Single-face deployment type three-dimensional human body measurement method, device and system based on 3D structured light

Abstract

The invention discloses a single-side deployment type three-dimensional human body measurement method, device and system based on 3D structured light. The 3D structured light camera module emits light with structural characteristics to the front and back of a to-be-detected human body in sequence, and captures the light reflected by the front and back to generate front depth map data and back depth map data. And carrying out point cloud conversion and spatial registration to obtain a point cloud data model. And matching the point cloud data model with the parameterized human body model, adjusting parameters of the parameterized human body model through an optimization algorithm, and minimizing the distance between the key point in the parameterized human body model and the corresponding key point in the point cloud data model. And finally, determining three-dimensional measurement data of the human body to be measured according to the parameterized human body model obtained after fitting is finished. The 3D structured light camera module can be deployed on one vertical face, the method is simple and convenient, a site does not need to be specially built to build an independent closed space, the user measurement experience is improved, and the cost of three-dimensional human body measurement is reduced.

Description

Technical Field

[0001] The present application relates to the field of computer vision technology, and in particular to a single-sided deployment three-dimensional human body measurement method, device and system based on 3D structured light. Background Art

[0002] In the digital age, the demand for accurate human body measurements is growing, especially in the fields of medicine, clothing customization, sports science, entertainment, etc. Existing human body measurement technologies, including traditional tape measurement or laser scanning-based methods, often cannot take into account both accuracy and efficiency of measurement. For example, although laser scanning has high accuracy, it is expensive and slow, and is not suitable for rapidly changing market demands.

[0003] At present, there are still some technical solutions for deploying camera arrays to achieve human body measurement. Such technical solutions usually require that several cameras be deployed three-dimensionally from multiple directions and angles in an independent and enclosed space, so as to collect rich data of the measured human body. In such solutions, it costs a lot to specially create a venue and build an independent and enclosed space. In addition, if a three-dimensional measurement of the human body is performed, the person being measured must first be placed independently in the space. Taking children as an example, such a measurement environment and measurement process may make people feel insecure or produce a sense of rejection, which will affect the user's measurement experience and is not conducive to the smooth progress of three-dimensional measurement. In addition, for independent and enclosed spaces, for safety reasons, special security monitoring products, such as security receivers, may also need to be deployed. It is not difficult to see that this will incur additional expenses.

[0004] How to achieve three-dimensional measurement of the human body in a relatively simple, convenient and low-cost way is a technical problem that needs to be solved urgently in the current field. Summary of the invention

[0005] Based on the above problems, the present application provides a single-sided deployed three-dimensional human body measurement method, device and system based on 3D structured light, aiming to improve the user's human body measurement experience and reduce the cost of three-dimensional human body measurement in a relatively simple and convenient deployment manner by using 3D structured light and a parametric human body model.

[0006] The embodiments of the present application disclose the following technical solutions:

[0007] In a first aspect, the present application provides a single-sided deployment three-dimensional human body measurement method based on 3D structured light, the method comprising:

[0008] By deploying a 3D structured light camera module on the first facade, light with structural characteristics is emitted to the front and back of the human body to be tested in turn, and the light reflected from the front and back of the human body to be tested is captured to generate front depth map data and back depth map data of the human body to be tested;

[0009] Converting the front depth map data into a first portion of point cloud data in a three-dimensional space, and converting the back depth map data into a second portion of point cloud data in the three-dimensional space;

[0010] Based on the first part of point cloud data and the second part of point cloud data, spatial registration is performed to obtain a point cloud data model corresponding to the human body to be measured;

[0011] Matching the point cloud data model with a parameterized human body model, and adjusting the parameters of the parameterized human body model by an optimization algorithm to minimize the distance between key points in the parameterized human body model and corresponding key points in the point cloud data model;

[0012] The three-dimensional measurement data of the human body to be measured is determined according to the parameterized human body model obtained after the fitting is completed.

[0013] In an optional implementation of the first aspect, N 3D structured light camera modules are specifically deployed on the first facade, each of the 3D structured light camera modules includes a transmitter and a receiver, and N is an integer greater than or equal to 2;

[0014] The capturing of the light reflected from the front side and the light reflected from the back side of the human body to be tested, and generating the front side depth map data and the back side depth map data of the human body to be tested, comprises:

[0015] The N receivers respectively capture the light reflected from the front side of the human body to be tested, and generate N groups of front depth map data of the human body to be tested; and the N receivers respectively capture the light reflected from the back side of the human body to be tested, and generate N groups of back depth map data of the human body to be tested;

[0016] The converting the front depth map data into a first portion of point cloud data in a three-dimensional space, and converting the back depth map data into a second portion of point cloud data in the three-dimensional space, comprises:

[0017] Using the N groups of front depth map data and the internal parameters corresponding to the N receivers, the N groups of front depth map data are converted one-to-one into N groups of first initial point cloud data; using the N groups of back depth map data and the internal parameters corresponding to the N receivers, the N groups of back depth map data are converted one-to-one into N groups of second initial point cloud data;

[0018] According to the external parameters corresponding to the N receivers, the N groups of first initial point cloud data are converted into the same three-dimensional space to obtain the first part of point cloud data; according to the external parameters corresponding to the N receivers, the N groups of second initial point cloud data are converted into the same three-dimensional space to obtain the second part of point cloud data.

[0019] In an optional implementation of the first aspect, converting the N groups of first initial point cloud data into the same three-dimensional space to obtain the first part of point cloud data includes:

[0020] After converting the N groups of first initial point cloud data into the same three-dimensional space, performing point cloud registration using an iterative closest point algorithm and a multi-view stereo matching technique, and obtaining the first part of point cloud data after the registration is completed;

[0021] The converting the N groups of second initial point cloud data into the same three-dimensional space to obtain the second part of point cloud data includes:

[0022] After converting the N groups of second initial point cloud data into the same three-dimensional space, point cloud registration is performed using an iterative closest point algorithm and a multi-view stereo matching technique, and the second part of point cloud data is obtained after the registration is completed.

[0023] In an optional implementation of the first aspect, performing spatial registration based on the first part of point cloud data and the second part of point cloud data to obtain a point cloud data model corresponding to the human body to be measured includes:

[0024] Performing spatial registration on the first portion of point cloud data and the second portion of point cloud data;

[0025] Based on the coverage of human key points, the integrity of the registered point cloud dataset is checked;

[0026] If the integrity check of the point cloud data set passes, data quality preprocessing is performed on the point cloud data set, and after the preprocessing is completed, a point cloud data model corresponding to the human body to be tested is obtained; the data quality preprocessing includes at least one of the following:

[0027] Filtering, background segmentation or outlier removal.

[0028] In an optional implementation of the first aspect, the 3D structured light camera module deployed on the first facade successively emits light with structural features to the front and back of the human body to be tested, and captures the light reflected from the front and back of the human body to be tested, to generate front depth map data and back depth map data of the human body to be tested, including:

[0029] Outputting a first measurement guidance prompt to guide the human body to be measured to move into the field of view of the 3D structured light camera module and face the first facade;

[0030] The 3D structured light camera module emits light with structural features to the front of the human body to be tested, and captures the light reflected from the front of the human body to be tested, so as to generate front depth map data of the human body to be tested;

[0031] Outputting a second measurement guidance prompt to guide the measured person to rotate the body with the back facing the first facade and still remain within the field of view of the 3D structured light camera module;

[0032] The 3D structured light camera module emits light with structural features to the back side of the human body to be measured, and captures the light reflected from the back side of the human body to be measured, so as to generate back side depth map data of the human body to be measured.

[0033] In an optional implementation of the first aspect, while emitting light with structural features to the front of the human body to be measured through the 3D structured light camera module and capturing light reflected from the front of the human body to be measured, the method further includes:

[0034] Feedback the front scanning progress of the human body to be tested by the 3D structured light camera module in real time through the human-computer interaction interface;

[0035] While emitting light with structural features to the back of the human body to be measured through the 3D structured light camera module and capturing the light reflected from the back of the human body to be measured, the method further includes:

[0036] Feedback the scanning progress of the back of the human body to be tested by the 3D structured light camera module in real time through the human-computer interaction interface;

[0037] The first measurement guidance prompt and the second measurement guidance prompt are both output through a human-computer interaction interface; wherein the second measurement guidance prompt is output by the human-computer interaction interface after the front scanning progress reaches 100%.

[0038] In an optional implementation of the first aspect, matching the point cloud data model with a parameterized human body model, and adjusting parameters of the parameterized human body model by an optimization algorithm to minimize distances between key points in the parameterized human body model and corresponding key points in the point cloud data model, includes:

[0039] Mapping the point cloud data model to the surface of the parameterized human body model to perform corresponding matching of key points;

[0040] Adjusting the parameters of the parameterized human body model by a nonlinear least squares optimization algorithm to minimize the distance between the key points in the parameterized human body model and the corresponding key points in the point cloud data model to complete the global fitting;

[0041] After the global fitting is completed, the parameterized human body model is processed in at least one of the following ways:

[0042] Local refinement of geometric features, simulation of gravity and contact forces based on soft tissue models, or enhancement of geometric features.

[0043] In an optional implementation of the first aspect, the method further includes:

[0044] Based on the point cloud data model, a fitting effect analysis is performed on the parameterized human body model obtained after fitting, and the fitting effect is displayed or a verification result of the fitting accuracy is output.

[0045] The second aspect of the present application provides a single-sided deployed three-dimensional human body measurement device based on 3D structured light, the device comprising:

[0046] A depth map generation module is used to emit light with structural characteristics to the front and back of the human body to be tested through a 3D structured light camera module deployed on the first facade, and capture the light reflected from the front and back of the human body to be tested, so as to generate front depth map data and back depth map data of the human body to be tested;

[0047] A point cloud conversion module, configured to convert the front depth map data into a first portion of point cloud data in a three-dimensional space, and to convert the back depth map data into a second portion of point cloud data in the three-dimensional space;

[0048] A registration module, configured to perform spatial registration based on the first part of point cloud data and the second part of point cloud data to obtain a point cloud data model corresponding to the human body to be measured;

[0049] A fitting module, used for matching the point cloud data model with a parameterized human body model, and adjusting the parameters of the parameterized human body model through an optimization algorithm to minimize the distance between key points in the parameterized human body model and corresponding key points in the point cloud data model;

[0050] The measurement data determination module is used to determine the three-dimensional measurement data of the human body to be measured according to the parameterized human body model obtained after the fitting is completed.

[0051] The third aspect of the present application provides a single-sided deployment three-dimensional human body measurement system based on 3D structured light, the system comprising: a processor, a memory, a 3D structured light camera module and a display; the memory, the 3D structured light camera module and the display are all communicatively connected to the processor;

[0052] The 3D structured light camera module is deployed on the first facade;

[0053] The memory stores a computer program;

[0054] The processor is used to run the computer program, and when the computer program is run, the single-side deployment three-dimensional human body measurement method based on 3D structured light as described in any implementation of the first aspect is executed;

[0055] The display is used to display a human-computer interaction interface, and outputs measurement guidance prompts for the front and back sides through the human-computer interaction interface.

[0056] In an optional implementation of the third aspect, N 3D structured light camera modules are specifically deployed on the first facade, each of the 3D structured light camera modules includes a transmitter and a receiver, and N is an integer greater than or equal to 2;

[0057] N receivers are arranged along the vertical direction, and the union of the visual fields of the N receivers covers the entire human body to be measured.

[0058] In an optional implementation manner of the third aspect, the exposure times of the infrared image and the depth image generated by the 3D structured light camera module are synchronized.

[0059] Compared with the prior art, this application has the following beneficial effects:

[0060] The present application proposes a single-sided deployment three-dimensional human body measurement method, device and system based on 3D structured light. The 3D structured light camera module successively emits light with structural features to the front and back of the human body to be measured, and captures the light reflected from the front and back to generate front depth map data and back depth map data. Then, point cloud conversion and spatial registration are performed to obtain a point cloud data model. The point cloud data model is matched with a parametric human body model, and the parameters of the parametric human body model are adjusted by an optimization algorithm to minimize the distance between the key points in the parametric human body model and the corresponding key points in the point cloud data model. Finally, the three-dimensional measurement data of the human body to be measured is determined according to the parametric human body model obtained after the fitting. The 3D structured light camera module emits light to the front and back of the human body to be measured and captures the reflected light, and successively performs operations such as point cloud conversion and spatial registration, thereby ensuring the accuracy and precision of the point cloud data model in restoring the three-dimensional structure of the human body. Based on the point cloud data model, key points are matched and parameters are adjusted with the parametric human body model, which improves the fitting accuracy of the parametric human body model. It can be seen that the technical solution of the present application can achieve three-dimensional human body measurement more accurately. In addition, the 3D structured light camera module can be deployed on a facade, which is simple and convenient. There is no need to create a special venue to create an independent enclosed space. The user measurement experience is improved and the cost of three-dimensional human body measurement is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0062] Figure 1 A schematic diagram of a scenario for implementing three-dimensional human body measurement in an independent and enclosed space;

[0063] Figure 2 A schematic diagram of a single-sided deployment three-dimensional human body measurement implementation scenario;

[0064] Figure 3 A schematic diagram of another implementation scenario of single-sided deployment 3D human body measurement;

[0065] Figure 4 A schematic diagram of another implementation scenario of single-sided deployment three-dimensional human body measurement;

[0066] Figure 5 A flowchart of a single-sided deployment three-dimensional human body measurement method based on 3D structured light provided in an embodiment of the present application;

[0067] Figure 6A schematic diagram of a front-facing camera reminder for a human-computer interaction interface;

[0068] Figure 7 A schematic diagram of a back-side shooting reminder for a human-computer interaction interface;

[0069] Figure 8 A schematic diagram of a front progress reminder of a human-computer interaction interface;

[0070] Fig. 9 A schematic diagram of a progress reminder on the back of a human-computer interaction interface;

[0071] Fig.10 A schematic diagram of the structure of a single-sided deployed three-dimensional human body measurement device based on 3D structured light provided in an embodiment of the present application;

[0072] Fig.11 A schematic structural diagram of a single-sided deployed three-dimensional human body measurement system based on 3D structured light provided in an embodiment of the present application. DETAILED DESCRIPTION

[0073] At present, in the fields of medicine, clothing design and sports analysis, it is often necessary to determine high-precision human body measurement data to carry out specific design or analysis. These measurement data may include but are not limited to height, head circumference, upper and lower limb ratio, etc. Although some methods can achieve three-dimensional human body measurement, such as laser scanning, the implementation cost is high and the speed is slow. Although the method of deploying camera arrays can also achieve three-dimensional human body measurement, it requires an independent enclosed space for users to be placed in it, which results in a poor user measurement experience and is not conducive to the advancement of measurement work. In addition, the cost is still relatively high. For ease of understanding, please refer to Figure 1 . Figure 1 This is a schematic diagram of the implementation scenario of three-dimensional human body measurement in an independent and enclosed space. Figure 1 In the scene shown, several cameras are configured in a three-dimensional manner to collect data around the human body from multiple perspectives and realize three-dimensional body measurement.

[0074] In view of the above problems, the inventors proposed, after research, to realize the collection of depth map data of the front and back of the human body to be measured by deploying a 3D structured light camera module on a single side. The depth map data of the front and back are converted into point cloud data and then spatially aligned to form a point cloud data model corresponding to the human body to be measured. The parameterized human body model and the point cloud data model are matched based on key points, and the parameters of the parameterized human body model are adjusted to optimize the fitting effect of the parameterized human body model on the human body to be measured. Finally, the three-dimensional measurement data of the human body to be measured is determined based on the parameterized human body model obtained after the fitting is completed. The present application uses a relatively simple and convenient camera deployment method, which not only saves the cost of three-dimensional human body measurement, but also ensures the measurement precision and accuracy. At the same time, it can also improve the user's measurement experience.

[0075] Figures 2 to 4 Schematic diagrams of several implementation scenarios of single-sided 3D human body measurement are shown. In these implementation scenarios, a 3D structured light camera module required for 3D human body measurement is deployed on a facade. Figure 2 and Figure 3 The 3D structured light camera module on display includes 1 3D structured light camera module, Figure 2 and Figure 3 The different colors of the human body in the picture indicate that the shooting position has changed due to rotation. Figure 2 It means that the human body is facing the receiver. Figure 3 It means that the back of the human body is facing the receiver. Each 3D structured light camera module can include a transmitter and a receiver. The transmitter is mainly used to emit infrared light, and the receiver is used to form an infrared image based on the infrared light reflected by the human body, which is used to generate a depth map later. The exposure time of the infrared image and the depth map are synchronized. Figure 4 Two 3D structured light camera modules are shown in the figure, and the two receivers can be arranged in the vertical direction. Figure 5 The flowchart of a single-sided deployment three-dimensional human body measurement method based on 3D structured light provided in an embodiment of the present application. Figure 5 As shown, the method includes:

[0076] S501. By deploying a 3D structured light camera module on the first facade, light with structural characteristics is emitted to the front and back of the human body to be tested, and the light reflected from the front and the light reflected from the back of the human body to be tested are captured to generate front depth map data and back depth map data of the human body to be tested.

[0077] The 3D structured light camera module includes a transmitter and a receiver. The transmitter is used to emit light with structural features, and the receiver reflects the light to obtain an infrared image. The module can also be configured with a chip that can obtain a corresponding depth map based on the infrared image. As an example, the structural features here can be grids or stripes. In actual applications, the number of 3D structured light camera modules can be selected in combination with the field of view of a single receiver. For example, if the field of view of a single receiver is sufficient to cover the front of the person to be measured or the back of the person to be measured at a single time, one 3D structured light camera module can be selected; conversely, if the field of view of a single receiver cannot cover the entire front of the person to be measured or the back of the person to be measured, or, in order to reduce the impact of image distortion on the measurement, multiple 3D structured light camera modules can be selected. For example Figure 4 Generally, two 3D structured light camera modules are configured in a 3D structured light camera module and arranged along the vertical direction.

[0078] In general, the selected 3D structured light camera module is required to meet the standards of high resolution and lightweight for easy deployment and operation. The 3D structured light camera module needs to be able to support operation under a wide field of view to ensure that the depth map data of the front of the human body and the depth map data of the back of the human body are captured at one time. The receiver inside the module can be specially designed to increase the sensitivity to reflected light and be able to resist the interference of ambient light. As a result, the 3D structured light camera module can perform relatively stably under a variety of lighting conditions and can achieve clear acquisition of depth map data. The following is an introduction to several features of the 3D structured light camera module, namely adaptive algorithm, anti-interference ability, dynamic range optimization, and real-time feedback and adjustment.

[0079] Adaptive algorithm: The 3D structured light camera module has a built-in adaptive algorithm, which enables it to maintain stable performance under different lighting conditions. The algorithm can detect changes in ambient lighting and dynamically adjust the emission intensity of the emitter to ensure that the reflected light on the human body surface can be clearly captured even in strong or weak light environments.

[0080] Anti-interference ability: 3D structured light camera module has anti-interference ability, which can filter out the interference of ambient light and improve the signal-to-noise ratio of the image. This is achieved by using specific infrared bands and narrow-band filters to ensure that only the reflected light from the structured light is captured, while the interference of other light sources is effectively suppressed.

[0081] Dynamic range optimization: The image sensor of the 3D structured light camera module has a wide dynamic range, which can capture rich details under extreme lighting conditions. Combined with the adaptive algorithm, the sensor can automatically adjust the exposure and gain to ensure clear depth images in various environments.

[0082] Real-time feedback and adjustment: The system provides a real-time feedback mechanism. When image quality degradation is detected, the adjustment process is automatically triggered, including adjusting the emitter intensity, sensor exposure parameters, etc., to ensure that the output depth map quality is always maintained at the best state.

[0083] In this step, the first facade is a physical facade or a virtual facade where the 3D structured light camera module is arranged. Figures 2 to 4 The scene shown in the figure is a 3D structured light camera module deployed on a physical facade. In practical applications, the 3D structured light camera module can also be deployed on a height-adjustable bracket. In this case, the bracket can also be regarded as being on a virtual facade, and correspondingly, the 3D structured light camera module is also deployed on a virtual facade. The bracket allows the adjustment of the shooting angle and shooting distance to adapt to the test subjects of different heights and body shapes, and has strong flexibility.

[0084] For the case of N receivers, in order to ensure the accurate recognition performance of the subsequent 3D structured light camera module and the accuracy of the three-dimensional human body measurement results, it is necessary to pre-calibrate the parameters of the N receivers using a calibration plate. Among them, the receiver includes a sensor and a lens, and the calibration plate is used to calibrate the parameters of the receiver, specifically the lens. Each receiver is required to shoot the calibration plate separately, and record the image of the calibration plate at different viewing angles corresponding to each receiver, so as to calculate the internal parameters of the camera (referred to as internal parameters, such as focal length, principal point coordinates) and external parameters (referred to as external parameters, such as the position and posture of the camera relative to the calibration plate). This process can be specifically achieved with the help of computer vision algorithms to achieve geometric solution, and the specific implementation process will not be repeated for the time being.

[0085] For the case of N receivers, the exposure time of the infrared image and the depth map are synchronized. For a 3D structured light camera module, the synchronization mechanism between the transmitter and the receiver is the key. The precise synchronization of the emitted light and the exposure of the receiver can reduce image distortion and calculation and measurement errors. If a 3D structured light camera module including N receivers is used, after the N transmitters emit infrared light with structural features to the front of the human body to be tested, the N receivers respectively capture the light reflected from the front of the human body to be tested, and generate N sets of front depth map data of the human body to be tested. After the transmitter emits infrared light with structural features to the back of the human body to be tested, the N receivers respectively capture the light reflected from the back of the human body to be tested, and generate N sets of back depth map data of the human body to be tested. It should be noted that the shooting order of the front and back can be customized, for example, shooting the front first and then the back, or shooting the back first and then the front.

[0086] Before shooting, guidance prompts can be output to the user by devices such as a display or a speaker on site. For example, a first measurement guidance prompt is first output to guide the person to be measured to move into the field of view of the 3D structured light camera module, and face the first facade; then, the 3D structured light camera module emits light with structural features to the front of the person to be measured, and captures the light reflected from the front of the person to be measured to generate the front depth map data of the person to be measured. Then, a second measurement guidance prompt is output to guide the person to be measured to rotate the body with the back facing the first facade, and still remain within the field of view of the 3D structured light camera module; the 3D structured light camera module emits light with structural features to the back of the person to be measured, and captures the light reflected from the back of the person to be measured to generate the back depth map data of the person to be measured.

[0087] In case there is a display on site, the human-computer interaction interface can be displayed on the display. Figure 6 and Figure 7The following are reminder diagrams for front and back shooting of the human-computer interaction interface. Before actually starting the front shooting, you can show Figure 6 The human-computer interaction interface shown is used to remind the user to face the receiver in the required direction and / or posture. Figure 6 The words in the figure are used to exemplarily show the first measurement guide mentioned above. Similarly, before actually starting to shoot the back side, the Figure 7 The human-computer interaction interface shown is used to remind the user to face the receiver in the required direction and / or posture. Figure 7 The words in the figure exemplarily show the second measurement guidance prompt mentioned above. That is to say, both the first measurement guidance prompt and the second measurement guidance prompt can be output through the human-computer interaction interface.

[0088] While emitting light with structural features to the front of the human body to be tested through the 3D structured light camera module and capturing the light reflected from the front of the human body to be tested, the method may also include: providing real-time feedback on the scanning progress of the front of the human body to be tested by the 3D structured light camera module through a human-computer interaction interface. Figure 8 As shown in FIG. 1 , it is a schematic diagram of a front progress reminder of a human-machine interaction interface. While the 3D structured light camera module emits light with structural features to the back of the human body to be tested and captures the light reflected from the back of the human body to be tested, the method may also include: providing real-time feedback of the scanning progress of the back of the human body to be tested by the 3D structured light camera module through the human-machine interaction interface. Fig. 9 As shown, it is a schematic diagram of the back progress reminder of the human-computer interaction interface. When the front scanning progress and the back scanning progress have not reached 100% respectively, the user should follow the prompts in the human-computer interaction interface and maintain the required direction and posture to ensure the accuracy of the measurement results.

[0089] The second measurement guidance prompt is output by the human-computer interaction interface after the front scanning progress reaches 100%, which can be understood as: Figure 8 After the progress reaches 100% in the interface shown, further display Figure 7 The display of scanning progress in the human-computer interaction interface makes it easier for users to understand and grasp the real-time progress of the measurement, so that users can anticipate the remaining processes they need to cooperate with in 3D human body measurement. This also improves the user's measurement experience.

[0090] S502: Convert the front depth map data into a first portion of point cloud data in a three-dimensional space, and convert the back depth map data into a second portion of point cloud data in the three-dimensional space.

[0091] The following further introduces an example implementation of this step using the example of N receivers described above. In one example, S502 includes:

[0092] First, using the N groups of front depth map data and the internal parameters corresponding to the N receivers, the N groups of front depth map data are converted one-to-one into N groups of first initial point cloud data; using the N groups of back depth map data and the internal parameters corresponding to the N receivers, the N groups of back depth map data are converted one-to-one into N groups of second initial point cloud data.

[0093] Next, according to the external parameters corresponding to the N receivers, the N groups of first initial point cloud data are converted into the same three-dimensional space to obtain the first part of point cloud data; according to the external parameters corresponding to the N receivers, the N groups of second initial point cloud data are converted into the same three-dimensional space to obtain the second part of point cloud data.

[0094] The above conversion process from depth map to point cloud can be completed by other processing chips outside the camera module.

[0095] Through this process, the generation of point cloud is completed. The point cloud generation process involves converting the depth map into point cloud data in three-dimensional space. Each point cloud represents a point on the surface of the human body, and each point cloud contains its X, Y, and Z coordinates in three-dimensional space. In this step, the generated point cloud data can also be pre-processed, including filtering and denoising, background segmentation, and outlier removal to improve the quality of the point cloud data. This process ensures the accuracy and reliability of subsequent point cloud matching, splicing, and analysis.

[0096] Among them, N groups of first initial point cloud data are converted into the same three-dimensional space to obtain the first part of point cloud data, including: after converting the N groups of first initial point cloud data into the same three-dimensional space, using iterative closest point (ICP) algorithm and multi-view stereo matching technology to perform point cloud registration, and obtaining the first part of point cloud data after the registration is completed.

[0097] Among them, converting N groups of second initial point cloud data into the same three-dimensional space to obtain the second part of point cloud data, including: after converting the N groups of second initial point cloud data into the same three-dimensional space, using ICP algorithm and multi-view stereo matching technology to perform point cloud registration, and obtaining the second part of point cloud data after the registration is completed.

[0098] ICP algorithm: Iteratively optimizes the relative position and pose between two point clouds by minimizing the distance between them.

[0099] Multi-view stereo matching: In some cases, it may be necessary to use multi-view stereo matching technology in this application to optimize the fusion of point clouds, especially when the viewpoint difference is large or the feature points are not well matched.

[0100] Specifically, the correspondence between point clouds from different perspectives can be determined through feature point matching technology, and then the point clouds can be registered using the ICP algorithm to find the optimal rotation and translation transformations so that point cloud data from different perspectives can be accurately aligned.

[0101] S503: Perform spatial registration based on the first part of point cloud data and the second part of point cloud data to obtain a point cloud data model corresponding to the human body to be measured.

[0102] In S502, the registration of point clouds for different viewing angles is introduced. It is understandable that the first part of the point cloud data comes from the frontal shooting of the human body to be measured, and the second part of the point cloud data comes from the back shooting of the human body to be measured. In order to realize the construction of a complete human body point cloud data model, it is also necessary to comprehensively use the first part of the point cloud data and the second part of the point cloud data. This process can still use the ICP algorithm and multi-view stereo matching technology to ensure that the first part of the point cloud data and the second part of the point cloud data after registration can be fused. In this process, special attention is paid to processing the overlapping areas of the point clouds to ensure that the generated model has no gaps or overlaps, thereby improving the accuracy and surface quality of the model.

[0103] Next, an example implementation of S504 is provided. Specifically, based on the first part of point cloud data and the second part of point cloud data, spatial registration is performed to obtain a point cloud data model corresponding to the human body to be measured, including:

[0104] The first part of point cloud data and the second part of point cloud data are spatially registered; based on the coverage of key points of the human body, the integrity of the registered point cloud data set is checked; if the integrity check of the point cloud data set passes, the point cloud data set is preprocessed for data quality, and after the preprocessing is completed, a point cloud data model corresponding to the human body to be tested is obtained.

[0105] The key points of the human body correspond to the main joints and characteristic parts of the human body, such as the acromion, the center of the hip joint, the center of the knee joint, etc., and are the key references in model fitting. Therefore, after spatially registering the first part of the point cloud data with the second part of the point cloud data, it is necessary to analyze the coverage rate of the key points of the human body in the currently registered point cloud data set. In this way, the problem of poor subsequent model fitting accuracy due to insufficient data representing the key points of the human body in the point cloud data, which in turn leads to inaccurate three-dimensional human body measurement, can be avoided.

[0106] In practical applications, the threshold of integrity check can be pre-set, for example, the threshold is set to 80%. If the coverage rate of all preset human key points in the registered point cloud data set reaches more than 80%, it is determined that the integrity check has passed; otherwise, if the coverage rate cannot reach more than 80%, it is determined that the integrity check has not passed. It should be noted that the above thresholds can be customized according to the requirements of measurement accuracy, and can be set through the human-computer interaction interface, so as to enhance the user's actual measurement experience in this flexible setting method.

[0107] If the integrity check fails, the user can be further prompted to change posture for one or more further shooting and collection, thereby improving the coverage of the point cloud data set for key points of the human body and improving the three-dimensional measurement precision and accuracy of the human body to be measured.

[0108] On the premise that the integrity check of the point cloud data set passes, the point cloud data set is preprocessed for data quality. The specific preprocessing methods may include but are not limited to: filtering, background segmentation or outlier removal. In this way, the accuracy and effectiveness of the point cloud data in the point cloud data set can be improved, unnecessary interference can be reduced, and the results of 3D human body measurement or the accuracy of model fitting can be avoided to be affected by outliers and other reasons.

[0109] It should be noted that the registration results can also be verified in this application, for example, by visual inspection or calculation of the registration error to verify the accuracy of the registration results. If the registration results are not satisfactory, you can return to the feature extraction or matching steps for adjustment. In addition, this application can also perform system correction. For example, based on the verification results of the registration results, the system is fine-tuned, including adjusting the position of the 3D structured light camera module, recalibrating or optimizing the algorithm parameters.

[0110] S504: Match the point cloud data model with a parameterized human body model, and adjust the parameters of the parameterized human body model through an optimization algorithm to minimize the distance between key points in the parameterized human body model and corresponding key points in the point cloud data model.

[0111] In an example implementation of this step, specifically: mapping the point cloud data model to the surface of the parameterized human body model to perform corresponding matching of key points; adjusting the parameters of the parameterized human body model through a nonlinear least squares optimization algorithm to minimize the distance between the key points in the parameterized human body model and the corresponding key points in the point cloud data model to complete global fitting; processing the parameterized human body model after the global fitting in at least one of the following ways: local refinement of geometric features, simulation of gravity and contact forces based on a soft tissue model, or enhancement of geometric features.

[0112] By minimizing the distance between the key points in the parametric human body model and the corresponding key points in the point cloud data model, as the optimization target of the parametric human body model in the model fitting stage, the model fitting effect can be significantly improved, and the model can be close to the spatial characteristics of the key points in the point cloud data model. In addition, the parametric human body model can accurately capture and simulate the characteristics of the key points of the human body to be tested.

[0113] The nonlinear least squares optimization algorithm is only used as an example optimization algorithm used when adjusting the parameters of the parameterized human body model in the embodiments of the present application. In practical applications, other possible optimization algorithms can also be selected in combination with specific optimization requirements. Here, the optimization algorithm is not limited.

[0114] After the global fitting, the parametric human body model can be optimized in terms of its ability to express human features by local refinement, simulation of gravity and contact forces based on the soft tissue model, or enhancement of geometric features, thereby improving the effect of three-dimensional human body measurement.

[0115] Specifically, local areas of the model are refined and local geometric features, such as muscle protrusions or joint depressions, are adjusted to better match the point cloud data. Soft tissue models are introduced to simulate the effects of gravity and contact forces on the shape of the human body to further improve fitting accuracy. Geometric features of the model, such as wrinkles, skin folds and other details, are enhanced to improve the realism of the model. Spatial transformations and constraints are used to correct deviations caused by local adjustments to ensure overall spatial consistency.

[0116] After the specific execution of the above steps, the human body features are captured and mapped to the parameterized human body model, and the features of the human body to be measured are expressed through the model. The features are not limited to the features of key points, but also include detailed features such as human skin wrinkles, clothing wrinkles, muscle bulges, wrinkles, etc. Based on this, the final determination of the three-dimensional measurement data can be completed through the following step S505.

[0117] S505: Determine the three-dimensional measurement data of the human body to be measured according to the parameterized human body model obtained after fitting.

[0118] The parametric human body model represents the shape and posture of the human body through human body parameters. Therefore, by reading the parameters of the parametric human body model, the specific measurement data representing the three-dimensional characteristics of the human body can be determined by calculation.

[0119] In the method embodiment introduced above, the 3D structured light camera module emits light and captures reflected light on the front and back of the human body to be measured, and successively performs operations such as point cloud conversion and spatial registration, thereby ensuring the accuracy and precision of the point cloud data model in restoring the three-dimensional structure of the human body. Based on the point cloud data model, the key points are matched and the parameters are adjusted with a parametric human body model, which improves the fitting accuracy of the parametric human body model. It can be seen that the technical solution of the present application can achieve three-dimensional human body measurement more accurately. In addition, the 3D structured light camera module can be deployed on a facade, which is simple and convenient. There is no need to specially create a venue to create an independent and enclosed space. The user measurement experience is improved and the cost of three-dimensional human body measurement is reduced.

[0120] As mentioned above, for the case of N receivers, in order to ensure the accurate recognition performance of the subsequent 3D structured light camera module and the accuracy of the 3D human body measurement results, it is necessary to calibrate the parameters of the N receivers using a calibration plate in advance. The following is an introduction to the process of using the calibration plate to calculate the external parameters for registration. Through this process, it is also ensured that the point cloud data finally converted from the images taken by different receivers can be correctly aligned, thus forming a unified point cloud data model.

[0121] In this application, a calibration plate with obvious features is used, such as a calibration plate containing a chessboard, a dot array or other geometric patterns, so as to facilitate accurate identification of feature points from the perspective of different receivers. Each receiver of the 3D structured light camera module independently shoots the calibration plate, and extracts the feature points on the calibration plate through an image processing algorithm. The positions of these feature points in the image need to be accurately detected for subsequent matching and calculation. For the binocular camera module, a matching algorithm can be used to achieve feature point matching between the images of the calibration plate taken by two receivers. A matching algorithm can be used to find corresponding feature point pairs between two sets of images. In addition, robust algorithms such as Random Sample Consensus (RANSAC) can be used to eliminate erroneous matching points to ensure that only correct matching points are used for subsequent calculations. For the binocular camera module, it is necessary to calculate the external parameters of each receiver relative to the calibration plate. Using the matched feature point pairs and the camera internal parameters, the external parameters of each receiver are solved by the algorithm. The registration results are usually expressed in the form of rotation matrices and translation vectors, which are then used for the fusion of point cloud data. It should be noted that, for the calibration process, the transmitter may emit infrared light and the receiver may collect the infrared light for calibration.

[0122] In practical applications, the single-sided deployment 3D human body measurement method based on 3D structured light can further include:

[0123] Based on the point cloud data model, the parametric human body model obtained after fitting is analyzed for fitting effect, and the fitting effect is displayed or the verification result of the fitting accuracy is output. Specifically, the display of the fitting effect or the output of the verification result of the fitting accuracy can be realized through a human-computer interaction interface.

[0124] Here are some examples of how to implement the fitting effect analysis:

[0125] 1. Key point error analysis: By comparing the key point positions predicted by the model with the key point positions in the actual point cloud data, the error value is calculated to evaluate the accuracy of the model.

[0126] 2. Statistical verification: Use statistical methods such as root mean square error (RMSE) or mean absolute error (MAE) to quantify the fitting accuracy and evaluate the overall fitting effect.

[0127] 3. Iterative optimization: If the fitting accuracy does not reach the preset threshold, adjust the model parameters and repeat the fitting process until the expected accuracy is achieved.

[0128] Finally, in the human-computer interaction interface, 3D visualization tools can be used to intuitively display the fitting effect between the model and the point cloud data, conduct visual evaluation, achieve qualitative analysis, and quickly identify potential problems.

[0129] Based on the above introduction to the single-sided deployment three-dimensional human body measurement method based on 3D structured light in this application, accordingly, this application also provides a single-sided deployment three-dimensional human body measurement device based on 3D structured light. Fig.10 , is a schematic diagram of the structure of a single-sided deployment three-dimensional human body measurement device based on 3D structured light. Fig.10 ,The device includes a depth map generation module, a point cloud conversion module, a registration module, a fitting module and a measurement data determination module.

[0130] A depth map generation module is used to emit light with structural characteristics to the front and back of the human body to be tested through a 3D structured light camera module deployed on the first facade, and capture the light reflected from the front and back of the human body to be tested, so as to generate front depth map data and back depth map data of the human body to be tested;

[0131] A point cloud conversion module, configured to convert the front depth map data into a first portion of point cloud data in a three-dimensional space, and to convert the back depth map data into a second portion of point cloud data in the three-dimensional space;

[0132] A registration module, configured to perform spatial registration based on the first part of point cloud data and the second part of point cloud data to obtain a point cloud data model corresponding to the human body to be measured;

[0133] A fitting module, used for matching the point cloud data model with a parameterized human body model, and adjusting the parameters of the parameterized human body model through an optimization algorithm to minimize the distance between key points in the parameterized human body model and corresponding key points in the point cloud data model;

[0134] The measurement data determination module is used to determine the three-dimensional measurement data of the human body to be measured according to the parameterized human body model obtained after the fitting is completed.

[0135] In an optional implementation, N 3D structured light camera modules are specifically deployed on the first facade, each 3D structured light camera module includes a transmitter and a receiver, and N is an integer greater than or equal to 2;

[0136] The depth map generation module is specifically used for:

[0137] The N receivers respectively capture the light reflected from the front side of the human body to be tested, and generate N groups of front depth map data of the human body to be tested; and the N receivers respectively capture the light reflected from the back side of the human body to be tested, and generate N groups of back depth map data of the human body to be tested;

[0138] The point cloud conversion module is specifically used for:

[0139] Using the N groups of front depth map data and the internal parameters corresponding to the N receivers, the N groups of front depth map data are converted one-to-one into N groups of first initial point cloud data; using the N groups of back depth map data and the internal parameters corresponding to the N receivers, the N groups of back depth map data are converted one-to-one into N groups of second initial point cloud data;

[0140] According to the external parameters corresponding to the N receivers, the N groups of first initial point cloud data are converted into the same three-dimensional space to obtain the first part of point cloud data; according to the external parameters corresponding to the N receivers, the N groups of second initial point cloud data are converted into the same three-dimensional space to obtain the second part of point cloud data.

[0141] In an optional implementation, the point cloud conversion module is specifically used to:

[0142] After converting the N groups of first initial point cloud data into the same three-dimensional space, performing point cloud registration using an iterative closest point algorithm and a multi-view stereo matching technique, and obtaining the first part of point cloud data after the registration is completed;

[0143] After converting the N groups of second initial point cloud data into the same three-dimensional space, point cloud registration is performed using an iterative closest point algorithm and a multi-view stereo matching technique, and the second part of point cloud data is obtained after the registration is completed.

[0144] In an optional implementation, the registration module is specifically used to:

[0145] Performing spatial registration on the first portion of point cloud data and the second portion of point cloud data;

[0146] Based on the coverage of human key points, the integrity of the registered point cloud dataset is checked;

[0147] If the integrity check of the point cloud data set passes, data quality preprocessing is performed on the point cloud data set, and after the preprocessing is completed, a point cloud data model corresponding to the human body to be tested is obtained; the data quality preprocessing includes at least one of the following:

[0148] Filtering, background segmentation or outlier removal.

[0149] In an optional implementation, the depth map generation module includes: a prompt output unit and a map data generation unit;

[0150] A prompt output unit, configured to output a first measurement guidance prompt to guide the human body to be measured to move into the field of view of the 3D structured light camera module and face the first facade;

[0151] A map data generating unit, configured to emit light with structural features to the front of the human body to be tested through the 3D structured light camera module, and capture the light reflected from the front of the human body to be tested, so as to generate front depth map data of the human body to be tested;

[0152] The prompt output unit is further used to output a second measurement guidance prompt to guide the measured person to rotate the body with the back facing the first facade and still remain within the field of view of the 3D structured light camera module;

[0153] The image data generating unit is also used to emit light with structural characteristics to the back side of the human body to be tested through the 3D structured light camera module, and capture the light reflected from the back side of the human body to be tested, so as to generate the back side depth map data of the human body to be tested.

[0154] In an optional implementation, the single-sided deployment three-dimensional human body measurement device based on 3D structured light further includes a feedback module; the feedback module is used to provide real-time feedback on the front scanning progress of the 3D structured light camera module on the human body to be measured through a human-computer interaction interface while emitting light with structural features to the front of the human body to be measured through the 3D structured light camera module and capturing light reflected from the front of the human body to be measured; and the feedback module is also used to provide real-time feedback on the back scanning progress of the 3D structured light camera module on the human body to be measured through the human-computer interaction interface while emitting light with structural features to the back of the human body to be measured through the 3D structured light camera module and capturing light reflected from the back of the human body to be measured;

[0155] The first measurement guidance prompt and the second measurement guidance prompt are both output through a human-computer interaction interface; wherein the second measurement guidance prompt is output by the human-computer interaction interface after the front scanning progress reaches 100%.

[0156] In an optional implementation, the fitting module is specifically used to:

[0157] Mapping the point cloud data model to the surface of the parameterized human body model to perform corresponding matching of key points;

[0158] Adjusting the parameters of the parameterized human body model by a nonlinear least squares optimization algorithm to minimize the distance between the key points in the parameterized human body model and the corresponding key points in the point cloud data model to complete the global fitting;

[0159] After the global fitting is completed, the parameterized human body model is processed in at least one of the following ways:

[0160] Local refinement of geometric features, simulation of gravity and contact forces based on soft tissue models, or enhancement of geometric features.

[0161] In an optional implementation, the single-sided deployment three-dimensional human body measurement device based on 3D structured light further includes:

[0162] The fitting verification module is used to analyze the fitting effect of the parameterized human body model obtained after fitting based on the point cloud data model, and to display the fitting effect or output the verification result of the fitting accuracy.

[0163] Based on the single-sided deployment three-dimensional human body measurement method based on 3D structured light and the single-sided deployment three-dimensional human body measurement device based on 3D structured light introduced above, the present application also provides a single-sided deployment three-dimensional human body measurement system based on 3D structured light. Fig.11 Figure 1 is a schematic diagram of the system structure. Fig.11As shown, the system includes: a processor, a memory, a 3D structured light camera module and a display; the memory, the 3D structured light camera module and the display are all communicatively connected to the processor;

[0164] The 3D structured light camera module is deployed on the first facade;

[0165] The memory stores a computer program;

[0166] The processor is used to run the computer program, and when the computer program is run, the single-side deployment three-dimensional human body measurement method based on 3D structured light as described in the method embodiment is executed;

[0167] The display is used to display a human-computer interaction interface, and outputs measurement guidance prompts for the front and back sides through the human-computer interaction interface.

[0168] In an optional implementation, N 3D structured light camera modules are specifically deployed on the first facade, each 3D structured light camera module includes a transmitter and a receiver; the N receivers are arranged along the vertical direction, and the union of the field of view of the N receivers covers the entire human body to be tested.

[0169] In an optional implementation, the exposure time of the infrared image generated by the 3D structured light camera module is synchronized. The exposure time of the N receivers is synchronized.

[0170] In summary, an innovative human body measurement technology is proposed in this application. By capturing the front and back depth maps of the human body on a single-sided 3D structured light camera module, the accurate capture and three-dimensional reconstruction of the human body size and shape are achieved. Combining the key point detection of the human skeleton with the parametric human body model and using efficient automated processing software, not only the equipment configuration is simplified and the cost is reduced, but also the convenience and accuracy of the measurement are improved. According to the accuracy requirements, binocular or multi-eye camera modules are selectively used to complete high-quality point cloud stitching to generate a more accurate three-dimensional model of the human body. This method is particularly suitable for application fields that require high-precision human body measurement, such as medical, clothing design, sports science, entertainment and games. This wide applicability ensures the market potential and commercial value of the technology and provides innovative solutions for different industries. Such as medical, clothing design and motion analysis, it has broad application prospects and commercial value. The following introduces several applicable fields and scenarios of the technical solution of this application.

[0171] Medical field: Customize personalized orthopedic braces and prostheses for patients, and accurately locate tumors on the body surface. In the medical field, it can improve the customization accuracy of orthopedic braces and prostheses, reduce medical costs and improve patient satisfaction.

[0172] Clothing design: It provides accurate body shape data, supports personalized clothing design and customization, and improves the fit and comfort of clothing. In the clothing industry, it can provide more accurate body shape data and promote the development of personalized customized clothing.

[0173] Sports Science: Providing athletes with accurate body measurement data to help optimize training plans and competitive performance.

[0174] Entertainment and Games: Create realistic 3D character models for enhanced visual and interactive experiences in movies, video games, and virtual reality.

[0175] The salient feature of the present application is its simplified camera deployment design. Unlike traditional multi-camera arrays or complex scanning equipment, the present invention only needs to deploy a 3D structured light camera module on a single side to complete the depth map shooting of the front and back of the human body. This design greatly reduces the complexity and cost of system deployment and simplifies the operation process. The efficient automated processing software module can guide users to shoot the front and back depth maps of the human body and fit these depth map data with the parametric human body model. Through automated data processing and model optimization, accurate measurement of various lengths and circumferences of the human body can be achieved, improving the efficiency and accuracy of the measurement. Accurately detect the key points of the human skeleton and combine them with the parametric human body model. This combination not only improves the fitting accuracy of the human body model, but also enables the model to more realistically reflect the human body characteristics of the individual. Through the accurate identification of the key points of the human skeleton (such as joints and limb endpoints), the present method can ensure that the parametric human body model can maintain a high degree of authenticity and accuracy in both dynamic and static conditions. Guided by the software, users can easily complete the depth scan of the front and back of the human body. The system automatically synchronizes the point cloud data of the front and back and performs spatial registration to ensure the consistency and continuity of the two sets of data (in the case of binocular camera modules) in three-dimensional space, thereby significantly improving the accuracy of the human body model.

[0176] The technical solution of this application realizes the combination of high precision and high efficiency. By integrating 3D structured light technology with parametric human body models, the present invention not only improves the accuracy of human body measurement, but also significantly improves the measurement efficiency through automated processing. This combination ensures that the construction and analysis of the human body model can be completed quickly while maintaining measurement accuracy.

[0177] The technical solution of this application achieves improved flexibility and adaptability. This technical solution provides the flexibility to choose to use a monocular or binocular / multi-camera system according to actual measurement needs. This flexibility allows the system to be adjusted according to different application scenarios and accuracy requirements, thereby optimizing resource allocation and improving the adaptability of the system.

[0178] The technical solution of this application realizes design simplification and cost reduction. By designing a single-sided 3D structured light camera module, the present invention reduces hardware requirements and reduces the complexity and cost of system deployment. In addition, the introduction of the automated processing software module further simplifies the operation process and reduces the need for professional operators, thereby reducing the overall implementation cost.

[0179] In modern application scenarios, human body measurement must not only adapt to static measurement, but also be able to capture dynamic changes and meet personalized needs. The technical solution of this application can adapt to the changes in body shape and posture of different individuals through the flexible adjustment of the parametric human body model, and provide a tailored measurement solution for each person, so as to better serve the fields of personalized clothing design, customized medical equipment manufacturing, etc.

[0180] In the single-sided deployed three-dimensional human body measurement system based on 3D structured light, modules are independently set and can be adjusted by the processor based on different scenarios and needs, which reduces the cost of maintenance and upgrades while ensuring the efficiency and accuracy of data collection and processing.

[0181] The processor can use advanced image processing algorithms, including the Iterative Closest Point (ICP) algorithm, to perform image denoising, enhancement and 3D reconstruction. The system compares and matches the processed data with the preset parametric human body model to optimize the accuracy of the results. In addition, the system also performs pre-processing of point cloud data, including filtering and denoising, background segmentation and outlier removal, to improve the quality of point cloud data and provide an accurate data basis for subsequent model fitting.

[0182] The system guides the operator to complete a deep scan of the front and back of the human body. After the front scan is completed, the system provides clear instructions through an intuitive human-computer interaction interface to guide the object to rotate to the back position for back scanning. The system synchronizes the point cloud data of the front and back, and performs spatial registration to ensure the consistency and continuity of the two sets of data in three-dimensional space. The intuitive and easy-to-use human-computer interaction interface allows the operator to easily set scanning parameters, start and stop the scanning process, and view measurement results in real time. The interface design simplifies the operation process and lowers the technical threshold, allowing non-professionals to use it effectively, thereby expanding the application scope and market acceptance of the technology. The operator can monitor the scanning progress and adjust the scanning parameters or retry according to the system prompts to ensure the integrity and accuracy of the data. The system also provides a data integrity verification function to mark and process missing or abnormal data to ensure the quality of the final output data.

[0183] It should be noted that each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments. The device and system embodiments described above are merely schematic, in which the units described as separate components may or may not be physically separated, and the components indicated as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0184] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A single-sided deployment three-dimensional human body measurement method based on 3D structured light, characterized in that: The method comprises: By deploying a 3D structured light camera module on the first facade, light with structural characteristics is emitted to the front and back of the human body to be tested in turn, and the light reflected from the front and back of the human body to be tested is captured to generate front depth map data and back depth map data of the human body to be tested; Converting the front depth map data into a first portion of point cloud data in a three-dimensional space, and converting the back depth map data into a second portion of point cloud data in the three-dimensional space; Based on the first part of point cloud data and the second part of point cloud data, spatial registration is performed to obtain a point cloud data model corresponding to the human body to be measured; Matching the point cloud data model with a parameterized human body model, and adjusting the parameters of the parameterized human body model by an optimization algorithm to minimize the distance between key points in the parameterized human body model and corresponding key points in the point cloud data model; The three-dimensional measurement data of the human body to be measured is determined according to the parameterized human body model obtained after the fitting is completed.

2. The method according to claim 1, characterized in that Specifically, N 3D structured light camera modules are deployed on the first facade, each of which includes a transmitter and a receiver; N is an integer greater than or equal to 2; The capturing of the light reflected from the front side and the light reflected from the back side of the human body to be tested, and generating the front side depth map data and the back side depth map data of the human body to be tested, comprises: The N receivers respectively capture the light reflected from the front side of the human body to be tested, and generate N groups of front depth map data of the human body to be tested; and the N receivers respectively capture the light reflected from the back side of the human body to be tested, and generate N groups of back depth map data of the human body to be tested; The converting the front depth map data into a first portion of point cloud data in a three-dimensional space, and converting the back depth map data into a second portion of point cloud data in the three-dimensional space, comprises: Using the N groups of front depth map data and the internal parameters corresponding to the N receivers, the N groups of front depth map data are converted one-to-one into N groups of first initial point cloud data; using the N groups of back depth map data and the internal parameters corresponding to the N receivers, the N groups of back depth map data are converted one-to-one into N groups of second initial point cloud data; According to the external parameters corresponding to the N receivers, the N groups of first initial point cloud data are converted into the same three-dimensional space to obtain the first part of point cloud data; according to the external parameters corresponding to the N receivers, the N groups of second initial point cloud data are converted into the same three-dimensional space to obtain the second part of point cloud data.

3. The method according to claim 2, characterized in that The converting the N groups of first initial point cloud data into the same three-dimensional space to obtain the first part of point cloud data includes: After converting the N groups of first initial point cloud data into the same three-dimensional space, performing point cloud registration using an iterative closest point algorithm and a multi-view stereo matching technique, and obtaining the first part of point cloud data after the registration is completed; The converting the N groups of second initial point cloud data into the same three-dimensional space to obtain the second part of point cloud data includes: After converting the N groups of second initial point cloud data into the same three-dimensional space, point cloud registration is performed using an iterative closest point algorithm and a multi-view stereo matching technique, and the second part of point cloud data is obtained after the registration is completed.

4. The method according to claim 1, characterized in that: The performing of spatial registration based on the first part of point cloud data and the second part of point cloud data to obtain a point cloud data model corresponding to the human body to be measured includes: Performing spatial registration on the first portion of point cloud data and the second portion of point cloud data; Based on the coverage of human key points, the integrity of the registered point cloud dataset is checked; If the integrity check of the point cloud data set passes, data quality preprocessing is performed on the point cloud data set, and after the preprocessing is completed, a point cloud data model corresponding to the human body to be tested is obtained; the data quality preprocessing includes at least one of the following: Filtering, background segmentation or outlier removal.

5. The method according to claim 1, characterized in that The 3D structured light camera module deployed on the first facade successively emits light with structural features to the front and back of the human body to be tested, and captures the light reflected from the front and back of the human body to be tested, to generate front depth map data and back depth map data of the human body to be tested, including: Outputting a first measurement guidance prompt to guide the human body to be measured to move into the field of view of the 3D structured light camera module and face the first facade; The 3D structured light camera module emits light with structural features to the front of the human body to be tested, and captures the light reflected from the front of the human body to be tested, so as to generate front depth map data of the human body to be tested; Outputting a second measurement guidance prompt to guide the measured person to rotate the body with the back facing the first facade and still remain within the field of view of the 3D structured light camera module; The 3D structured light camera module emits light with structural features to the back side of the human body to be measured, and captures the light reflected from the back side of the human body to be measured, so as to generate back side depth map data of the human body to be measured.

6. The method according to claim 5, characterized in that While emitting light with structural features to the front of the human body to be measured through the 3D structured light camera module and capturing light reflected from the front of the human body to be measured, the method further includes: Feedback the front scanning progress of the human body to be tested by the 3D structured light camera module in real time through the human-computer interaction interface; While emitting light with structural features to the back of the human body to be measured through the 3D structured light camera module and capturing the light reflected from the back of the human body to be measured, the method further includes: Feedback the scanning progress of the back of the human body to be tested by the 3D structured light camera module in real time through the human-computer interaction interface; The first measurement guidance prompt and the second measurement guidance prompt are both output through a human-computer interaction interface; wherein the second measurement guidance prompt is output by the human-computer interaction interface after the front scanning progress reaches 100%.

7. The method according to claim 1, characterized in that The matching of the point cloud data model with the parameterized human body model and adjusting the parameters of the parameterized human body model by an optimization algorithm to minimize the distance between key points in the parameterized human body model and corresponding key points in the point cloud data model include: Mapping the point cloud data model to the surface of the parameterized human body model to perform corresponding matching of key points; Adjusting the parameters of the parameterized human body model by a nonlinear least squares optimization algorithm to minimize the distance between the key points in the parameterized human body model and the corresponding key points in the point cloud data model to complete the global fitting; After the global fitting, the parameterized human body model is processed in at least one of the following ways: Local refinement of geometric features, simulation of gravity and contact forces based on soft tissue models, or enhancement of geometric features.

8. The method according to claim 1, characterized in that Also includes: Based on the point cloud data model, a fitting effect analysis is performed on the parameterized human body model obtained after fitting, and the fitting effect is displayed or a verification result of the fitting accuracy is output.

9. A single-sided deployment three-dimensional human body measurement device based on 3D structured light, characterized in that: The device comprises: A depth map generation module is used to emit light with structural characteristics to the front and back of the human body to be tested through a 3D structured light camera module deployed on the first facade, and capture the light reflected from the front and back of the human body to be tested, so as to generate front depth map data and back depth map data of the human body to be tested; A point cloud conversion module, configured to convert the front depth map data into a first portion of point cloud data in a three-dimensional space, and to convert the back depth map data into a second portion of point cloud data in the three-dimensional space; A registration module, configured to perform spatial registration based on the first part of point cloud data and the second part of point cloud data to obtain a point cloud data model corresponding to the human body to be measured; A fitting module, used for matching the point cloud data model with a parameterized human body model, and adjusting the parameters of the parameterized human body model through an optimization algorithm to minimize the distance between key points in the parameterized human body model and corresponding key points in the point cloud data model; The measurement data determination module is used to determine the three-dimensional measurement data of the human body to be measured according to the parameterized human body model obtained after the fitting is completed.

10. A single-sided deployment three-dimensional human body measurement system based on 3D structured light, characterized in that: include: Processor, memory, 3D structured light camera module and display; The memory, the 3D structured light camera module and the display are all communicatively connected to the processor; The 3D structured light camera module is deployed on the first facade; The memory stores a computer program; The processor is used to run the computer program, and when the computer program is run, the single-sided deployment three-dimensional human body measurement method based on 3D structured light is executed as described in any one of claims 1 to 8; The display is used to display a human-computer interaction interface, and outputs measurement guidance prompts for the front and back sides through the human-computer interaction interface.

11. The system according to claim 10, characterized in that Specifically, N 3D structured light camera modules are deployed on the first facade, each of which includes a transmitter and a receiver; N is an integer greater than or equal to 2; The N receivers are arranged along the vertical direction, and the union of the visual fields of the N receivers covers the entire human body to be measured.

12. The method according to claim 10, characterized in that The exposure times of the infrared image and the depth image generated by the 3D structured light camera module are synchronized.

Images