Intelligent body measurement method and system based on multi-sensor fusion

Abstract

The invention discloses an intelligent body measurement method and system based on multi-sensor fusion, belongs to the technical field of multi-source data fusion, and realizes high-precision and comprehensive measurement of body characteristics of a target object by comprehensively using various data sources such as images, videos and gravity sensors. The method comprises the following steps: acquiring original data of a target object, and carrying out preprocessing and feature extraction to obtain a plurality of measurement data sets such as images, videos and gravity; and then, fusing the data sets to construct a three-dimensional intelligent body measurement model of the target object in different attitudes, and performing optimization by using an SCAPE model. According to the method, the precision and comprehensiveness of body measurement are improved, intelligence and automation are also realized, the working efficiency is improved, and powerful technical support is provided for the fields of clothing customization, body shape analysis and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-source data fusion, and more particularly to an intelligent body measurement method and system based on multi-sensor fusion. Background Art

[0002] In the pursuit of personalized clothing customization, accurate body measurement technology has become the key to achieving this goal. Although the traditional manual measurement method meets basic needs to a certain extent, its accuracy is limited by the experience and skills of the measurer, and the process is cumbersome and inefficient. With the advancement of technology, advanced equipment such as 3D scanners have been introduced into the field of clothing design. They scan the human body through laser or grating technology to obtain more accurate 3D data. However, the high cost and operational complexity of such equipment limit its widespread application in ordinary clothing design studios or retail stores.

[0003] In order to overcome these limitations, emerging technologies such as smart body-measuring garments and 3D smart body-measuring systems have emerged in recent years. Smart body-measuring garments combine technologies such as pressure sensors and RFID tags to obtain body data by wearing body-measuring garments and taking photos. Although it provides a new way of measuring body, the accuracy of the data is still affected by many factors such as shooting angle and light. The 3D smart body-measuring system uses high-precision 3D scanning equipment combined with artificial intelligence algorithms to achieve fast and accurate body measurement, but the high cost and difficulty of data processing are still obstacles to its promotion.

[0004] Therefore, providing a low-cost and high-accuracy intelligent body measurement method and system is an urgent problem to be solved by those skilled in the art. Summary of the invention

[0005] In view of this, the present invention provides an intelligent body measurement method and system based on multi-sensor fusion, which integrates and utilizes data from multiple sensors to achieve comprehensive and accurate body measurement of a target object.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] An intelligent body measurement method based on multi-sensor fusion, comprising:

[0008] Performing data collection on a target object to obtain original data of the target object;

[0009] Preprocessing the raw data, extracting key features from the preprocessed raw data, and classifying the preprocessed raw data according to the key features to obtain multiple measurement data sets, wherein the measurement data sets include image data sets, video data sets, and gravity data sets;

[0010] The image data set, the video data set, and the gravity data set are integrated to construct an intelligent body model of the target object in different postures;

[0011] A body measurement report is automatically generated according to the intelligent body measurement model.

[0012] Preferably, preprocessing the raw data includes:

[0013] Establishing a reference coordinate system, and performing spatial position alignment on the original data based on the reference coordinate system to obtain spatially aligned data;

[0014] The data acquisition frequency is extracted based on the sensor, and a reference frequency is established, and the time dimension alignment is performed on the spatial alignment data according to the reference frequency and the data acquisition frequency.

[0015] Preferably, fusing the image dataset, the video dataset, and the gravity dataset to construct an intelligent body model of the target object in different postures includes:

[0016] acquiring the weight of the target object based on the gravity data set;

[0017] Acquire posture data of the target object based on the video data set;

[0018] constructing a three-dimensional model of the target object based on the image data set;

[0019] Optimizing the three-dimensional model based on the body weight to obtain an accurate three-dimensional model;

[0020] An intelligent body model of the target object in different postures is constructed based on the posture data, the precise three-dimensional model and the video data set.

[0021] Preferably, acquiring the posture data of the target object based on the video data set includes:

[0022] Using the local maximum value method of inter-frame difference and the adaptive similarity threshold method to screen the key frames of the video data set, and obtain the gesture recognition frame image;

[0023] The openpose posture estimation algorithm is used to extract the skeleton key point features of the posture recognition frame image to obtain human skeleton information, which is the posture data.

[0024] Preferably, constructing the three-dimensional model of the target object based on the image data set includes:

[0025] Inputting the image data set into a preset feature extraction model to obtain human body contour features;

[0026] Inputting the human body contour features into the first encoding layer of a pre-trained three-dimensional modeling model to obtain vertex features, joint rotation features, and shape features of the target object;

[0027] Inputting device parameters for acquiring image data into the first coding layer to obtain device characteristics;

[0028] Inputting the human body contour features, the vertex features, the joint point rotation features, the shape features and the device features into a second coding layer in the three-dimensional modeling model to obtain coding features;

[0029] According to the coding features, a vertex coordinate set of each model vertex and a joint point coordinate set of each joint point contained in the three-dimensional model corresponding to the target object are obtained;

[0030] Three-dimensional modeling is performed according to the vertex coordinate sets of each model vertex and the joint point coordinate sets of each joint point to obtain a three-dimensional model of the target object.

[0031] Preferably, constructing an intelligent body model of the target object in different postures based on the posture data and the accurate three-dimensional model includes:

[0032] Based on the posture data, the SCAPE model is used to obtain the three-dimensional coordinates of the joint points and the joint points of the skeleton in different postures;

[0033] Adjusting the precise three-dimensional model based on the three-dimensional coordinates of the joint points and the joint points of the skeleton to obtain a posture three-dimensional model;

[0034] The posture three-dimensional model is optimized according to the video data set to obtain intelligent body models in different postures.

[0035] On the other hand, the present invention also discloses an intelligent body measurement system based on multi-sensor fusion, comprising:

[0036] A data acquisition module is used to collect data from a target object and obtain the original data of the target object;

[0037] A data set module, used to preprocess the raw data, extract key features from the preprocessed raw data, and classify the preprocessed raw data according to the key features to obtain multiple measurement data sets, wherein the measurement data sets include image data sets, video data sets, and gravity data sets;

[0038] A modeling module, used to fuse the image data set, the video data set, and the gravity data set to construct an intelligent body model of the target object in different postures;

[0039] The output module is used to automatically generate a measurement report based on the intelligent measurement model.

[0040] Preferably, the modeling module includes:

[0041] a weight calculation unit, configured to obtain the weight of the target object based on the gravity data set;

[0042] A posture acquisition unit, used for acquiring posture data of the target object based on the video data set;

[0043] A three-dimensional model unit, configured to construct a three-dimensional model of the target object based on the image data set;

[0044] An optimization unit, configured to optimize the three-dimensional model based on the body weight to obtain an accurate three-dimensional model;

[0045] An intelligent body model unit is used to construct an intelligent body model of the target object in different postures based on the posture data, the precise three-dimensional model and the video data set.

[0046] Preferably, the posture acquisition unit includes:

[0047] A screening unit, used to screen key frames of the video data set by using a local maximum value method of inter-frame difference and an adaptive similarity threshold method to obtain a gesture recognition frame image;

[0048] The information extraction unit is used to extract the skeleton key point features of the posture recognition frame image using the openpose posture estimation algorithm to obtain human skeleton information, and the human skeleton information is posture data.

[0049] Preferably, the intelligent volume model unit comprises:

[0050] A coordinate calculation unit, for obtaining three-dimensional coordinates of joint points and skeleton joint points in different postures using a SCAPE model based on the posture data;

[0051] A posture conversion unit, used for adjusting the precise three-dimensional model based on the three-dimensional coordinates of the joint points and the joint points of the skeleton to obtain a posture three-dimensional model;

[0052] The model optimization unit is used to optimize the posture three-dimensional model according to the video data set to obtain intelligent body models in different postures.

[0053] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses an intelligent measurement method and system based on multi-sensor fusion, which realizes high-precision and comprehensive measurement of the body shape characteristics of the target object by comprehensively using multiple data sources such as images, videos and gravity sensors. The present invention first uses multiple sensors to collect the original data of the target object, and ensures the accuracy and consistency of the data through preprocessing and alignment of spatial and temporal dimensions. In the data fusion stage, the present invention combines the gravity data set to obtain weight information, uses the video data set to capture posture data, and builds a three-dimensional model of the target object based on the image data set. The three-dimensional model is optimized by weight information to obtain an accurate three-dimensional model. Finally, combined with the posture data and video data set, the advanced SCAPE model is used to construct an intelligent measurement model of the target object in different postures, which not only significantly improves the accuracy and comprehensiveness of the measurement, but also realizes the intelligence and automation of the measurement process, greatly improving the work efficiency. At the same time, the present invention has strong adaptability to target objects with different postures, can bring users a more convenient and accurate measurement experience, and provide strong technical support for the fields of clothing customization and body shape analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention 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 embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0055] Figure 1 The present invention provides a flow chart of the intelligent body measurement method.

[0056] Figure 2 This is a framework diagram of the intelligent body measurement system provided by the present invention. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] The embodiment of the present invention discloses an intelligent body measurement method based on multi-sensor fusion, such as Figure 1 As shown, including:

[0059] Step 1: Collect data on the target object and obtain the original data of the target object.

[0060] Comprehensive data collection of the target object is the first step in building an intelligent measurement model. We use a variety of advanced equipment such as high-definition cameras, dynamic video capture equipment, and high-precision gravity sensors to observe and measure the target object in detail, thereby obtaining detailed and rich original data. In order to obtain more accurate data, the same data can be obtained using a variety of different sensors or equipment.

[0061] Step 2: Preprocess the raw data, extract key features from the preprocessed raw data, and classify the preprocessed raw data according to the key features to obtain multiple measurement data sets, including image data sets, video data sets, and gravity data sets.

[0062] Preprocessing includes but is not limited to data cleaning, denoising, and format unification to ensure data accuracy and consistency. On the basis of preprocessing, key features in the data are further extracted, which can distinguish data types. According to these extracted key features, the preprocessed raw data is scientifically classified to form multiple measurement data sets, such as image data sets, video data sets, and gravity data sets, which provide strong support for subsequent data fusion and model construction.

[0063] Step 3: Fuse the image dataset, video dataset, and gravity dataset to build an intelligent body measurement model of the target object in different postures.

[0064] In the data fusion stage, the complementary advantages of images, videos and gravity data are fully utilized, and they are organically integrated to build an intelligent body measurement model that can accurately reflect the body shape characteristics of the target object in different postures. This model is not only highly accurate and robust, but can also be flexibly adjusted and expanded according to actual needs, making the designed clothing more in line with human posture.

[0065] Step 4: Automatically generate a measurement report based on the intelligent measurement model.

[0066] The generated measurement report contains the body data of the target object in various postures, including normal state, daily state, working state, etc. It also conducts a comprehensive and in-depth analysis of its body characteristics, providing strong data support and decision-making basis for subsequent clothing customization, body management and other work.

[0067] Furthermore, the raw data is preprocessed including:

[0068] A reference coordinate system is established, and the original data is spatially aligned based on the reference coordinate system to obtain spatially aligned data; the reference coordinate system provides a unified reference framework to ensure that all original data are consistent and comparable in spatial position.

[0069] The data acquisition frequency is extracted based on the sensor, and a reference frequency is established. The time dimension alignment is performed on the spatially aligned data according to the reference frequency and the data acquisition frequency. In addition to the alignment of the spatial position, the alignment of the time dimension is also an indispensable part of data preprocessing. Since different sensors may have different data acquisition frequencies, these frequencies need to be extracted and a reference frequency needs to be established. The selection of the reference frequency is usually based on the needs and characteristics of the data to ensure that the alignment in the time dimension can accurately reflect the actual changes in the data.

[0070] Specifically, the image dataset, video dataset, and gravity dataset are integrated to construct an intelligent volume model of the target object in different postures, including:

[0071] Obtaining the weight of the target object based on the gravity data set;

[0072] The posture data of the target object is obtained based on the video dataset; the video dataset records the dynamic performance of the target object in different time periods and scenes. Through image processing and machine learning technology, the posture changes of the target object can be accurately captured, including the position and angle of the limbs and the changes of various parts of the body in different postures.

[0073] Build a 3D model of the target object based on the image dataset;

[0074] The three-dimensional model is optimized based on the weight to obtain an accurate three-dimensional model; the preliminary three-dimensional model constructed using the image data set often has certain errors or deficiencies, so it needs to be optimized. In this step, the embodiment of the present invention makes full use of the previously obtained weight information, and adjusts the parameters such as the volume and density of the model to make it more consistent with the actual situation of the target object. After the optimization process, we obtain an accurate three-dimensional model, which is highly similar to the target object in shape, appearance and physical properties.

[0075] Based on posture data, accurate 3D models and video data sets, an intelligent measurement model of the target object in different postures is constructed. This model can not only accurately reflect the shape and appearance changes of the target object in different postures, but also dynamically adjust and optimize according to needs. Through this model, it is more convenient to carry out virtual fitting, size measurement and personalized customization, providing strong support for the development of the clothing industry and other related fields.

[0076] Furthermore, the posture data of the target object is obtained based on the video data set, including:

[0077] The local maximum value method of inter-frame difference and the adaptive similarity threshold method are used to screen the key frames of the video dataset to obtain the gesture recognition frame images.

[0078] Specifically, in order to efficiently extract the most representative frames when processing video data, the embodiment of the present invention adopts a strategy combining the local maximum method of inter-frame difference and the adaptive similarity threshold method. First, the key frames are roughly extracted by the local maximum method of inter-frame difference, and then the adaptive similarity threshold method is used for further fine extraction to obtain the final gesture recognition frame image.

[0079] Further, in the rough extraction stage, the following steps are performed:

[0080] Get all the image frames contained in the video dataset, and for each two adjacent image frames in the video, calculate the inter-frame difference between them, so as to obtain the inter-frame difference intensity corresponding to the two image frames. The purpose of this step is to find the degree of change between adjacent frames.

[0081] Based on the obtained multiple inter-frame differential intensities, the average inter-frame differential intensity is calculated. On this basis, the local maximum of the average inter-frame differential intensity is further determined. The image frames corresponding to these local maxima often represent significant change points in the video, and therefore can be regarded as key frames where the posture changes.

[0082] An image frame with a local maximum value is used as an initial key frame, and the obtained multiple initial key frames are sorted in order of frame numbers to generate an initial key frame sequence.

[0083] After obtaining the initial key frame sequence, we further use the adaptive similarity threshold method to extract the key frames. The specific steps are as follows:

[0084] For each initial key frame, its size is scaled to 8×8 to obtain the adjusted image frame.

[0085] The adjusted image frame is processed with 64 gray levels to obtain a pre-processed image frame corresponding to the initial key frame. The pre-processed image frame contains 64 pixels, each of which has a gray value.

[0086] Determine the grayscale value of each pixel in the preprocessed image frame corresponding to each initial key frame, and calculate the comparison value corresponding to the initial key frame based on the current grayscale threshold and the grayscale value corresponding to the initial key frame. The grayscale threshold here is a dynamic value, which will be adjusted according to the change range of the grayscale average value of the initial key frame whose corresponding comparison value has been determined. The comparison value is the number of pixels in the preprocessed image frame whose grayscale value is greater than or equal to the current grayscale threshold.

[0087] Compare the comparison values ​​of two adjacent initial key frames. If the difference between the comparison value of the initial key frame with the previous frame number and the comparison value of the initial key frame with the next frame number is less than or equal to the preset comparison threshold, then it is considered that the content similarity of the two key frames is high, so only the initial key frame with the previous frame number is retained as the gesture recognition frame image. If the difference is greater than the comparison threshold, it is considered that the content of the two key frames has changed significantly, so both initial key frames are retained as gesture recognition frame images.

[0088] The openpose posture estimation algorithm is used to extract the skeleton key point features of the posture recognition frame image to obtain the human skeleton information, which is the posture data.

[0089] Furthermore, a 3D model of the target object is constructed based on the image dataset, including:

[0090] Input the image data set into the preset feature extraction model to obtain the human body contour features;

[0091] Input the human body contour features into the first encoding layer of the pre-trained 3D modeling model to obtain the vertex features, joint rotation features, and shape features of the target object;

[0092] Inputting device parameters for acquiring image data into the first coding layer to obtain device features;

[0093] Inputting the human body contour features, vertex features, joint rotation features, shape features, and device features into the second coding layer of the three-dimensional modeling model to obtain coding features;

[0094] According to the coding features, a vertex coordinate set of each model vertex and a joint point coordinate set of each joint point contained in the three-dimensional model corresponding to the target object are obtained;

[0095] Three-dimensional modeling is performed according to the vertex coordinate set of each model vertex and the joint point coordinate set of each joint point to obtain a three-dimensional model of the target object.

[0096] In another embodiment, constructing an intelligent body model of a target object in different postures based on posture data and an accurate three-dimensional model includes:

[0097] Based on the posture data, the SCAPE model is used to obtain the three-dimensional coordinates of the joint points and the skeleton joint points in different postures;

[0098] The precise three-dimensional model is adjusted based on the three-dimensional coordinates of the joint points and the skeleton joint points to obtain a posture three-dimensional model;

[0099] The posture 3D model is optimized according to the video data set to obtain the intelligent body model in different postures.

[0100] On the other hand, the present invention also discloses an intelligent body measurement system based on multi-sensor fusion, such as Figure 2 As shown, including:

[0101] A data acquisition module is used to collect data from a target object and obtain the original data of the target object;

[0102] The data set module is used to preprocess the raw data, extract key features from the preprocessed raw data, and classify the preprocessed raw data according to the key features to obtain multiple measurement data sets, including image data sets, video data sets, and gravity data sets;

[0103] Modeling module, used to fuse image data sets, video data sets, and gravity data sets to build intelligent volume models of target objects in different postures;

[0104] The output module is used to automatically generate a measurement report based on the intelligent measurement model.

[0105] Specifically, the modeling module includes:

[0106] A weight calculation unit, used for obtaining the weight of the target object based on the gravity data set;

[0107] A posture acquisition unit, used for acquiring posture data of a target object based on a video data set;

[0108] A three-dimensional model unit, used for constructing a three-dimensional model of a target object based on an image data set;

[0109] An optimization unit, used for optimizing the three-dimensional model based on body weight to obtain an accurate three-dimensional model;

[0110] The intelligent volume model unit is used to construct an intelligent volume model of the target object in different postures based on posture data, accurate three-dimensional model and video data set.

[0111] Specifically, the posture acquisition unit includes:

[0112] A screening unit, used to screen key frames of a video data set using a local maximum value method of inter-frame differences and an adaptive similarity threshold method to obtain a gesture recognition frame image;

[0113] The information extraction unit is used to extract the skeleton key point features of the posture recognition frame image using the openpose posture estimation algorithm to obtain the human skeleton information, which is the posture data.

[0114] Preferably, the intelligent volume model unit comprises:

[0115] A coordinate calculation unit, used to obtain the three-dimensional coordinates of the joint points and the skeleton joint points in different postures using the SCAPE model based on the posture data;

[0116] A posture conversion unit, used for adjusting the precise three-dimensional model based on the three-dimensional coordinates of the joint points and the joint points of the skeleton to obtain a posture three-dimensional model;

[0117] The model optimization unit is used to optimize the posture three-dimensional model according to the video data set to obtain the intelligent volume model under different postures.

[0118] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0119] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent body measurement method based on multi-sensor fusion, characterized in that: include: Performing data collection on a target object to obtain original data of the target object; Preprocessing the raw data, extracting key features from the preprocessed raw data, and classifying the preprocessed raw data according to the key features to obtain multiple measurement data sets, wherein the measurement data sets include image data sets, video data sets, and gravity data sets; The image data set, the video data set, and the gravity data set are integrated to construct an intelligent body model of the target object in different postures; A body measurement report is automatically generated according to the intelligent body measurement model.

2. The intelligent body measurement method based on multi-sensor fusion according to claim 1, characterized in that: Preprocessing the raw data includes: Establishing a reference coordinate system, and performing spatial position alignment on the original data based on the reference coordinate system to obtain spatially aligned data; The data acquisition frequency is extracted based on the sensor, and a reference frequency is established, and the time dimension alignment is performed on the spatial alignment data according to the reference frequency and the data acquisition frequency.

3. The intelligent body measurement method based on multi-sensor fusion according to claim 1, characterized in that: The step of fusing the image data set, the video data set, and the gravity data set to construct an intelligent body model of the target object in different postures includes: acquiring the weight of the target object based on the gravity data set; Acquire posture data of the target object based on the video data set; constructing a three-dimensional model of the target object based on the image data set; Optimizing the three-dimensional model based on the body weight to obtain an accurate three-dimensional model; An intelligent body model of the target object in different postures is constructed based on the posture data, the precise three-dimensional model and the video data set.

4. The intelligent body measurement method based on multi-sensor fusion according to claim 3 is characterized in that: Acquiring the posture data of the target object based on the video data set includes: Using the local maximum value method of inter-frame difference and the adaptive similarity threshold method to screen the key frames of the video data set, and obtain the gesture recognition frame image; The openpose posture estimation algorithm is used to extract the skeleton key point features of the posture recognition frame image to obtain human skeleton information, which is the posture data.

5. The intelligent body measurement method based on multi-sensor fusion according to claim 3 is characterized in that: Constructing a three-dimensional model of the target object based on the image data set includes: Inputting the image data set into a preset feature extraction model to obtain human body contour features; Inputting the human body contour features into the first encoding layer of a pre-trained three-dimensional modeling model to obtain vertex features, joint rotation features, and shape features of the target object; Inputting device parameters for acquiring image data into the first coding layer to obtain device characteristics; Inputting the human body contour features, the vertex features, the joint point rotation features, the shape features and the device features into a second coding layer in the three-dimensional modeling model to obtain coding features; According to the coding features, a vertex coordinate set of each model vertex and a joint point coordinate set of each joint point contained in the three-dimensional model corresponding to the target object are obtained; Three-dimensional modeling is performed according to the vertex coordinate sets of each model vertex and the joint point coordinate sets of each joint point to obtain a three-dimensional model of the target object.

6. The intelligent body measurement method based on multi-sensor fusion according to claim 3 is characterized in that: Building an intelligent body model of the target object in different postures based on the posture data and the accurate three-dimensional model includes: Based on the posture data, the SCAPE model is used to obtain the three-dimensional coordinates of the joint points and the joint points of the skeleton in different postures; Adjusting the precise three-dimensional model based on the three-dimensional coordinates of the joint points and the joint points of the skeleton to obtain a posture three-dimensional model; The posture three-dimensional model is optimized according to the video data set to obtain intelligent body models in different postures.

7. An intelligent body measurement system based on multi-sensor fusion, characterized in that: A data acquisition module is used to collect data from a target object and obtain the original data of the target object; A data set module, used to preprocess the raw data, extract key features from the preprocessed raw data, and classify the preprocessed raw data according to the key features to obtain multiple measurement data sets, wherein the measurement data sets include image data sets, video data sets, and gravity data sets; A modeling module, used to fuse the image data set, the video data set, and the gravity data set to construct an intelligent body model of the target object in different postures; The output module is used to automatically generate a measurement report based on the intelligent measurement model.

8. The intelligent body measurement system based on multi-sensor fusion according to claim 7, characterized in that: The modeling module includes: a weight calculation unit, configured to obtain the weight of the target object based on the gravity data set; A posture acquisition unit, used for acquiring posture data of the target object based on the video data set; A three-dimensional model unit, configured to construct a three-dimensional model of the target object based on the image data set; An optimization unit, configured to optimize the three-dimensional model based on the body weight to obtain an accurate three-dimensional model; An intelligent body model unit is used to construct an intelligent body model of the target object in different postures based on the posture data, the precise three-dimensional model and the video data set.

9. The intelligent body measurement system based on multi-sensor fusion according to claim 8, characterized in that: The posture acquisition unit includes: A screening unit, used to screen key frames of the video data set by using a local maximum value method of inter-frame difference and an adaptive similarity threshold method to obtain a gesture recognition frame image; The information extraction unit is used to extract the skeleton key point features of the posture recognition frame image using the openpose posture estimation algorithm to obtain human skeleton information, and the human skeleton information is posture data.

10. The intelligent body measurement system based on multi-sensor fusion according to claim 8, characterized in that: The intelligent volume model unit includes: A coordinate calculation unit, for obtaining three-dimensional coordinates of joint points and skeleton joint points in different postures using a SCAPE model based on the posture data; A posture conversion unit, used for adjusting the precise three-dimensional model based on the three-dimensional coordinates of the joint points and the joint points of the skeleton to obtain a posture three-dimensional model; The model optimization unit is used to optimize the posture three-dimensional model according to the video data set to obtain intelligent body models in different postures.

Images