CN116671898A - Hololens-based three-dimensional human body measurement method and system - Google Patents
Hololens-based three-dimensional human body measurement method and system Download PDFInfo
- Publication number
- CN116671898A CN116671898A CN202310652604.3A CN202310652604A CN116671898A CN 116671898 A CN116671898 A CN 116671898A CN 202310652604 A CN202310652604 A CN 202310652604A CN 116671898 A CN116671898 A CN 116671898A
- Authority
- CN
- China
- Prior art keywords
- human body
- point cloud
- point
- dimensional
- dimensional human
- 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
- 238000000691 measurement method Methods 0.000 title claims abstract description 9
- 238000005259 measurement Methods 0.000 claims abstract description 76
- 238000000034 method Methods 0.000 claims abstract description 49
- 230000006870 function Effects 0.000 claims abstract description 28
- 238000007781 pre-processing Methods 0.000 claims abstract description 21
- 238000005070 sampling Methods 0.000 claims abstract description 16
- 238000000605 extraction Methods 0.000 claims abstract description 11
- 230000004927 fusion Effects 0.000 claims abstract description 11
- 238000005457 optimization Methods 0.000 claims abstract description 7
- 238000006243 chemical reaction Methods 0.000 claims abstract description 4
- 230000011218 segmentation Effects 0.000 claims description 30
- 238000013461 design Methods 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 11
- 230000009466 transformation Effects 0.000 claims description 11
- 208000013860 rhabdoid tumor of the kidney Diseases 0.000 claims description 10
- 238000012800 visualization Methods 0.000 claims description 10
- 238000009499 grossing Methods 0.000 claims description 9
- 238000003064 k means clustering Methods 0.000 claims description 8
- 230000008447 perception Effects 0.000 claims description 8
- 238000012805 post-processing Methods 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 230000003213 activating effect Effects 0.000 claims description 2
- 238000010276 construction Methods 0.000 claims description 2
- 238000001514 detection method Methods 0.000 claims description 2
- 230000001131 transforming effect Effects 0.000 claims description 2
- 238000011872 anthropometric measurement Methods 0.000 claims 1
- 230000000007 visual effect Effects 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 8
- 230000000295 complement effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 101000958041 Homo sapiens Musculin Proteins 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013079 data visualisation Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 102000046949 human MSC Human genes 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/107—Measuring physical dimensions, e.g. size of the entire body or parts thereof
- A61B5/1072—Measuring physical dimensions, e.g. size of the entire body or parts thereof measuring distances on the body, e.g. measuring length, height or thickness
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/107—Measuring physical dimensions, e.g. size of the entire body or parts thereof
- A61B5/1079—Measuring physical dimensions, e.g. size of the entire body or parts thereof using optical or photographic means
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Pathology (AREA)
- Medical Informatics (AREA)
- Dentistry (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a holonens-based three-dimensional human body measurement method, which comprises the following steps: step 1, acquiring rough three-dimensional human body grid data; step 2, preprocessing coarse three-dimensional human body grid data; step 3, performing conversion, approximate voxel sampling, key point extraction and optimization, coarse registration, fine registration and fusion to obtain a complete human body point cloud; step 4, performing measurement, namely completing interception, projection, extraction of convex hull vertexes and Euclidean distance summation; and 5, designing a three-dimensional human body measurement system to realize the functions of the steps 1 to 4. The invention also discloses a HoloLens-based three-dimensional human body measuring device which is used for realizing the HoloLens-based three-dimensional human body measuring method. The method and the device reduce the measurement cost and the acquisition difficulty of the three-dimensional size of the human body and obviously improve the measurement precision.
Description
Technical Field
The invention belongs to the technical field of three-dimensional human body measurement, relates to a three-dimensional human body measurement method based on holonens, and further relates to a three-dimensional human body measurement system based on holonens.
Background
Along with development and introduction of three-dimensional measurement technology, human body measurement is changed from traditional contact measurement to non-contact measurement, and the non-contact measurement can utilize a computer algorithm to realize automatic measurement on three-dimensional human body data acquired by scanning equipment, and has the characteristics of simplicity and convenience in operation, strong expansibility, high precision, high efficiency and the like.
How to perform high-efficiency and high-precision registration is a key preprocessing step for guaranteeing the integrity of three-dimensional human body data in the three-dimensional human body measurement technology. Under the condition of acquiring complete three-dimensional human body data, whether the three-dimensional segmentation technology is accurate or not can directly relate to accurate measurement of human body dimensions. The complete three-dimensional model can be divided into sub-parts, the dividing speed and accuracy are critical to the analysis of the characteristic dimension of the human body, and the three-dimensional division becomes a very important step in the measurement technology of three-dimensional human body data.
Although the three-dimensional human body data measurement technology has basic application in various fields, the three-dimensional human body data measurement technology is still used in academic and industrial fields and has a distance from popularization in life, so that the three-dimensional human body data measurement technology has a huge progress space. The holonens scanning data is used for three-dimensional human body data measurement, and the holonens scanning data is a research direction with popularization and application values.
Disclosure of Invention
The invention aims to provide a holonens-based three-dimensional human body measurement method, which solves the problems that in the prior art, an additional environment needs to be built and three-dimensional data is difficult to acquire rapidly, and reduces the three-dimensional human body measurement cost and the operation difficulty.
It is another object of the present invention to provide a holonens-based three-dimensional anthropometric system.
The technical scheme adopted by the invention is that the three-dimensional human body measurement method based on HoloLens is implemented according to the following steps:
step 1, acquiring rough three-dimensional human body grid data;
step 2, preprocessing coarse three-dimensional human body grid data;
step 3, performing conversion, approximate voxel sampling, key point extraction and optimization, coarse registration, fine registration and fusion to obtain a complete human body point cloud;
step 4, performing measurement, namely completing interception, projection, extraction of convex hull vertexes and Euclidean distance summation;
and 5, realizing the visualization of each function of the steps 1 to 4 by using a three-dimensional human body measurement system.
The invention adopts another technical proposal that a three-dimensional human body measuring device based on HoloLens is used for realizing the three-dimensional human body measuring method based on HoloLens,
the Hololens-based three-dimensional human body measuring device integrates the algorithm functions of the steps 1 to 4, develops a Hololens scanning data-oriented three-dimensional human body measuring system, and the main control flow of the three-dimensional human body measuring system is described as follows:
firstly recompilating the VTK by using CMake locally, and adding Qt support for the VTK; then copying the 'QVTKWIdgetPlugin. Dll' generated by VTK compiling to the designator path of msvc_64 version of Qt; after the operation is finished, a QVTKWIDget control is added in the Qt Designer of msvc_64 version;
newly creating a QtWidgetsapplication project on a VS, opening a UI file in the project by using a Qt Designer to perform interface design, selecting a corresponding control from a part list, manually dragging the corresponding control to a design window on the right side, and designing the display position and the area of a button by using reasonable layout according to a system module and function demand logic;
and (3) carrying out the architecture design of the three-dimensional human body measurement system based on VS2019 and Qt, respectively integrating the algorithm functions in the steps 1-4 and correspondingly dividing the algorithm functions into a preprocessing module, a point cloud registration module and a three-dimensional human body segmentation and measurement module for three-dimensional human body grid data, developing corresponding system program sub-interfaces for different functional modules, respectively presetting a plurality of UI keys for preprocessing, post-processing and measuring the three-dimensional human body data, and facilitating the use of users.
The invention has the beneficial effects that the invention comprises two aspects:
1) The problems of poor convenience, non-ideal visual effect and the like in a default Hololens equipment space mapping grid data acquisition mode can be avoided by deploying a data acquisition mode of the MRTK space perception system; the adopted preprocessing flow for denoising and smoothing of the rough three-dimensional human body grid data can solve the problems of low resolution, multiple noise holes, uneven topological structure and the like of the three-dimensional human body grid data obtained after the Hololens is deployed with the MRTK space sensing system; the adopted Point cloud registration method based on the Modified RANSAC and the Point-to-Plane ICP carries out high-efficiency and high-precision registration fusion on the Point cloud data converted from the three-dimensional human body grid data, so that the registration speed can be greatly improved while higher registration precision is ensured; the method is characterized in that a K-Means clustering method is used, human body point cloud is divided into 8 parts in a self-defined initial clustering center mode, measurement of corresponding parts is completed based on local human body point cloud after clustering segmentation, global measurement problems on a whole point cloud model are converted into local measurement problems, and measurement efficiency is improved; the Hololens-based three-dimensional human body measurement system is developed based on the realized three-dimensional data preprocessing, post-processing and human body measurement technology design, performs preprocessing on three-dimensional human body grid data acquired by Hololens, registers point cloud data converted from the three-dimensional human body grid data, performs point cloud segmentation on the complete point cloud after registration and fusion, and performs automatic measurement based on the segmented local human body point cloud, thereby meeting the measurement requirement of a human body in a standard posture for Hololens scanning data.
2) The method not only reduces the measurement cost and the acquisition difficulty of the three-dimensional size of the human body, but also provides a preprocessing flow, a point cloud registration method, a point cloud clustering segmentation method and an automatic measurement method after clustering segmentation, which are suitable for the three-dimensional human body grid data acquired by Hololens equipment, and the combined use can further improve the efficiency of three-dimensional human body measurement and obviously improve the measurement precision.
Drawings
FIG. 1 is a schematic overall flow diagram of the method of the present invention;
fig. 2a is a holonens scanned original human three-dimensional data visualization, fig. 2b is an outlier removal visualization, fig. 2c is a heavy topology visualization, fig. 2d is a smoothing visualization, fig. 2e is a hole complement visualization, fig. 2f is a poisson disc sampling visualization;
FIG. 3a is a visualized image for Point cloud data, FIG. 3b is a Modified approximate voxel sampling visualized image, FIG. 3c is an ISS3D keypoint extraction and optimization visualized image, FIG. 3D is a Modified RANSAC coarse registration visualized image, FIG. 3e is a Point-to-Plane ICP fine registration visualized image, FIG. 3f is a Point cloud fusion visualized image;
FIG. 4a is a complete human point cloud visual image, FIG. 4b is a K-Means cluster segmentation visual image, and FIG. 4c is a visual image of a K-Means custom initial cluster center, respectively;
fig. 5a is a view of obtaining a local point cloud visual image of a part to be measured, fig. 5b is a view of intercepting a point cloud visual image of the height of the part to be measured, fig. 5c is a view of plane projection of the point cloud visual image of the height to be measured, and fig. 5d is a view of extracting convex hull vertices visual image by a convex hull algorithm.
Detailed Description
The invention will be described in detail below with reference to the drawings and the detailed description.
Referring to fig. 1, the three-dimensional anthropometric method of the present invention is implemented as follows:
step 1, acquiring coarse three-dimensional human body grid data,
an MRTK space perception system is deployed on Hololens equipment, (the MRTK space perception system is Spatial Awareness System, which belongs to the prior art),
firstly, starting and activating a space perception system in an MRTK tool on a Unity platform, configuring and compiling relevant attributes of a space grid observer (Spatial Mesh Observer) in the MRTK space perception system, and deploying the observer on Hololens by using VS after compiling and generating items;
then, the HoloLens equipment is worn, and the wearer checks the reconstruction state of the three-dimensional human body model at the gazing position in real time to acquire original three-dimensional human body data and obtain rough three-dimensional human body grid data.
Step 2, preprocessing coarse three-dimensional human body grid data, wherein the specific process is as follows:
2.1 Denoising in a mode of connectivity among triangular patches and limiting the number of triangular patches of the triangular cluster;
2.2 Under the premise of ensuring the precision, utilizing a unified local smoothing operator to optimize and output the vertex position and the edge direction in the rough three-dimensional human body grid data, re-grid the curved surface to an isotropic triangle grid, and re-dividing the grid topology to ensure that the grid has a uniform and regular topological structure;
2.3 Using Taubin algorithm to popularize classical discrete Fourier analysis into two-dimensional discrete surface signals, namely defining a function on the surface of a polyhedron with arbitrary topology, simplifying the surface smoothing problem into a low-pass filtering problem, and carrying out smoothing filtering treatment to obtain three-dimensional human body grid data with smooth surface and no obvious shrinkage;
2.4 For each hole of the three-dimensional human body grid data obtained in the step 2.3), firstly finding out all boundary edges of the holes, and then sequencing the boundary edges; finding out two boundary edges with the smallest angle from all the boundary edges after sequencing, and (additionally) adding a third boundary edge to form a new triangular patch; and finally, setting a termination condition, and continuously iterating to complete the completion of all the holes.
Step 3, converting, approximate voxel sampling, extracting and optimizing key points, coarse registration, fine registration and fusion to obtain a complete human body point cloud,
the Point cloud registration method based on the Modified RANSAC and the Point-to-Plane ICP carries out registration fusion on the Point cloud data converted from the three-dimensional human body grid data, and the specific process is as follows:
3.1 The three-dimensional human body grid data acquired by Hololens equipment and subjected to pretreatment is converted into a point cloud format through poisson disk sampling, and a source point cloud is defined as P 1 The target point cloud is P 2 ；
3.2 To source point cloud P 1 Improved approximate voxel sampling, construction of a new point set P using a hash function to quickly approximate the centroid of non-empty voxels a {p a A=1, 2,. }; traversing point cloud P using Kd-tree algorithm a Find and point cloud P a Nearest neighbor point of each point in (3)Constitute a new point set P b {p b B=1, 2,. }; for the down-sampled point cloud P b Extracting ISS3D key points to form a new point set P c {p c C=1, 2,. }; optimizing key points by using a direction vector threshold value to form a new point set P d {p d ,d＝1,2,...}；
Similarly, the target point cloud P 2 Improved approximate voxel sampling and ISS3D key point extraction and optimization to form a new point set P q {p q ,q＝1,2,...}；
3.3 Point set P according to source point cloud d N (n is more than or equal to 3) random sampling points p i I e {1,., n } and point set P of the target point cloud q Corresponding point q in (a) i I epsilon {1, & gt, n }, establishing FPFH characteristic description, solving an approximate transformation matrix by utilizing a Modified RANSAC algorithm, and performing coarse registration;
3.4 Fine registration is carried out on the Point cloud after coarse registration by using the Point-to-Plane ICP algorithm to obtain an optimal transformation matrix M opt (the approximate transformation matrix is the initial imprecise transformation matrix obtained by coarse registration, and the optimal transformation matrix is the final optimal transformation matrix solved on the basis of the approximate transformation matrix), at this time, the source point cloud P 1 Interior point to target point cloud P 2 The distance of the plane where the corresponding point is located is minimum, and the point in the source point cloud is p i ＝(p ix ,p iy ,p iz ,1) T The corresponding point in the target point cloud is q i ＝(q ix ,q iy ,q iz ,1) T Point p i Where the unit normal vector is n i ＝(n ix ,n iy ,n iz ,0) T ；
Iterative solution of an optimal transformation matrix M opt M is carried out on the source point cloud opt Transforming, under a specific selected error metric, to minimize the distance from a point in the source point cloud to a plane in which a point corresponding to the target point cloud is located, where the expression is:
3.5 After finishing the fine registration, fusing the well registered point clouds to obtain the complete human body point cloud.
Step 4, measuring, completing interception, projection and Euclidean distance summation,
the method for clustering the human body point cloud obtained in the step 3.5) is divided into eight parts by using a K-Means clustering method and by a self-defining initial clustering center mode, and then the measurement of corresponding parts is respectively completed based on the local human body point cloud segmented by the K-Means clustering method, wherein the specific process is as follows:
4.1 The initial clustering center is customized, the center of eight segmentation parts are selected through the 'Point list picking' function of CloudCompare software, the eight segmentation parts are respectively a head, a left arm, a right arm, a chest, a waist, a hip, a left leg and a right leg, the coordinate value of the center of the eight segmentation parts is set as the initial clustering center of K-Means clustering, iteration is continued, and the average value of cluster sample points generated by each iteration is used as the clustering center until the clustering result tends to be stable;
4.2 Calculating the height H of the detection object according to the maximum value and the minimum value of the z coordinate in the data of the human body point cloud; determining the height of the part to be measured according to the linear relation between the height of each characteristic part of the human body and the height H; the three circumferences are the maximum circumferences on the horizontal slice within a certain range of the corresponding height positions respectively, and the corresponding division point clouds are intercepted;
preferably, point clouds with the height of the characteristic part being 0.01m up and down are intercepted, and the data of the subsequent girth measurement are ensured to approach to a true value by the point clouds in enough height;
4.3 Plane projection is carried out on the intercepted segmentation point cloud to obtain a plane point set which approximates human body circumference, a RANSAC algorithm is adopted to carry out plane fitting, and the intercepted segmentation point cloud is projected on a fitting plane;
4.4 Searching a minimum convex set surrounding the characteristic point cloud by using a convex hull algorithm, namely, a convex polygon with the minimum area in which a certain point set is completely contained;
4.5 Calculating Euclidean distance between adjacent vertexes of the convex polygon obtained by cutting the contour, summing to obtain measured circumference, and generating a convex hull vertex point set (X) i ,Y i ) The expression of the sum of Euclidean distances L is:
wherein X is i ，Y i Is a coordinate expression of the convex hull vertex point set, i=1, 2,3, …, and n represents the point set.
Step 5, using a three-dimensional human body measurement system to realize the visualization of each function of the steps 1 to 4,
the algorithm functions of the steps 1 to 4 are integrated, a three-dimensional human body measurement system facing HoloLens scanning data is designed and developed, the three-dimensional human body measurement system is a software program, and the following three small steps describe the main control flow (or called as software architecture) of the three-dimensional human body measurement system, and the method is specifically described as follows:
5.1 Recompilation of the VTK locally using CMake and adding Qt support for the VTK; then copying the QVTKWIdgetPlugin. Dll generated by VTK compilation (taking Release generation as an example) to the designator path of msvc_64 version of Qt; after the operation is finished, a QVTKWIDget control is added in the Qt Designer of msvc_64 version;
5.2 Newly creating a QtWidgetsapplication project on the VS, opening a UI file in the project by using a Qt Designer to perform interface design, selecting a corresponding control from a part list, manually dragging the corresponding control to a design window on the right side, and designing the display position and the area of a button by using reasonable layout according to a system module and function demand logic;
5.3 The algorithm functions of the steps 1 to 4 are respectively integrated and correspondingly divided into different functional modules, namely a preprocessing module of three-dimensional human body grid data, a point cloud registration module and a three-dimensional human body segmentation and measurement module, corresponding system program sub-interfaces are developed for the different functional modules, and a plurality of UI keys for preprocessing, post-processing and measuring the three-dimensional human body data are respectively preset, so that the system is convenient for a user to use.
The method can preprocess three-dimensional human body grid data; simultaneously registering and fusing to generate complete three-dimensional human body point cloud data; the segmentation of the complete human body point cloud and the automatic measurement of the local human body point cloud based on the clustering segmentation can be completed; the combined use can effectively finish the basic dimension measurement of the three-dimensional human body under the standard posture, has strong expansibility, and the measurement result can be automatically displayed in a text display box of the three-dimensional human body measurement system interface.
The device is embodied in the step 5, which is equivalent to integrating all functions of the steps 1 to 4 to develop three interfaces, namely a three-dimensional human body grid preprocessing interface, a three-dimensional human body point cloud registration interface and a three-dimensional human body segmentation and human body measurement interface, so as to form a complete system. The related technical processes of each step 1 to step 4 correspond to one function button in the three interfaces developed in step 5, the developed functions support real-time visualization and can be converted into images, fig. 2a, fig. 2b, fig. 2c, fig. 2d, fig. 2e and fig. 2f are respectively computer-visualized images of three-dimensional human grid data preprocessing and format conversion developed by a three-dimensional human grid preprocessing interface, fig. 3a, fig. 3b, fig. 3c, fig. 3d, fig. 3e and fig. 3f are respectively computer-visualized images of three-dimensional human point cloud data registration and fusion developed by a three-dimensional human point cloud registration interface, fig. 4a, fig. 4b and fig. 4c are respectively computer-visualized images of three-dimensional human point cloud cluster segmentation developed by a three-dimensional human body segmentation and human body measurement interface, and fig. 5a, fig. 5b, fig. 5c and fig. 5d are respectively computer-visualized images of automatic measurement of the circumferences developed by a three-dimensional human body segmentation and human body measurement interface.
Example 1
The measurement object Human1 of example 1;
firstly, measuring the calculated chest circumference, waistline, hip circumference and other size data according to the step process of the method; then, the comparison is made with the measurement truth value, and the specific comparison result is referred to table 1.
TABLE 1 example 1 automatic measurement result analysis
In the specific implementation process, this embodiment 1 is shown by visualization through the following drawings:
the visualized image of the three-dimensional human body grid data preprocessing flow step comprises six small images, wherein fig. 2a is holonens scanning original human body three-dimensional data, fig. 2b is removing outlier noise, fig. 2c is heavy topology, fig. 2d is smoothing processing, fig. 2e is hole complement, and fig. 2f is poisson disk sampling.
The visualized image of the Point cloud registration step based on Modified RANSAC and Point-to-Plane ICP comprises six small images, wherein fig. 3a is visualized for Point cloud data, fig. 3b is improved approximate voxel sampling, fig. 3c is ISS3D key Point extraction and optimization, fig. 3D is Modified RANSAC coarse registration, fig. 3e is Point-to-Plane ICP fine registration, and fig. 3f is Point cloud fusion.
The 3D human body data display image comprises three small images, wherein, FIG. 4a is an image of a complete human body point cloud, FIG. 4b is K-Means clustering segmentation, and FIG. 4c is a K-Means custom initial clustering center.
The visualized image of the girth measurement after the point cloud segmentation and the cluster segmentation comprises four small images, wherein, fig. 5a is a local point cloud of a part to be detected, fig. 5b is a point cloud of intercepting the height of the part to be detected, fig. 5c is a plane projection point cloud of the height to be detected, and fig. 5d is a convex hull vertex extraction method.
Example 2
The measurement object Human2 of example 2;
firstly, measuring the calculated chest circumference, waistline, hip circumference and other size data according to the step process of the method; and then compared with the measured true values, and the specific comparison results refer to table 2.
TABLE 2 example 2 automatic measurement result analysis
The visual image during the procedure of example 2 is omitted here and is not shown.
Example 3
The measurement object Human3 of example 3;
firstly, measuring the calculated chest circumference, waistline, hip circumference and other size data according to the step process of the method;
and then compared with the measured true value, and the specific comparison result refers to table 3.
TABLE 3 example 3 automatic measurement result analysis
The visual image during the procedure of example 3 is omitted here and is not shown.
By comparing the dimensional data of chest circumference, waistline, hip circumference and the like measured by the three embodiments with the measurement true values, it is obvious that the measurement error percentage of the three circumferences measured by the method is less than 1.4%, and the feasibility and the accuracy of the method are fully illustrated.
In summary, the main purpose of the invention is to improve the efficiency and precision of three-dimensional human body measurement, firstly, deploy an MRTK space sensing system on Hololens equipment to complete the acquisition of three-dimensional human body grid data, then, pre-process the three-dimensional human body grid data, then, convert the three-dimensional human body grid data into point cloud data for registration, then, perform point cloud segmentation on the registered and fused complete human body point cloud and perform automatic measurement on the three-dimensional size of the human body based on the segmented local human body point cloud, finally integrate the functions, complete Qt development based on VS, and design and develop the three-dimensional human body measurement system based on Hololens. The method has complete functions, reduces the measurement cost and the acquisition difficulty of the human body size, and has high measurement efficiency, high measurement precision and high intelligent degree. Compared with the existing method, the method can effectively finish the basic size measurement of the three-dimensional human body data under the standard posture on the premise of lower cost and faster efficiency, has strong expansibility, and provides a new technical means for the three-dimensional human body measurement.
Claims (8)
1. The holonens-based three-dimensional human body measurement method is characterized by comprising the following steps of:
step 1, acquiring rough three-dimensional human body grid data;
step 2, preprocessing coarse three-dimensional human body grid data;
step 3, performing conversion, approximate voxel sampling, key point extraction and optimization, coarse registration, fine registration and fusion to obtain a complete human body point cloud;
step 4, performing measurement, namely completing interception, projection, extraction of convex hull vertexes and Euclidean distance summation;
and 5, realizing the visualization of each function of the steps 1 to 4 by using a three-dimensional human body measurement system.
2. The holonens-based three-dimensional anthropometric method according to claim 1, wherein in step 1, the specific process is: an MRTK spatial perception system is deployed on a holonens device,
firstly, starting and activating a space perception system in an MRTK tool on a Unity platform, configuring and compiling relevant attributes of a space grid observer in the MRTK space perception system, and deploying the observer on Hololens by using VS after compiling and generating a project;
then, the HoloLens equipment is worn, and the wearer checks the reconstruction state of the three-dimensional human body model at the gazing position in real time to acquire original three-dimensional human body data and obtain rough three-dimensional human body grid data.
3. The holonens-based three-dimensional anthropometric method according to claim 1, wherein in step 2, the specific process is:
2.1 Denoising in a mode of connectivity among triangular patches and limiting the number of triangular patches of the triangular cluster;
2.2 Utilizing a unified local smoothing operator to optimize and output vertex positions and edge directions in rough three-dimensional human body grid data, re-grid the curved surface to an isotropic triangle grid, and re-dividing the grid topology to ensure that the grid has a uniform and regular topological structure;
2.3 Using Taubin algorithm to popularize classical discrete Fourier analysis into two-dimensional discrete surface signals, namely defining a function on the surface of a polyhedron with arbitrary topology, simplifying the surface smoothing problem into a low-pass filtering problem, and carrying out smoothing filtering treatment to obtain three-dimensional human body grid data;
2.4 For each hole of the three-dimensional human body grid data obtained in the step 2.3), firstly finding out all boundary edges of the holes, and then sequencing the boundary edges; finding two boundary edges with the smallest angle from all the boundary edges after sequencing, and adding a third boundary edge to form a new triangular patch; and finally, setting a termination condition, and continuously iterating to complete the completion of all the holes.
4. The holonens-based three-dimensional anthropometric measurement method according to claim 1, wherein in step 3, a Point cloud registration method based on Modified RANSAC and Point-to-Plane ICP performs registration fusion on Point cloud data converted from three-dimensional anthropometric mesh data, and the specific process is as follows:
3.1 The three-dimensional human body grid data acquired by Hololens equipment and subjected to pretreatment is converted into a point cloud format through poisson disk sampling, and a source point cloud is defined as P 1 The target point cloud is P 2 ；
3.2 To source point cloud P 1 Improved approximate voxel sampling, construction of a new point set P using a hash function to quickly approximate the centroid of non-empty voxels a {p a A=1, 2,. }; traversing point cloud P using Kd-tree algorithm a Find and point cloud P a The nearest neighbor point of each point in (a) constitutes a new point set P b {p b B=1, 2,. }; for the down-sampled point cloud P b Extracting ISS3D key points to form a new point set P c {p c C=1, 2,. }; optimizing key points by using a direction vector threshold value to form a new point set P d {p d ,d＝1,2,...}；
Similarly, the target point cloud P 2 Improved approximate voxel sampling and ISS3D key point extraction and optimization to form new point setP q {p q ,q＝1,2,...}；
3.3 Point set P according to source point cloud d N (n is more than or equal to 3) random sampling points p i I e {1,., n } and point set P of the target point cloud q Corresponding point q in (a) i I epsilon {1, & gt, n }, establishing FPFH characteristic description, solving an approximate transformation matrix by utilizing a Modified RANSAC algorithm, and performing coarse registration;
3.4 Fine registration is carried out on the Point cloud after coarse registration by using the Point-to-Plane ICP algorithm to obtain an optimal transformation matrix M opt At this time, the source point cloud P 1 Interior point to target point cloud P 2 The distance of the plane where the corresponding point is located is minimum, and the point in the source point cloud is p i ＝(p ix ,p iy ,p iz ,1) T The corresponding point in the target point cloud is q i ＝(q ix ,q iy ,q iz ,1) T Point p i Where the unit normal vector is n i ＝(n ix ,n iy ,n iz ,0) T ；
Iterative solution of an optimal transformation matrix M opt M is carried out on the source point cloud opt Transforming, under a specific selected error metric, to minimize the distance from a point in the source point cloud to a plane in which a point corresponding to the target point cloud is located, where the expression is:
3.5 After finishing the fine registration, fusing the well registered point clouds to obtain the complete human body point cloud.
5. The holomens-based three-dimensional human body measurement method according to claim 1, wherein in step 4, a K-Means clustering method is used, the human body point cloud obtained in step 3.5) is divided into eight parts by Means of a self-defined initial clustering center, and the measurement of corresponding parts is completed based on the local human body point cloud divided by the K-Means clustering method, respectively, wherein the specific process is as follows:
4.1 The initial clustering center is customized, the center of eight segmentation parts are selected through the 'Point list picking' function of CloudCompare software, the eight segmentation parts are respectively a head, a left arm, a right arm, a chest, a waist, a hip, a left leg and a right leg, the coordinate value of the center of the eight segmentation parts is set as the initial clustering center of K-Means clustering, iteration is continued, and the average value of cluster sample points generated by each iteration is used as the clustering center until the clustering result tends to be stable;
4.2 Calculating the height H of the detection object according to the maximum value and the minimum value of the z coordinate in the data of the human body point cloud; determining the height of the part to be measured according to the linear relation between the height of each characteristic part of the human body and the height H; the three circumferences are the maximum circumferences on the horizontal slice within a certain range of the corresponding height positions respectively, and the corresponding division point clouds are intercepted;
4.3 Plane projection is carried out on the intercepted segmentation point cloud to obtain a plane point set which approximates human body circumference, a RANSAC algorithm is adopted to carry out plane fitting, and the intercepted segmentation point cloud is projected on a fitting plane;
4.4 Searching a minimum convex set surrounding the characteristic point cloud by using a convex hull algorithm, namely, a convex polygon with the minimum area in which a certain point set is completely contained;
4.5 Calculating Euclidean distance between adjacent vertexes of the convex polygon obtained by cutting the contour, summing to obtain measured circumference, and generating a convex hull vertex point set (X) i ,Y i ) The expression of the sum of Euclidean distances L is:
wherein X is i ，Y i Is a coordinate expression of the convex hull vertex point set, i=1, 2,3, …, and n represents the point set.
6. The holonens-based three-dimensional anthropometric method according to claim 1, wherein in step 5, the specific process is:
integrating the algorithm functions of the steps 1 to 4, designing and developing a holonens scanning data-oriented three-dimensional human body measurement system, and describing the main control flow of the three-dimensional human body measurement system by the following three small steps, wherein the following description is as follows:
5.1 Recompilation of the VTK locally using CMake and adding Qt support for the VTK; then copying the 'QVTKWIdgetPlugin. Dll' generated by VTK compiling to the designator path of msvc_64 version of Qt; after the operation is finished, a QVTKWIDget control is added in the Qt Designer of msvc_64 version;
5.2 Newly creating a QtWidgetsapplication project on the VS, opening a UI file in the project by using a Qt Designer to perform interface design, selecting a corresponding control from a part list, manually dragging the corresponding control to a design window on the right side, and designing the display position and the area of a button according to a system module and a logical reasonable layout of functional requirements;
5.3 The algorithm functions of the steps 1 to 4 are respectively integrated and correspondingly divided into a preprocessing module, a point cloud registration module and a three-dimensional human body segmentation and measurement module for three-dimensional human body grid data based on VS2019 and Qt, corresponding system program sub-interfaces are developed for different functional modules, and a plurality of UI keys for preprocessing, post-processing and measuring the three-dimensional human body data are respectively preset, so that the system is convenient for users to use.
7. A holonens-based three-dimensional anthropometric apparatus for implementing the holonens-based three-dimensional anthropometric method of claim 1, characterized in that:
the Hololens-based three-dimensional human body measuring device integrates the algorithm functions of the steps 1-4, and a three-dimensional human body measuring system facing Hololens scanning data is developed.
8. The holonens-based three-dimensional anthropometric apparatus of claim 7, wherein the main control flow of the three-dimensional anthropometric system is as follows:
firstly recompilating the VTK by using CMake locally, and adding Qt support for the VTK; then copying the 'QVTKWIdgetPlugin. Dll' generated by VTK compiling to the designator path of msvc_64 version of Qt; after the operation is finished, a QVTKWIDget control is added in the Qt Designer of msvc_64 version;
newly creating a QtWidgetsapplication project on a VS, opening a UI file in the project by using a Qt Designer to perform interface design, selecting a corresponding control from a part list, manually dragging the corresponding control to a design window on the right side, and designing the display position and the area of a button by using reasonable layout according to a system module and function demand logic;
and (3) carrying out the architecture design of the three-dimensional human body measurement system based on VS2019 and Qt, respectively integrating the algorithm functions in the steps 1-4 and correspondingly dividing the algorithm functions into a preprocessing module, a point cloud registration module and a three-dimensional human body segmentation and measurement module for three-dimensional human body grid data, developing corresponding system program sub-interfaces for different functional modules, respectively presetting a plurality of UI keys for preprocessing, post-processing and measuring the three-dimensional human body data, and facilitating the use of users.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310652604.3A CN116671898A (en) | 2023-06-02 | 2023-06-02 | Hololens-based three-dimensional human body measurement method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310652604.3A CN116671898A (en) | 2023-06-02 | 2023-06-02 | Hololens-based three-dimensional human body measurement method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116671898A true CN116671898A (en) | 2023-09-01 |
Family
ID=87790312
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310652604.3A Pending CN116671898A (en) | 2023-06-02 | 2023-06-02 | Hololens-based three-dimensional human body measurement method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116671898A (en) |
-
2023
- 2023-06-02 CN CN202310652604.3A patent/CN116671898A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | From laser-scanned data to feature human model: a system based on fuzzy logic concept | |
| Guo et al. | A survey on algorithms of hole filling in 3D surface reconstruction | |
| CN111696210A (en) | Point cloud reconstruction method and system based on three-dimensional point cloud data characteristic lightweight | |
| Attene et al. | Polygon mesh repairing: An application perspective | |
| Sun et al. | Cloud data modelling employing a unified, non-redundant triangular mesh | |
| Kim et al. | Garment pattern generation from body scan data | |
| CN101510316B (en) | Deformation design method for product STL model | |
| Chen et al. | Reverse engineering in the design of turbine blades–a case study in applying the MAMDP | |
| Hsieh | An efficient development of 3D surface registration by Point Cloud Library (PCL) | |
| Cohen et al. | Virtual reconstruction of archeological vessels using expert priors and intrinsic differential geometry information | |
| CN107730587A (en) | One kind is based on picture quick three-dimensional Interactive Modeling method | |
| Lai et al. | An incremental approach to feature aligned quad dominant remeshing | |
| Yingjie et al. | Improved moving least squares algorithm for directed projecting onto point clouds | |
| Tong et al. | 3D point cloud initial registration using surface curvature and SURF matching | |
| CN111862176B (en) | Three-dimensional oral cavity point cloud orthodontic front and back accurate registration method based on palatine fold | |
| CN112686916A (en) | Curved surface reconstruction system based on heterogeneous multi-region CT scanning data processing | |
| Zhong et al. | Triple screening point cloud registration method based on image and geometric features | |
| CN116671898A (en) | Hololens-based three-dimensional human body measurement method and system | |
| Huang et al. | Automatic CAD model reconstruction from multiple point clouds for reverse engineering | |
| Sahillioğlu et al. | Coarse-to-fine surface reconstruction from silhouettes and range data using mesh deformation | |
| Jarvis et al. | 3D shape reconstruction of small bodies from sparse features | |
| CN113298919A (en) | Skeleton extraction method of three-dimensional plant point cloud model | |
| Jia et al. | 3D personalized human modeling and deformation technology for garment CAD | |
| Zheng et al. | A new 3D contour extraction method for tooth cavity in a dental CAD/CAM system | |
| Pang et al. | Developing an image-based 3D model editing method |
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 |