点，进而达到了提高匹配的准确性的效果。
可选地，在本公开实施例提供的三维扫描的处理装置中，确定单元包括：第一获取子单元，被配置为在确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对之前，获取三帧二维图像中的多个线条，其中，线条由多个像素点组成，第一确定子单元，被配置为将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与选定点匹配的多个候选匹配点；重建子单元，被配置为基于三角法原理将选定点与多个候选匹配点进行三维重建，得到多个第一候选三维点；第二确定子单元，被配置为将满足预设条件的第一候选三维点确定为第二候选三维点，其中，第二候选三维点有多个；第一组成子单元，被配置为将第二候选三维点对应的候选匹配点与选定点组成匹配点对。
可选地，在本公开实施例提供的三维扫描的处理装置中，该装置还包括：获取单元，被配置为在验证三组匹配点对之间的匹配一致性之前，获取第二候选三维点对应的多个光平面，其中，光平面为选定点对应的光平面；第二确定单元，被配置为从多个光平面中确定选定点对应的目标光平面。
可选地，在本公开实施例提供的三维扫描的处理装置中，第二确定单元包括：第二获取子单元，被配置为获取与选定点在同一线条上的多个像素点对应的多个光平面；计算子单元，被配置为计算每个光平面出现的次数，将出现次数最多的光平面作为目标光平面。
可选地，在本公开实施例提供的三维扫描的处理装置中，验证单元包括：将目标光平面对应的匹配点作为目标匹配点；第二组成子单元，被配置为将选定点与目标匹配点组成目标匹配点对，得到三组目标匹配点对；验证子单元，被配置为验证三组目标匹配点对之间的一致性。
可选地，在本公开实施例提供的三维扫描的处理装置中，三帧二维图像分别为第一帧二维图像，第二帧二维图像和第三帧二维图像，验证子单元包括：第一获取模块，被配置为获取第一帧二维图像中的选定点，以及选定点经过第一帧二维图像与第二帧二维图像匹配确定的目标光平面一；第二获取模块，被配置为获取第一帧二维图像中的选定点，以及选定点经过第一帧二维图像与第三帧二维图像匹配确定的目标光平面二；第三获取模块，被配置为获取第一帧二维图像中的选定点在第二帧二维图像中的目标匹配点；第四获取模块，被配置为获取第二帧二维图像中目标匹配点经过第二帧二维图像与第三帧二维图像匹配确定的目标光平面三；判断模块，被配置为判断目标光平面一，目标光平面二和目标光平面三是否为同一光平面；当目标光平面一，目标光平面二和目标光平面三为同一光平面时，则三组目标匹配点对具有匹配一致性。
可选地，在本公开实施例提供的三维扫描的处理装置中，第一确定子单元包括：第五获取单元，被配置为获取选定点对应到另一帧二维图像的极线方程；将极线方程 与另一帧二维图像中的多个线条的交点作为多个候选匹配点。
所述三维扫描的处理装置包括处理器和存储器，上述投射单元，采集单元，第一确定单元，验证单元和重建单元等均作为程序单元存储在存储器中，由处理器执行存储在存储器中的上述程序单元来实现相应的功能。
处理器中包含内核，由内核去存储器中调取相应的程序单元。内核可以设置一个或以上，通过调整内核参数来实现被测物体三维点的重建。
存储器可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)，存储器包括至少一个存储芯片。
本发明实施例提供了一种计算机可读存储介质，其上存储有程序，该程序被处理器执行时实现三维扫描的处理方法。
本发明实施例提供了一种处理器，处理器用于运行程序，其中，程序运行时执行三维扫描的处理方法。
本发明实施例提供了一种设备，设备包括处理器、存储器及存储在存储器上并可在处理器上运行的程序，处理器执行程序时实现以下步骤：通过图案投射器投射多线至被测物体表面；通过三个相机采集被测物体表面的二维图像，对应得到三帧二维图像；确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到被测物体表面的三维点。
可选地，确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对，包括：获取三帧二维图像中的多个线条，其中，线条由多个像素点组成，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与选定点匹配的多个候选匹配点；基于三角法原理将选定点与多个候选匹配点进行三维重建，得到多个第一候选三维点；将满足预设条件的第一候选三维点确定为第二候选三维点，其中，第二候选三维点有多个；将第二候选三维点对应的候选匹配点与选定点组成匹配点对。
可选地，在验证三组匹配点对之间的匹配一致性之前，该方法还包括：获取第二候选三维点对应的多个光平面，其中，光平面为选定点对应的光平面；从多个光平面中确定选定点对应的目标光平面。
可选地，从多个光平面中确定选定点对应的目标光平面，包括：获取与选定点在同一线条上的多个像素点对应的多个光平面；计算每个光平面出现的次数，将出现次数最多的光平面作为目标光平面。
可选地，验证三组匹配点对之间的匹配一致性，包括：将目标光平面对应的匹配点作为目标匹配点；将选定点与目标匹配点组成目标匹配点对，得到三组目标匹配点对；验证三组目标匹配点对之间的一致性。
可选地，三帧二维图像分别为第一帧二维图像，第二帧二维图像和第三帧二维图像，验证三组目标匹配点对之间的匹配一致性包括:获取第一帧二维图像中的选定点，以及选定点经过第一帧二维图像与第二帧二维图像匹配确定的目标光平面一；获取第一帧二维图像中的选定点，以及选定点经过第一帧二维图像与第三帧二维图像匹配确定的目标光平面二；获取第一帧二维图像中的选定点在第二帧二维图像中的目标匹配点；获取第二帧二维图像中目标匹配点经过第二帧二维图像与第三帧二维图像匹配确定的目标光平面三；判断目标光平面一，目标光平面二和目标光平面三是否为同一光平面；当目标光平面一，目标光平面二和目标光平面三为同一光平面时，则三组目标匹配点对具有匹配一致性。
可选地，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与选定点匹配的多个候选匹配点，包括：获取选定点对应到另一帧二维图像的极线方程；将极线方程与另一帧二维图像中的多个线条的交点作为多个候选匹配点。本文中的设备可以是服务器、PC、PAD、手机等。
本公开还提供了一种计算机程序产品，当在数据处理设备上执行时，适于执行初始化有如下方法步骤的程序：通过图案投射器投射多线至被测物体表面；通过三个相机采集被测物体表面的二维图像，对应得到三帧二维图像；确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到被测物体表面的三维点。
可选地，确定三帧二维图像两两之间的匹配点对，对应得到三组匹配点对，包括：获取三帧二维图像中的多个线条，其中，线条由多个像素点组成，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与选定点匹配的多个候选匹配点；基于三角法原理将选定点与多个候选匹配点进行三维重建，得到多个第一候选三维点；将满足预设条件的第一候选三维点确定为第二候选三维点，其中，第二候选三维点有多个；将第二候选三维点对应的候选匹配点与选定点组成匹配点对。
可选地，在验证三组匹配点对之间的匹配一致性之前，该方法还包括：获取第二候选三维点对应的多个光平面，其中，光平面为选定点对应的光平面；从多个光平面中确定选定点对应的目标光平面。
可选地，从多个光平面中确定选定点对应的目标光平面，包括：获取与选定点在同一线条上的多个像素点对应的多个光平面；计算每个光平面出现的次数，将出现次 数最多的光平面作为目标光平面。
可选地，验证三组匹配点对之间的匹配一致性，包括：将目标光平面对应的匹配点作为目标匹配点；将选定点与目标匹配点组成目标匹配点对，得到三组目标匹配点对；验证三组目标匹配点对之间的一致性。
可选地，三帧二维图像分别为第一帧二维图像，第二帧二维图像和第三帧二维图像，验证三组目标匹配点对之间的匹配一致性包括:获取第一帧二维图像中的选定点，以及选定点经过第一帧二维图像与第二帧二维图像匹配确定的目标光平面一；获取第一帧二维图像中的选定点，以及选定点经过第一帧二维图像与第三帧二维图像匹配确定的目标光平面二；获取第一帧二维图像中的选定点在第二帧二维图像中的目标匹配点；获取第二帧二维图像中目标匹配点经过第二帧二维图像与第三帧二维图像匹配确定的目标光平面三；判断目标光平面一，目标光平面二和目标光平面三是否为同一光平面；当目标光平面一，目标光平面二和目标光平面三为同一光平面时，则三组目标匹配点对具有匹配一致性。
可选地，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与选定点匹配的多个候选匹配点，包括：获取选定点对应到另一帧二维图像的极线方程；将极线方程与另一帧二维图像中的多个线条的交点作为多个候选匹配点。
本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中，计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
存储器可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)。存储器是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括，但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带，磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质，可用于存储可以被计算设备访问的信息。按照本文中的界定，计算机可读介质不包括暂存电脑可读媒体(transitory media)，如调制的数据信号和载波。
还需要说明的是，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、商品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、商品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括要素的过程、方法、商品或者设备中还存在另外的相同要素。
本领域技术人员应明白，本申请的实施例可提供为方法、系统或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
以上仅为本申请的实施例而已，并不用于限制本申请。对于本领域技术人员来说，本申请可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本申请的权利要求范围之内。
工业实用性
本公开实施例所提供的三维扫描的处理方法、装置和三维扫描设备，通过图案投射器投射多线至被测物体表面；通过三个相机采集所述被测物体表面的二维图像，对应得到三帧二维图像；确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证所述匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点，解决了相关技术中双目激光扫描的扫描效率无法满足一些高扫描效率要求的场景，提高扫描效率的方法是通过增加扫描线数，但是在双目扫描系统中增加扫描线数，会导致出现匹配的准确性急剧下降的问题。通过三个相机采集被测物体表面的二维图像，得到三帧二维图像，将三帧二维图像两两进行匹配，得到三组匹配点对，对具有匹配一致性的匹配点对进行三维重建，得到待测物体表面的三维点，进而达到了提高匹配的准确性的效果。

Claims (11)

  1. 一种三维扫描的处理方法，包括：
    通过图案投射器投射多线至被测物体表面；
    通过三个相机采集所述被测物体表面的二维图像，对应得到三帧二维图像；
    确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；
    验证所述匹配点对之间的匹配一致性；
    对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点。
  2. 根据权利要求1所述的方法，其中，确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对，包括：
    获取所述三帧二维图像中的多个线条，其中，所述线条由多个像素点组成，
    将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与所述选定点匹配的多个候选匹配点；
    基于三角法原理将所述选定点与所述多个候选匹配点进行三维重建，得到多个第一候选三维点；
    将满足预设条件的第一候选三维点确定为第二候选三维点，其中，所述第二候选三维点有多个；
    将所述第二候选三维点对应的所述候选匹配点与所述选定点组成所述匹配点对。
  3. 根据权利要求2所述的方法，其中，在验证所述三组匹配点对之间的匹配一致性之前，所述方法还包括：
    获取所述第二候选三维点对应的多个光平面，其中，所述光平面为所述选定点对应的光平面；
    从所述多个光平面中确定所述选定点对应的目标光平面。
  4. 根据权利要求3所述的方法，其中，从所述多个光平面中确定所述选定点对应的目标光平面，包括：
    获取与所述选定点在同一线条上的多个像素点对应的多个光平面；
    计算每个光平面出现的次数，将出现次数最多的光平面作为目标光平面。
  5. 根据权利要求4所述的方法，其中，验证所述三组匹配点对之间的匹配一致性，包括：
    将所述目标光平面对应的匹配点作为目标匹配点；
    将所述选定点与所述目标匹配点组成目标匹配点对，得到三组目标匹配点对；
    验证所述三组目标匹配点对之间的一致性。
  6. 根据权利要求5所述的方法，其中，其中，所述三帧二维图像分别为第一帧二维图像，第二帧二维图像和第三帧二维图像，验证所述三组目标匹配点对之间的匹配一致性包括:
    获取所述第一帧二维图像中的选定点，以及所述选定点经过所述第一帧二维图像与所述第二帧二维图像匹配确定的目标光平面一；
    获取所述第一帧二维图像中的选定点，以及所述选定点经过所述第一帧二维图像与所述第三帧二维图像匹配确定的目标光平面二；
    获取所述第一帧二维图像中的选定点在所述第二帧二维图像中的目标匹配点；
    获取所述第二帧二维图像中所述目标匹配点经过所述第二帧二维图像与所述第三帧二维图像匹配确定的目标光平面三；
    判断所述目标光平面一，所述目标光平面二和所述目标光平面三是否为同一光平面；
    当所述目标光平面一，所述目标光平面二和所述目标光平面三为同一光平面时，则所述三组目标匹配点对具有匹配一致性。
  7. 根据权利要求2所述的方法，其中，将一帧二维图像中线条上一个像素点作为选定点，确定另一帧二维图像中与所述选定点匹配的多个候选匹配点，包括：
    获取所述选定点对应到所述另一帧二维图像的极线方程；
    将所述极线方程与所述另一帧二维图像中的多个线条的交点作为所述多个候选匹配点。
  8. 一种三维扫描设备，其中，所述三维扫描设备中包括三个相机，将所述三个相机两两组合，得到三个双目系统；其中，所述三个双目系统用于采集被测物体表面的二维图像，对应得到三帧二维图像，其中，确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；验证所述匹配点对之间的匹配一致性；对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点。
  9. 一种三维扫描的处理装置，包括：
    投射单元，被配置为通过图案投射器投射多线至被测物体表面；
    采集单元，被配置为通过三个相机采集所述被测物体表面的二维图像，对应得到三帧二维图像；
    第一确定单元，被配置为确定所述三帧二维图像两两之间的匹配点对，对应得到三组匹配点对；
    验证单元，被配置为验证所述匹配点对之间的匹配一致性；
    重建单元，被配置为对具有匹配一致性的匹配点对进行三维重建，得到所述被测物体表面的三维点。
  10. 一种计算机可读存储介质，其中，所述存储介质包括存储的程序，其中，所述程序执行权利要求1至7中任意一项所述的三维扫描的处理方法。
  11. 一种处理器，其中，所述处理器用于运行程序，其中，所述程序运行时执行权利要求1至7中任意一项所述的三维扫描的处理方法。
PCT/CN2022/139716 2021-12-17 2022-12-16 三维扫描的处理方法、装置和三维扫描设备 WO2023109960A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111556738.2 2021-12-17
CN202111556738.2A CN116266379A (zh) 2021-12-17 2021-12-17 三维扫描的处理方法、装置和三维扫描设备

Publications (1)

Publication Number Publication Date
WO2023109960A1 true WO2023109960A1 (zh) 2023-06-22

Family

ID=86743932

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/139716 WO2023109960A1 (zh) 2021-12-17 2022-12-16 三维扫描的处理方法、装置和三维扫描设备

Country Status (2)

Country Link
CN (1) CN116266379A (zh)
WO (1) WO2023109960A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108288292A (zh) * 2017-12-26 2018-07-17 中国科学院深圳先进技术研究院 一种三维重建方法、装置及设备
US20190353477A1 (en) * 2016-09-14 2019-11-21 Hangzhou Scantech Company Limited Three-dimensional sensor system and three-dimensional data acquisition method
WO2021088481A1 (zh) * 2019-11-08 2021-05-14 南京理工大学 一种基于条纹投影的高精度动态实时360度全方位点云获取方法
CN113390357A (zh) * 2021-07-08 2021-09-14 南京航空航天大学 一种基于双目多线结构光的铆钉齐平度测量方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190353477A1 (en) * 2016-09-14 2019-11-21 Hangzhou Scantech Company Limited Three-dimensional sensor system and three-dimensional data acquisition method
CN108288292A (zh) * 2017-12-26 2018-07-17 中国科学院深圳先进技术研究院 一种三维重建方法、装置及设备
WO2021088481A1 (zh) * 2019-11-08 2021-05-14 南京理工大学 一种基于条纹投影的高精度动态实时360度全方位点云获取方法
CN113390357A (zh) * 2021-07-08 2021-09-14 南京航空航天大学 一种基于双目多线结构光的铆钉齐平度测量方法

Also Published As

Publication number Publication date
CN116266379A (zh) 2023-06-20

Similar Documents

Publication Publication Date Title
Zioulis et al. Spherical view synthesis for self-supervised 360 depth estimation
KR102096806B1 (ko) 3차원 센서 시스템 및 3차원 데이터 획득방법
US20200334840A1 (en) Three-Dimensional Scanning System and Scanning Method Thereof
JP6897563B2 (ja) 画像処理装置と画像処理方法およびプログラム
CN109461181A (zh) 基于散斑结构光的深度图像获取方法及系统
US20190012804A1 (en) Methods and apparatuses for panoramic image processing
US20210044787A1 (en) Three-dimensional reconstruction method, three-dimensional reconstruction device, and computer
US20120176478A1 (en) Forming range maps using periodic illumination patterns
US9025862B2 (en) Range image pixel matching method
CN111678459B (zh) 三维扫描方法、三维扫描系统和计算机可读存储介质
JP2018511874A (ja) ３次元モデリング方法及び装置
CN105241397A (zh) 基于结构光的实时测量拼接方法及其设备
CN109974623B (zh) 基于线激光和双目视觉的三维信息获取方法和装置
JP6580761B1 (ja) 偏光ステレオカメラによる深度取得装置及びその方法
CN106225676B (zh) 三维测量方法、装置及系统
CN102903101B (zh) 使用多台相机进行水面数据采集与重建的方法
JP2017510793A (ja) ２つのカメラからの曲線の集合の構造化光整合
WO2018120168A1 (zh) 一种视觉检测方法及系统
WO2018001252A1 (zh) 一种投射单元及包括该单元的拍摄装置、处理器、成像设备
JP2010276433A (ja) 撮像装置、画像処理装置及び距離計測装置
CN110825079A (zh) 一种地图构建方法及装置
Zhang et al. Synthetic aperture based on plenoptic camera for seeing through occlusions
JP2013101464A (ja) 画像処理装置および画像処理方法
US11803982B2 (en) Image processing device and three-dimensional measuring system
CN105430375B (zh) 一种检测全景摄像机的拼接适用距离的方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22906705

Country of ref document: EP

Kind code of ref document: A1