WO2023030036A1 - 基于多摄像头的人体经络识别方法和装置及人体经络调理设备 - Google Patents

基于多摄像头的人体经络识别方法和装置及人体经络调理设备 Download PDF

Info

Publication number
WO2023030036A1
WO2023030036A1 PCT/CN2022/113488 CN2022113488W WO2023030036A1 WO 2023030036 A1 WO2023030036 A1 WO 2023030036A1 CN 2022113488 W CN2022113488 W CN 2022113488W WO 2023030036 A1 WO2023030036 A1 WO 2023030036A1
Authority
WO
WIPO (PCT)
Prior art keywords
meridian
human body
recognition
camera
identification
Prior art date
Application number
PCT/CN2022/113488
Other languages
English (en)
French (fr)
Inventor
王亮
Original Assignee
中科尚易健康科技(北京)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中科尚易健康科技(北京)有限公司 filed Critical 中科尚易健康科技(北京)有限公司
Publication of WO2023030036A1 publication Critical patent/WO2023030036A1/zh

Links

Images

Definitions

  • the present application relates to the technical field of intelligent medical conditioning machinery and equipment, in particular to a multi-camera-based human meridian recognition method and device and human body meridian conditioning equipment.
  • Deep learning is a new field in machine learning. Its motivation is to establish and simulate the neural network of the human brain for analysis and learning, and to analyze the data by imitating the mechanism of the human brain through the established neural network, such as: images, sounds and text etc. Realizing the recognition of human meridians by using deep learning has a very good effect on improving the efficiency of human meridian identification and acupoint selection.
  • the neural network is used to identify the meridians of the human body, since the collected images are collected by a single camera, and a single camera cannot obtain all the images of the human body, this makes it impossible to use depth
  • the recognition results obtained when learning to recognize human meridians are not high enough in accuracy and completeness.
  • the present application proposes a multi-camera-based human meridian recognition method, which can effectively improve the accuracy and completeness of human meridian recognition results.
  • a multi-camera-based human meridian recognition method including:
  • All the meridian recognition results are integrated and processed to obtain the final human meridian recognition result.
  • the camera calibration is used to obtain the transformation matrix from each camera to the mechanical arm in the human body meridian conditioning equipment, and the transformation matrix is performed based on the transformation matrix. Coordinate transformation of each meridian identification result.
  • the coordinate conversion of each meridian identification result is performed, including:
  • the three-dimensional coordinates of each recognition target in the corresponding camera coordinate system are multiplied by the conversion matrix to obtain the three-dimensional coordinates of each recognition target in the robot arm coordinate system.
  • a multi-camera-based human meridian recognition device including: an image acquisition module, an image recognition module and a result integration module;
  • the image acquisition module is configured to acquire the human body images collected by the cameras arranged on the human body meridian conditioning equipment;
  • the image recognition module is configured to input each of the human body images into a pre-built meridian recognition neural network, and use the meridian recognition neural network to recognize each of the human body images to obtain the meridian of each of the human body images recognition result;
  • the result integration module is configured to integrate and process each of the meridian identification results to obtain the final human meridian identification result.
  • a coordinate conversion processing module is also included;
  • the coordinate conversion processing module is configured to perform coordinate conversion processing on each of the meridian identification results when the result integration module integrates each of the meridian identification results to obtain the final human body meridian identification result.
  • the coordinate conversion processing module when configured to perform coordinate conversion on each of the meridian recognition results, through camera calibration, obtain the mechanical A transformation matrix of the arm, performing coordinate transformation of each of the meridian identification results based on the transformation matrix.
  • the coordinate conversion processing module includes a coordinate acquisition submodule, a depth information addition submodule, and a coordinate conversion submodule;
  • the coordinate acquisition submodule is configured to acquire the two-dimensional coordinates of each identification target in the corresponding camera coordinate system in the meridian identification result;
  • the depth information adding sub-module is configured to add corresponding depth information to the two-dimensional coordinates of each of the recognition targets and convert them into three-dimensional coordinates of the recognition targets in the corresponding camera coordinate system;
  • the coordinate transformation sub-module is configured to multiply the three-dimensional coordinates of each recognition target in the corresponding camera coordinate system by point multiplying the conversion matrix to obtain the three-dimensional coordinates of each recognition target in the robot arm coordinate system.
  • a human body meridian conditioning device including:
  • memory for storing processor-executable instructions
  • the processor is configured to implement any one of the aforementioned methods when executing the executable instructions
  • the multiple cameras are respectively set at different positions of the human body meridian conditioning device.
  • the number of the cameras is four;
  • the four cameras are respectively arranged at the left side, the right front side and the top position of the conditioning bed in the human body meridian conditioning equipment.
  • multiple cameras are arranged on the equipment for regulating the meridians of the human body, and the multiple cameras collect images of the human body from different angles, and then based on the images collected from different angles
  • the human body image is used to identify the meridians of the human body, and then the meridian recognition results recognized by the human body images from different angles are integrated to obtain the final human meridian recognition result, which makes the obtained human meridian recognition result more complete.
  • Fig. 1 shows the flow chart of the human body meridian recognition method based on multi-camera of the embodiment of the present application
  • FIG. 2 shows a schematic diagram of the installation structure of each camera in the human body meridian conditioning device in the multi-camera-based human meridian recognition method according to the embodiment of the present application;
  • FIG. 3 shows a structural block diagram of a multi-camera-based human meridian recognition device according to an embodiment of the present application.
  • Fig. 1 shows a flowchart of a multi-camera-based human meridian recognition method according to an embodiment of the present application.
  • the method includes: Step S100 , acquiring human body images collected by cameras configured on the human body meridian conditioning device.
  • the device for regulating the meridians of the human body may be an instrument for regulating the meridians of the human body, as shown in FIG. 2 .
  • the method in the embodiment of the present application can be specifically applied to an instrument for regulating the meridians of the human body.
  • a plurality of cameras are installed on the human body meridian conditioning equipment, and each camera is respectively installed at different positions of the human body meridian conditioning equipment, so that each camera can collect human body images from different directions.
  • step S200 can be executed to input each human body image into the pre-built meridian recognition neural network, and the meridian recognition neural network will analyze each human body The image is recognized, and the meridian recognition result of each human body image is obtained. Furthermore, through step S300, the recognition results of the meridians are integrated to obtain the final human meridian recognition result.
  • multiple cameras are arranged on the equipment for regulating the meridians of the human body, and the multiple cameras collect images of the human body from different angles, and then based on different angles
  • the collected human body images are used to identify the meridians of the human body, and then the meridian recognition results identified from the human body images from different angles are integrated to obtain the final human meridian recognition results, which makes the obtained human meridian recognition results more complete.
  • the constructed meridian recognition neural network can adopt a conventional target recognition network model in the field, and will not be repeated here.
  • the meridian recognition neural network needs to be trained first.
  • the training samples in the training sample set constructed are various human body images collected through actual shooting and from the network, and combined with the human body acupoint map in each human body image The image data of the corresponding human acupoints are marked in .
  • the process of integrating the meridian recognition results of each human body image includes an operation of performing coordinate transformation on the meridian recognition results of each human body image.
  • the meridian recognition results of each human body image are two-dimensional coordinates in the coordinate system of each camera, and the coordinate systems of different cameras are different.
  • the coordinate system of each camera is also different from the coordinate system of the mechanical arm in the human body meridian conditioning equipment. Therefore, in order to facilitate the integration of the meridian recognition results of each human body image, it is necessary to unify the coordinate systems of the meridian recognition results of each human body image.
  • the coordinates of the meridian recognition results of each human body image can be converted into coordinates in the coordinate system of the mechanical arm in the human body meridian conditioning device, so that the mechanical arm performs massage according to the identified acupoints Or conditioning can smoothly and accurately identify the location of each acupuncture point.
  • the transformation matrix from each camera to the mechanical arm in the human body meridian conditioning device can be obtained through camera calibration, and the coordinate transformation of each meridian recognition result can be performed based on the transformation matrix. That is, by calibrating the coordinate system of each camera and the coordinate system of the mechanical arm in the human body meridian conditioning equipment, and then obtaining the conversion between each camera and the mechanical arm according to the calibrated coordinate system of each camera and the coordinate system of the mechanical arm matrix.
  • the image acquisition device-robot coordinate transformation matrix can be realized by calibrating the coordinates of the image acquisition device and the robot arm respectively, and then performing corresponding calculations based on the calibrated image acquisition device coordinates and the robot arm coordinates.
  • the coordinate calibration of the image acquisition device and the mechanical arm can be performed in a checkerboard manner.
  • the image acquisition device obtains the three-dimensional coordinates of the center of the checkerboard under the coordinates of the image acquisition device
  • the robotic arm obtains the center of the checkerboard at
  • the manipulator walks according to a certain 3*3 grid, and obtains the center coordinates of the checkerboard under the coordinate system of the image acquisition equipment of 9 groups and the coordinates of the center of the checkerboard under the coordinate system of the manipulator.
  • the calculation can obtain a transformation matrix from the coordinates of the image acquisition device to the coordinates of the manipulator (that is, the coordinate transformation matrix of the image acquisition device-manipulator)
  • the coordinate conversion of the recognition results of each meridian can be realized in the following manner.
  • each identification target refers to the acupuncture points of the human body identified from the image of the human body.
  • the two-dimensional coordinates of each recognition target in the corresponding camera coordinate system can be obtained according to the position of each body acupoint identified in the human body image in the human body image.
  • the corresponding depth information can be added through the Azure Kinect SDK, here also No further details will be given.
  • the coordinate data of each recognition target is still the coordinate data in the corresponding camera coordinate system. Therefore, it is also necessary to transform the coordinate data of each recognition target into the coordinate system of the robot arm so that the robot arm can smoothly and accurately identify the position of each acupuncture point of the human body.
  • the conversion of each recognition target from the camera coordinate system to the robot arm coordinate system can be realized by means of dot product conversion matrix.
  • dot product conversion matrix for different meridian recognition results of human body images, different coordinate transformation matrices between the camera and the robotic arm need to be used.
  • each camera is located at a different position of the human body meridian conditioning device, namely camera A, camera B, camera C and camera D.
  • the camera A collects the human body image PictureA
  • the camera B collects the human body image PictureB
  • the camera C collects the human body image PictureC
  • the camera D collects the human body image PictureD.
  • the above-mentioned human body image PictureA, human body image PictureB, human body image PictureC and human body image PictureD are sequentially input into the meridian recognition neural network, and the meridian recognition of the above-mentioned human body images is carried out by the meridian recognition neural network, and the corresponding meridians are identified from each human body image.
  • Human body acupuncture points At this time, the identified coordinate data of each acupuncture point of the human body is in the corresponding camera coordinate system. Therefore, it is necessary to carry out coordinate transformation of each acupoint point in each human body image, and transform the coordinates of each acupoint point in each human body image into the mechanical arm coordinate system.
  • the coordinate calibration of each camera and the coordinate calibration of the manipulator can be realized by conventional coordinate calibration techniques in the field, and at the same time, based on the calibrated coordinate system of each camera and the coordinate system of the manipulator, the The transformation matrix between each camera and the mechanical arm can also be realized by conventional technical means in the field, and will not be repeated here.
  • the present application also provides a device for identifying human meridians based on cameras. Since the working principle of the multi-camera-based human meridian recognition device provided by the present application is the same or similar to the principle of the multi-camera-based human meridian recognition method of the present application, the repeated descriptions will not be repeated.
  • the multi-camera human meridian recognition device 100 includes an image acquisition module 110 , an image recognition module 120 and a result integration module 130 .
  • the image acquisition module 110 is configured to acquire human body images collected by cameras arranged on the human body meridian conditioning equipment.
  • the image recognition module 120 is configured to input each human body image into the pre-built meridian recognition neural network, recognize each human body image by the meridian recognition neural network, and obtain the meridian recognition result of each human body image.
  • the result integration module 130 is configured to integrate the recognition results of the meridians to obtain the final human meridian recognition result.
  • a coordinate conversion processing module (not shown in the figure) is also included.
  • the coordinate conversion processing module is configured to perform coordinate conversion processing on each meridian identification result when the result integration module 130 integrates each meridian identification result to obtain the final human body meridian identification result.
  • the coordinate conversion processing module when configured to perform coordinate conversion on the recognition results of each meridian, obtains the conversion matrix from each camera to the mechanical arm in the human meridian conditioning equipment through camera calibration, based on the conversion matrix Carry out coordinate transformation of each meridian identification result.
  • the coordinate conversion processing module includes a coordinate acquisition submodule, a depth information addition submodule, and a coordinate conversion submodule (none of which are shown in the figure).
  • the coordinate obtaining sub-module is configured to obtain the two-dimensional coordinates of each recognition target in the corresponding camera coordinate system in the meridian recognition result.
  • the depth information adding sub-module is configured to add corresponding depth information to the two-dimensional coordinates of each recognition target and then convert them into three-dimensional coordinates of the recognition target in the corresponding camera coordinate system.
  • the coordinate transformation sub-module is configured to obtain the three-dimensional coordinates of each recognition target in the robot arm coordinate system by multiplying the three-dimensional coordinates of each recognition target in the corresponding camera coordinate system by a point-product transformation matrix.
  • a human body meridian regulating device 200 is also provided.
  • the human body meridian regulating device 200 of the embodiment of the present application includes a processor (not shown in the figure), a camera, and a memory (not shown in the figure) for storing instructions executable by the processor.
  • the processor is configured to implement any one of the multi-camera-based human meridian recognition methods described above when executing executable instructions.
  • processors may be one or more.
  • an input device and an output device both not shown in the figure
  • the processor, the memory, the input device and the output device may be connected through a bus or in other ways, which are not specifically limited here.
  • the memory can be used to store software programs, computer-executable programs and various modules, such as the programs or modules corresponding to the intelligent conditioning method in the embodiment of the present application.
  • the processor executes various functional applications and data processing of the human body meridian conditioning device by running the software programs or modules stored in the memory.
  • the input device can be used to receive input numbers or signals.
  • the signal may be a key signal related to user setting and function control of the device/terminal/server.
  • the output device may include a display device such as a display screen.
  • the multiple cameras are respectively set at different positions of the human body meridian conditioning device.
  • the number of cameras can be set to four, and the four cameras are respectively set at the left side, the right side, the head side and the top position of the conditioning bed in the human body meridian conditioning equipment.
  • At least one camera 220 is set above one side of the conditioning bed 210 , that is, at least one camera 220 is set above the left side of the conditioning bed 210 .
  • the camera 220 located above the left side of the conditioning bed 210 is located obliquely above the left side of the human body, and the meridians on the left side and the inner side of the right leg of the human body can be obtained through the camera 220;
  • the camera 220 located above the left side of the conditioning bed 210 is located obliquely above the right side of the human body, and the meridians on the right side and inside of the left leg of the human body can be obtained through the camera 220.
  • At least one camera 220 is set above the other side of the conditioning bed 210 , that is, at least one camera 220 is set above the right side of the conditioning bed 210 .
  • the camera 220 located above the right side of the conditioning bed 210 is located obliquely above the right side of the human body, and the meridians on the right side and the inner side of the left leg of the human body can be obtained through the camera 220;
  • the camera 220 located above the right side of the conditioning bed 210 is located obliquely above the left side of the human body.
  • the camera 220 can obtain the meridians of the left side and the inner side of the right leg of the human body. At least one camera 220 is set above one end of the conditioning bed 210 , that is, at least one camera 220 is set above the head end of the conditioning bed 210 . The camera 220 arranged above the head end of the conditioning bed 210 is located obliquely above the head of the human body. When the human body is lying supine on the conditioning bed 210, the meridians of the top of the head, face and shoulders can be obtained through the camera 220; At least one camera 220 is disposed directly above the conditioning bed 210 , that is, at least one camera 220 is disposed directly above the conditioning bed 210 .
  • the camera 220 can obtain the meridians of the front side of the human body; It should be pointed out that the camera 220 is a depth camera or a 3D camera. On the whole, when the human body is lying on its back or prone, all the meridians of the human body can be recognized in all directions, which is beneficial for the massager 120 to locate each meridian and effectively improve the conditioning effect.
  • each of the camera 220 , the driver 128 and the robot arm 140 is electrically connected to the controller.
  • Each camera 220 can transmit acquired meridian information to the controller.
  • the controller controls the mechanical arm 140 to work to control the massager 120 to locate and find each meridian.
  • the controller controls the first motor 122 to work through the driver 128 .
  • a plurality of cameras 220 cooperate with each other, and adopt a multi-angle method to identify human meridians in different regions, which effectively improves the recognition accuracy, reduces the influence of image distortion, reduces the risk of occlusion, and improves the positioning accuracy of the massager 120, thereby improving conditioning effect.
  • a first support rod 230 and a second support rod 240 are also included.
  • At least one camera 220 is fixed on the side of the first fixing portion 231 of one of the first support rods 230 facing directly above the conditioning bed 210 , and the first fixing portion 231 of the other first support rod 230 is facing the front of the conditioning bed 210 .
  • At least one camera 220 is fixed on the upper side.
  • the first support rod 230 and the second support rod 240 are arc-shaped overall, and have better mechanical strength.
  • the first fixing part 231 is a "convex" structure, and a lighting device can also be installed on the first fixing part 231 to improve the conditioning environment.
  • There is one second support rod 240 which is arranged vertically and is arranged at one end of the conditioning bed 210 , and its upper part is bent towards the top of the conditioning bed 210 to form a second fixing portion 241 . It should be pointed out that the second support rod 240 is arranged at the head end of the conditioning bed 210 , that is, the second support rod 240 is arranged at the end of the bed board 112 with a larger cross section.
  • At least one camera 220 is fixed on the bottom surface of the second fixing part 241 of the second support rod 240 , and at least one camera 220 is fixed on the side facing the conditioning bed 210 in the middle.
  • the orthographic projection of the second fixing part 241 from top to bottom is a "convex" structure, and an illuminating device is installed on the bottom surface of the second fixing part 241 to improve the conditioning environment.
  • the lighting device includes a plurality of LED lamps, and the plurality of LEDs are arranged in an array.
  • the angle between the axial direction of the detection port of the camera 220 fixed on the middle part of the second support rod 240 and the plane where the bed board 112 is located is 30-60 degrees
  • the axis of the detection port of the camera 220 fixed on the second fixing part 241 The included angle between the direction and the plane of the bed board 112 is 85-95 degrees
  • the included angle between the axial direction of the detection port of the camera 220 fixed on the first fixing part 231 and the plane of the bed board 112 is 30-60 degrees.

Landscapes

  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

本申请涉及一种基于多摄像头的人体经络识别方法和装置及人体经络调理设备，其中方法包括：获取配置在人体经络调理设备上的各摄像头采集到的人体图像；将各人体图像输入至预先构建的经络识别神经网络中，由经络识别神经网络对各人体图像进行识别，得到各人体图像的经络识别结果；将各经络识别结果进行整合处理得到最终的人体经络识别结果。其相较于相关技术中从单一角度采集到的人体图像进行经络识别所得到的经络识别结果来说，有效增加了人体图像多角度的识别，这也就有效避免了单一角度采集人体图像导致经络识别遗漏的情况，最终有效保证了人体经络识别结果的准确性和完整性。

Description

基于多摄像头的人体经络识别方法和装置及人体经络调理设备 技术领域
本申请涉及智能医疗调理机械设备技术领域，尤其涉及一种基于多摄像头的人体经络识别方法和装置及人体经络调理设备。
背景技术
深度学习是机器学习中的一个新的领域，其动机在于建立、模拟人脑进行分析学习的神经网络，通过所建立的神经网络来模仿人脑的机制进行数据的解析，如：图像、声音和文本等。通过采用深度学习的方式来实现人体经络的识别，对于提高人体经络识取穴的效率具有很好的效果。但是，在相关技术中，通过所建立的神经网络进行人体经络的识别时，由于所采集到的图像是通过单个相机采集到的，而单个相机不能获取人体的全部影像，这就使得在采用深度学习进行人体经络识别时得到的识别结果在准确度和完整性上不够高。
发明内容
有鉴于此，本申请提出了一种基于多摄像头的人体经络识别方法，可以有效提高人体经络识别结果的准确度和完整性。
根据本申请的一方面，提供了一种基于多摄像头的人体经络识别方法，包括：
获取配置在人体经络调理设备上的各摄像头采集到的人体图像；
将各所述人体图像输入至预先构建的经络识别神经网络中，由所述经络识别神经网络对各所述人体图像进行识别，得到各所述人体图像的经络识别结果；
将各所述经络识别结果进行整合处理得到最终的人体经络识别结果。
在一种可能的实现方式中，将各所述经络识别结果进行整合处理得到最 终的人体经络识别结果时，包括将各所述经络识别结果进行坐标转换的操作。
在一种可能的实现方式中，将各所述经络识别结果进行坐标转换时，通过摄像头标定，获取各所述摄像头到所述人体经络调理设备中机械臂的转换矩阵，基于所述转换矩阵进行各所述经络识别结果的坐标转换。
在一种可能的实现方式中，基于各所述摄像头到所述人体经络调理设备中机械臂的转换矩阵，进行各所述经络识别结果的坐标转换，包括：
获取所述经络识别结果中的各识别目标在对应的摄像头坐标系下的二维坐标；
对各所述识别目标的二维坐标添加对应的深度信息后转换为所述识别目标在对应的摄像头坐标系下的三维坐标；
将各所述识别目标在对应的摄像头坐标系下的三维坐标通过点乘所述转换矩阵，得到各所述识别目标在机械臂坐标系下的三维坐标。
根据本申请的另一方面，还提供了一种基于多摄像头的人体经络识别装置，包括：图像获取模块、图像识别模块和结果整合模块；
所述图像获取模块，被配置为获取配置在人体经络调理设备上的各摄像头采集到的人体图像；
所述图像识别模块，被配置为将各所述人体图像输入至预先构建的经络识别神经网络中，由所述经络识别神经网络对各所述人体图像进行识别，得到各所述人体图像的经络识别结果；
所述结果整合模块，被配置为将各所述经络识别结果进行整合处理得到最终的人体经络识别结果。
在一种可能的实现方式中，还包括坐标转换处理模块；
所述坐标转换处理模块，被配置为在所述结果整合模块将各所述经络识别结果进行整合处理得到最终的人体经络识别结果时，将各所述经络识别结果进行坐标转换处理。
在一种可能的实现方式中，所述坐标转换处理模块，在别配置为将各所述经络识别结果进行坐标转换时，通过摄像头标定，获取各所述摄像头到所 述人体经络调理设备中机械臂的转换矩阵，基于所述转换矩阵进行各所述经络识别结果的坐标转换。
在一种可能的实现方式中，所述坐标转换处理模块包括坐标获取子模块、深度信息添加子模块和坐标转换子模块；
所述坐标获取子模块，被配置为获取所述经络识别结果中的各识别目标在对应的摄像头坐标系下的二维坐标；
所述深度信息添加子模块，被配置为对各所述识别目标的二维坐标添加对应的深度信息后转换为所述识别目标在对应的摄像头坐标系下的三维坐标；
所述坐标转换子模块，被配置为将各所述识别目标在对应的摄像头坐标系下的三维坐标通过点乘所述转换矩阵，得到各所述识别目标在机械臂坐标系下的三维坐标。
根据本申请的另一方面，还提供了一种人体经络调理设备，包括：
处理器；
用于存储处理器可执行指令的存储器；
其中，所述处理器被配置为执行所述可执行指令时实现前面任一所述的方法；
还包括：
摄像头；
所述摄像头为多个，且多个所述摄像头分别设置在所述人体经络调理设备的不同位置处。
在一种可能的实现方式中，所述摄像头的个数为四个；
四个所述摄像头分别设置在所述人体经络调理设备中调理床的左侧、右侧前侧和顶部位置处。
本申请实施例的方法，在进行人体经络识别时，通过在人体经络调理设备上配置多个摄像头，由多个摄像头分别从不同的角度进行人体图像的采集，然后再基于从不同角度采集到的人体图像进行人体经络的识别，进而再将由 不同角度的人体图像识别出的经络识别结果进行整合得到最终的人体经络识别结果，这就使得所得到的人体经络识别结果更加完整。相较于相关技术中从单一角度采集到的人体图像进行经络识别所得到的经络识别结果来说，有效增加了人体图像多角度的识别，这也就有效避免了单一角度采集人体图像导致经络识别遗漏的情况，最终有效保证了人体经络识别结果的准确性和完整性。
根据下面参考附图对示例性实施例的详细说明，本申请的其它特征及方面将变得清楚。
附图说明
包含在说明书中并且构成说明书的一部分的附图与说明书一起示出了本申请的示例性实施例、特征和方面，并且用于解释本申请的原理。
图1示出本申请实施例的基于多摄像头的人体经络识别方法的流程图；
图2示出本申请实施例的基于多摄像头的人体经络识别方法中，各摄像头在人体经络调理设备中的安装结构示意图；
图3示出本申请实施例的基于多摄像头的人体经络识别装置的结构框图。
具体实施方式
以下将参考附图详细说明本申请的各种示例性实施例、特征和方面。附图中相同的附图标记表示功能相同或相似的元件。尽管在附图中示出了实施例的各种方面，但是除非特别指出，不必按比例绘制附图。
在这里专用的词“示例性”意为“用作例子、实施例或说明性”。这里作为“示例性”所说明的任何实施例不必解释为优于或好于其它实施例。
另外，为了更好的说明本申请，在下文的具体实施方式中给出了众多的具体细节。本领域技术人员应当理解，没有某些具体细节，本申请同样可以实施。在一些实例中，对于本领域技术人员熟知的方法、手段、元件和电路未作详细描述，以便于凸显本申请的主旨。
图1示出根据本申请一实施例的基于多摄像头的人体经络识别方法的流程图。如图1所示，该方法包括：步骤S100，获取配置在人体经络调理设备上的各摄像头采集到的人体图像。此处，需要说明的是，在本申请实施例的方法中，人体经络调理设备可以为用于实现人体经络调理的仪器，如图2所示。本申请实施例的方法具体可以应用到进行人体经络调理的仪器中。在人体经络调理设备上安装多个摄像头，每个摄像头分别安装在人体经络调理设备的不同位置处，从而使得各摄像头能够从不同方位进行人体图像的采集。
在通过配置在人体经络调理设备上的各摄像头采集到不同方位的人体图像后，即可执行步骤S200，将各人体图像输入至预先构建的经络识别神经网络中，由经络识别神经网络对各人体图像进行识别，得到各人体图像的经络识别结果。进而再通过步骤S300，将各经络识别结果进行整合处理得到最终的人体经络识别结果。
由此，本申请实施例的方法，在进行人体经络识别时，通过在人体经络调理设备上配置多个摄像头，由多个摄像头分别从不同的角度进行人体图像的采集，然后再基于从不同角度采集到的人体图像进行人体经络的识别，进而再将由不同角度的人体图像识别出的经络识别结果进行整合得到最终的人体经络识别结果，这就使得所得到的人体经络识别结果更加完整。相较于相关技术中从单一角度采集到的人体图像进行经络识别所得到的经络识别结果来说，有效增加了人体图像多角度的识别，这也就有效避免了单一角度采集人体图像导致经络识别遗漏的情况，最终有效保证了人体经络识别结果的准确性和完整性。
应当指出的是，在本申请实施例的方法中，在将各人体图像输入至预先构建的经络识别神经网络中，由经络识别神经网络对各人体图像进行识别得到各人体图像的经络识别结果时，所构建的经络识别神经网络可以采用本领域常规的目标识别网络模型，此处不再进行赘述。
同时，还应当指出的是，在使用预先构建的经络识别神经网络进行人体经络的识别时，经络识别神经网络需要先进行训练。对于经络识别神经网络 的训练，则需要构建相应的训练样本集，并对训练样本集中的各训练样本进行标注。此处，需要指出的是，在本申请实施例的方法中，所构建的训练样本集中的训练样本为通过实际拍摄和从网络收集到的各种人体图像，并结合人体穴位图在各人体图像中标注出相应的人体穴位的图像数据。
进一步地，在通过将各摄像头采集到的不同角度的人体图像输入至经络识别神经网络中，由经络识别神经网络识别出各人体图像中的穴位点之后，即可将各人体图像的经络识别结果进行整合后作为最终的人体经络识别结果。在一种可能的实现方式中，对各人体图像的经络识别结果的整合处理过程中包括有将各人体图像的经络识别结果进行坐标转换的操作。
这是由于各人体图像的经络识别结果均是在各摄像头坐标系下的二维坐标，不同的摄像头的坐标系有所不同。同时，各摄像头的坐标系与人体经络调理设备中机械臂坐标系也有所不同。由此，为了便于进行各人体图像的经络识别结果的整合，需要将各人体图像的经络识别结果的坐标系进行统一。在一种可能的实现方式中，可以通过将各人体图像的经络识别结果的坐标全部转换为人体经络调理设备中机械臂坐标系下的坐标，从而在机械臂根据识别出来的各穴位点进行按摩或调理时能够顺利准确的识别出各穴位点的位置。
具体的，将各经络识别结果进行坐标转换时，可以通过摄像头标定，获取各摄像头到人体经络调理设备中机械臂的转换矩阵，基于转换矩阵进行各经络识别结果的坐标转换。即，通过对各摄像头进行坐标系标定，以及对人体经络调理设备中机械臂进行坐标系标定，进而再根据所标定的各摄像头的坐标系以及机械臂的坐标系得到各摄像头与机械臂的转换矩阵。
其中，图像采集设备—机械臂的坐标转换矩阵可以通过分别对图像采集设备和机械臂进行坐标标定，然后再根据标定的图像采集设备坐标和机械臂坐标进行对应计算的方式来实现。在一种可能的实现方式中，对图像采集设备和机械臂的坐标标定则可以通过棋盘格的方式进行。
具体的，通过urx和opencv两个开源库，使用棋盘格对图像采集设备和机 械臂进行标定，图像采集设备获取棋盘格的中心在图像采集设备坐标下的三维坐标，机械臂获得棋盘格中心在机械臂坐标系下的三维坐标，机械臂按照一定的3*3网格进行行走，得到9组图像采集设备坐标系下的棋盘格中心坐标和机械臂坐标系下的棋盘格中心坐标，通过计算即可得出图像采集设备坐标到机械臂坐标的一个转换矩阵(即，图像采集设备—机械臂的坐标转换矩阵)
其中，应当指出的是，不同的摄像头对应存在与机械臂的一转换矩阵。在进行各人体图像中所识别出的穴位点的坐标转换时，根据采集各人体图像的摄像头与机械臂的转换矩阵进行。
更加具体的，基于各摄像头到人体经络调理设备中机械臂的转换矩阵，进行各经络识别结果的坐标转换时，可以通过以下方式来实现。
首先，获取经络识别结果中的各识别目标在对应的摄像头坐标系下的二维坐标。然后，对各识别目标的二维坐标添加对应的深度信息后转换为识别目标在对应的摄像头坐标系下的三维坐标。进而，再将各识别目标在对应的摄像头坐标系下的三维坐标通过点乘转换矩阵，得到各识别目标在机械臂坐标系下的三维坐标。
此处，需要说明的是，在获取经络识别结果中各识别目标在对应的摄像头坐标系下的二维坐标，可以直接根据由经络识别神经网络所识别出的穴位点在人体图像上的位置来实现。即，各识别目标指的即为由人体图像中识别出来的人体穴位点。各识别目标在对应的摄像头坐标系下的二维坐标，则可以根据由人体图像中所识别出来的各人体穴位点在人体图像中的位置来得到。此处，本领域技术人员可以理解的是，根据各人体穴位点在人体图像中的位置得到各人体穴位点在对应的摄像头坐标系下的二维坐标可以采用本领域常规技术手段来实现，如：直接根据各摄像头的内部参数和外部参数之间的映射关系来实现，此处不再进行赘述。
进一步地，对各识别目标的二维坐标添加对应的深度信息后转换为识别目标在对应的摄像头坐标系下的三维坐标，则可以通过Azure Kinect SDK来实现对应的深度信息的添加，此处也不再进行赘述。
在通过上述任一种方式将经络识别结果中各识别目标由二维坐标转换为三维坐标之后，此时各识别目标的坐标数据还是处于所对应的摄像头坐标系下的坐标数据。因此，还需要将各识别目标的坐标数据转换至机械臂坐标系下才能使得机械臂顺利准确的识别出各人体穴位点的位置。
根据前面所述，在本申请实施例的方法中，可以通过点乘转换矩阵的方式来实现各识别目标在摄像头坐标系至机械臂坐标系的转换。此处，应当指出的是，对于不同的人体图像的经络识别结果，需要采用不同的摄像头与机械臂之间的坐标转换矩阵。
举例来说，在人体经络调理设备中设置有四个摄像头，各摄像头分别位于人体经络调理设备的不同位置处，分别为摄像头A，摄像头B，摄像头C和摄像头D。其中，摄像头A采集到人体图像PictureA，摄像头B采集到人体图像PictureB，摄像头C采集到人体图像PictureC，摄像头D采集到人体图像PictureD。
将上述人体图像PictureA、人体图像PictureB、人体图像PictureC和人体图像PictureD依次输入至经络识别神经网络中，由经络识别神经网络分别对上述各人体图像进行经络识别，由各人体图像中识别出相应的人体穴位点。此时，所识别出的各人体穴位点的坐标数据均是处于其所对应的摄像头坐标系下的。因此，需要将各人体图像中的各人体穴位点进行坐标转换，均将各人体图像中的各人体穴位点的坐标转换至机械臂坐标系下。
在进行坐标转换时，对于由人体图像A中识别出的各人体穴位点，需要基于摄像头A与机械臂之间的转换矩阵进行转换。对于由人体图像B中识别出的各人体穴位点，则需要基于摄像头B与机械臂之间的转换矩阵进行转换。同理，对于由人体图像C中识别出的各人体穴位点，则需要基于摄像头C与机械臂之间的转换矩阵进行转换。对于由人体图像D中识别出的各人体穴位点，需要基于摄像头C与机械臂之间的转换矩阵进行转换。
此处，需要说明的是，对于各摄像头的坐标标定以及机械臂的坐标标定均可以采用本领域常规的坐标标定技术手段来实现，同时基于所标定的各摄 像头坐标系与机械臂坐标系所得到的各摄像头与机械臂之间的转换矩阵也均可以采用本领域的常规技术手段来实现，此处均不再进行赘述。
相应的，基于前面任一所述的基于多摄像头的人体经络识别方法，本申请还提供了一种基于所摄像头的人体经络识别装置。由于本申请提供的基于多摄像头的人体经络识别装置的工作原理，与本申请的基于多摄像头的人体经络识别方法的原理相同或相似，因此重复之处不再赘述。
参阅图3，本申请提供的基于多摄像头的人体经络识别装置100，包括图像获取模块110、图像识别模块120和结果整合模块130。其中，图像获取模块110，被配置为获取配置在人体经络调理设备上的各摄像头采集到的人体图像。图像识别模块120，被配置为将各人体图像输入至预先构建的经络识别神经网络中，由经络识别神经网络对各人体图像进行识别，得到各人体图像的经络识别结果。结果整合模块130，被配置为将各经络识别结果进行整合处理得到最终的人体经络识别结果。
在一种可能的实现方式中，还包括坐标转换处理模块(图中未示出)。其中，坐标转换处理模块，被配置为在结果整合模块130将各经络识别结果进行整合处理得到最终的人体经络识别结果时，将各经络识别结果进行坐标转换处理。
在一种可能的实现方式中，坐标转换处理模块，在别配置为将各经络识别结果进行坐标转换时，通过摄像头标定，获取各摄像头到人体经络调理设备中机械臂的转换矩阵，基于转换矩阵进行各经络识别结果的坐标转换。
在一种可能的实现方式中，坐标转换处理模块包括坐标获取子模块、深度信息添加子模块和坐标转换子模块(图中均未示出)。其中，坐标获取子模块，被配置为获取经络识别结果中的各识别目标在对应的摄像头坐标系下的二维坐标。深度信息添加子模块，被配置为对各识别目标的二维坐标添加对应的深度信息后转换为识别目标在对应的摄像头坐标系下的三维坐标。坐标转换子模块，被配置为将各识别目标在对应的摄像头坐标系下的三维坐标通过点乘转换矩阵，得到各识别目标在机械臂坐标系下的三维坐标。
更进一步地，根据本申请的另一方面，还提供了一种人体经络调理设备200。参阅图2，本申请实施例的人体经络调理设备200包括处理器(图中未示出)、摄像头以及用于存储处理器可执行指令的存储器(图中未示出)。其中，处理器被配置为执行可执行指令时实现前面任一所述的基于多摄像头的人体经络识别方法。
此处，应当指出的是，处理器的个数可以为一个或多个。同时，在本申请实施例的人体经络调理设备中，还可以包括输入装置和输出装置(图中均未示出)。其中，处理器、存储器、输入装置和输出装置之间可以通过总线连接，也可以通过其他方式连接，此处不进行具体限定。
存储器作为一种计算机可读存储介质，可用于存储软件程序、计算机可执行程序和各种模块，如：本申请实施例的智能调理方法所对应的程序或模块。处理器通过运行存储在存储器中的软件程序或模块，从而执行人体经络调理设备的各种功能应用及数据处理。
输入装置可用于接收输入的数字或信号。其中，信号可以为产生与设备/终端/服务器的用户设置以及功能控制有关的键信号。输出装置可以包括显示屏等显示设备。
摄像头为多个，且多个摄像头分别设置在人体经络调理设备的不同位置处。在一种可能的实现方式中，摄像头的个数可以设置为四个，四个摄像头分别设置在人体经络调理设备中调理床的左侧、右侧、头部侧和顶部位置处。
具体的，参见图2，至少一个摄像头220设于调理床210的一侧的上方，即，至少一个摄像头220设于调理床210的左侧的上方。当人体仰卧在调理床210上时，设于调理床210的左侧的上方的摄像头220位于人体的左侧面的斜上方，通过该摄像头220能够获取人体左侧及右腿内侧的经脉；当人体俯卧在调理床210上时，设于调理床210的左侧的上方的摄像头220位于人体的右侧面的斜上方，通过该摄像头220能够获取人体右侧及左腿内侧的经脉。至少一个摄像头220设于调理床210的另一侧的上方，即，至少一个摄像头220设于调理床210的右侧的上方。当人体仰卧在调理床210上时，设于调理床210 的右侧的上方的摄像头220位于人体的右侧面的斜上方，通过该摄像头220能够获取人体右侧及左腿内侧的经脉；当人体俯卧在调理床210上时，设于调理床210的右侧的上方的摄像头220位于人体的左侧面的斜上方，通过该摄像头220能够获取人体左侧及右腿内侧的经脉。至少一个摄像头220设于调理床210的一端的上方，即，至少一个摄像头220设于调理床210的床头端的上方。设于调理床210的床头端的上方的摄像头220位于人体头部的斜上方。当人体仰卧在调理床210上时，通过该摄像头220能够获取头顶、面部和肩部的经脉；当人体俯卧在调理床210上时，通过该摄像头220能够获取头顶、肩背的经脉。至少一个摄像头220设于调理床210的正上方，即，至少一个摄像头220设于调理床210的正上方。当人体仰卧在调理床210上时，通过该摄像头220能够获取人体的前侧面的经脉；当人体俯卧在调理床210上时，通过该摄像头220能够获取人体的后侧面的经脉。需要指出的是，摄像头220为深度摄像头或3D相机。整体上，在人体进行仰卧或俯卧时，均能够对人体经脉进行全方位的识别，有利于按摩器120定位寻找每一种经脉，有效地改善了调理效果。此处，需要说明的是，每个摄像头220、驱动器128以及机械臂140均电连接于控制器。每个摄像头220能够将获取的经脉信息传输给控制器。控制器控制机械臂140进行工作，以控制按摩器120进行定位寻找每一种经脉。控制器通过驱动器128控制第一电机122进行工作。而且，多个摄像头220相互配合工作，采用多角度方式对人体经脉进行分区域识别，有效地提高了识别精度，减少图像畸变的影响，降低遮挡风险，改善按摩器120的定位准度，从而改善了调理效果。
在本申请的人体经络调理设备200的一具体实施例中，还包括第一支撑杆230和第二支撑杆240。第一支撑杆230为两个，均竖直设置，分别设于调理床210的相对两侧，上部均朝向调理床210的正上方弯折以形成第一固定部231。其中一个第一支撑杆230的第一固定部231朝向调理床210的正上方的一侧面上固定有至少一个摄像头220，另一个第一支撑杆230的第一固定部231朝向调理床210的正上方的一侧面上固定有至少一个摄像头220。第一支撑杆 230和第二支撑杆240整体为弧形结构，具有较好地机械强度。第一固定部231为“凸”型结构，在第一固定部231上还能够安装照明装置，以改善调理环境。第二支撑杆240为一个，竖直设置，设于调理床210的一端，上部朝向调理床210的正上方弯折以形成第二固定部241。需要指出的是，第二支撑杆240设置在调理床210的床头端，即第二支撑杆240设于床板112横截面较大的一端。第二支撑杆240的第二固定部241的底面上固定有至少一个摄像头220，中部朝向调理床210的一侧面上固定有至少一个摄像头220。第二固定部241从上向下的正投影为“凸”型结构，在第二固定部241的底面上安装照明装置，以改善调理环境。照明装置包括多个LED灯，多个LED呈阵列式排布。另外，固定于第二支撑杆240的中部的摄像头220的探测口的轴向与床板112所在平面的夹角为30-60度，固定于第二固定部241上的摄像头220的探测口的轴向与床板112所在平面的夹角为85-95度，固定于第一固定部231上的摄像头220的探测口的轴向与床板112所在平面的夹角为30-60度。如此，有利于改善摄像头220对经脉进行获取时的获取效果，降低遮挡风险，改善按摩器120的定位准度，从而改善了调理效果。
以上已经描述了本申请的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术的技术改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。

Claims (10)

  1. 一种基于多摄像头的人体经络识别方法，其特征在于，包括：
    获取配置在人体经络调理设备上的各摄像头采集到的人体图像；
    将各所述人体图像输入至预先构建的经络识别神经网络中，由所述经络识别神经网络对各所述人体图像进行识别，得到各所述人体图像的经络识别结果；
    将各所述经络识别结果进行整合处理得到最终的人体经络识别结果。
  2. 根据权利要求1所述的方法，其特征在于，将各所述经络识别结果进行整合处理得到最终的人体经络识别结果时，包括将各所述经络识别结果进行坐标转换的操作。
  3. 根据权利要求2所述的方法，其特征在于，将各所述经络识别结果进行坐标转换时，通过摄像头标定，获取各所述摄像头到所述人体经络调理设备中机械臂的转换矩阵，基于所述转换矩阵进行各所述经络识别结果的坐标转换。
  4. 根据权利要求3所述的方法，其特征在于，基于各所述摄像头到所述人体经络调理设备中机械臂的转换矩阵，进行各所述经络识别结果的坐标转换，包括：
    获取所述经络识别结果中的各识别目标在对应的摄像头坐标系下的二维坐标；
    对各所述识别目标的二维坐标添加对应的深度信息后转换为所述识别目标在对应的摄像头坐标系下的三维坐标；
    将各所述识别目标在对应的摄像头坐标系下的三维坐标通过点乘所述转换矩阵，得到各所述识别目标在机械臂坐标系下的三维坐标。
  5. 一种基于多摄像头的人体经络识别装置，其特征在于，包括：图像获取模块、图像识别模块和结果整合模块；
    所述图像获取模块，被配置为获取配置在人体经络调理设备上的各摄像头采集到的人体图像；
    所述图像识别模块，被配置为将各所述人体图像输入至预先构建的经络 识别神经网络中，由所述经络识别神经网络对各所述人体图像进行识别，得到各所述人体图像的经络识别结果；
    所述结果整合模块，被配置为将各所述经络识别结果进行整合处理得到最终的人体经络识别结果。
  6. 根据权利要求5所述的装置，其特征在于，还包括坐标转换处理模块；
    所述坐标转换处理模块，被配置为在所述结果整合模块将各所述经络识别结果进行整合处理得到最终的人体经络识别结果时，将各所述经络识别结果进行坐标转换处理。
  7. 根据权利要求6所述的装置，其特征在于，所述坐标转换处理模块，在别配置为将各所述经络识别结果进行坐标转换时，通过摄像头标定，获取各所述摄像头到所述人体经络调理设备中机械臂的转换矩阵，基于所述转换矩阵进行各所述经络识别结果的坐标转换。
  8. 根据权利要求7所述的装置，其特征在于，所述坐标转换处理模块包括坐标获取子模块、深度信息添加子模块和坐标转换子模块；
    所述坐标获取子模块，被配置为获取所述经络识别结果中的各识别目标在对应的摄像头坐标系下的二维坐标；
    所述深度信息添加子模块，被配置为对各所述识别目标的二维坐标添加对应的深度信息后转换为所述识别目标在对应的摄像头坐标系下的三维坐标；
    所述坐标转换子模块，被配置为将各所述识别目标在对应的摄像头坐标系下的三维坐标通过点乘所述转换矩阵，得到各所述识别目标在机械臂坐标系下的三维坐标。
  9. 一种人体经络调理设备，其特征在于，包括：
    处理器；
    用于存储处理器可执行指令的存储器；
    其中，所述处理器被配置为执行所述可执行指令时实现权利要求1至4中任意一项所述的方法；
    还包括：
    摄像头；
    所述摄像头为多个，且多个所述摄像头分别设置在所述人体经络调理设备的不同位置处。
  10. 根据权利要求8所述的设备，其特征在于，所述摄像头的个数为四个；
    四个所述摄像头分别设置在所述人体经络调理设备中调理床的左侧、右侧前侧和顶部位置处。
PCT/CN2022/113488 2021-08-30 2022-08-19 基于多摄像头的人体经络识别方法和装置及人体经络调理设备 WO2023030036A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111004279.7A CN113807207A (zh) 2021-08-30 2021-08-30 基于多摄像头的人体经络识别方法和装置及人体经络调理设备
CN202111004279.7 2021-08-30

Publications (1)

Publication Number Publication Date
WO2023030036A1 true WO2023030036A1 (zh) 2023-03-09

Family

ID=78942053

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/113488 WO2023030036A1 (zh) 2021-08-30 2022-08-19 基于多摄像头的人体经络识别方法和装置及人体经络调理设备

Country Status (2)

Country Link
CN (1) CN113807207A (zh)
WO (1) WO2023030036A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113807207A (zh) * 2021-08-30 2021-12-17 中科尚易健康科技(北京)有限公司 基于多摄像头的人体经络识别方法和装置及人体经络调理设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110188699A (zh) * 2019-05-31 2019-08-30 安徽柏络智能科技有限公司 一种双目摄像头的人脸识别方法及系统
CN111414812A (zh) * 2020-03-03 2020-07-14 平安科技（深圳）有限公司 人体属性识别方法、系统、计算机设备及存储介质
CN112184705A (zh) * 2020-10-28 2021-01-05 成都智数医联科技有限公司 一种基于计算机视觉技术的人体穴位识别、定位及应用系统
CN113807207A (zh) * 2021-08-30 2021-12-17 中科尚易健康科技(北京)有限公司 基于多摄像头的人体经络识别方法和装置及人体经络调理设备

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5832106A (en) * 1996-05-22 1998-11-03 Electronics And Telecommunications Research Institute Method for camera calibration of range imaging system by use of neural network
CN107137225B (zh) * 2017-06-22 2023-09-01 山东中医药大学 个性化头面部穴位识别模型建立、定位的方法和系统
CN108765546A (zh) * 2018-04-18 2018-11-06 北京奇虎科技有限公司 穴位检测方法及装置
CN110232318A (zh) * 2019-05-06 2019-09-13 平安科技（深圳）有限公司 穴位识别方法、装置、电子设备及存储介质
CN110464633A (zh) * 2019-06-17 2019-11-19 深圳壹账通智能科技有限公司 穴位识别方法、装置、设备及存储介质
CN210405493U (zh) * 2019-06-28 2020-04-24 西安工程大学 一种矩阵式多目摄像头系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110188699A (zh) * 2019-05-31 2019-08-30 安徽柏络智能科技有限公司 一种双目摄像头的人脸识别方法及系统
CN111414812A (zh) * 2020-03-03 2020-07-14 平安科技（深圳）有限公司 人体属性识别方法、系统、计算机设备及存储介质
CN112184705A (zh) * 2020-10-28 2021-01-05 成都智数医联科技有限公司 一种基于计算机视觉技术的人体穴位识别、定位及应用系统
CN113807207A (zh) * 2021-08-30 2021-12-17 中科尚易健康科技(北京)有限公司 基于多摄像头的人体经络识别方法和装置及人体经络调理设备

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
MA, ZHEWEN; YU, HAOGUANG: "Visual Positioning Based Acupoint Tracking System for Massage Robots", ROBOT TECHNIQUE AND APPLICATION, no. 6, 30 November 2010 (2010-11-30), pages 33 - 35, XP009543948, ISSN: 1004-6437 *

Also Published As

Publication number Publication date
CN113807207A (zh) 2021-12-17

Similar Documents

Publication Publication Date Title
US10875187B2 (en) Robotic arm camera system and method
US10893850B2 (en) Methods and apparatuses for guiding collection of ultrasound data using motion and/or orientation data
JP5688825B2 (ja) カメラシステムの座標系の較正および変換
JP6869302B2 (ja) 検査位置の調整方法、調整装置、超音波プローブおよび端末
CN106920228B (zh) 脑图谱与脑图像的配准方法及装置
WO2023030036A1 (zh) 基于多摄像头的人体经络识别方法和装置及人体经络调理设备
WO2021078065A1 (zh) 乳房三维点云重建方法、装置、存储介质及计算机设备
US20100249595A1 (en) System and method for automatic calibration of tracked ultrasound
US20150071491A1 (en) Method And Device For Optically Determining A Position And/Or Orientation Of An Object In Space
JP7241933B2 (ja) 画像レジストレーション方法及びそれに関係するモデルトレーニング方法、デバイス、装置
CN112669371A (zh) 一种艾灸机器人控制装置、系统、设备及存储介质
WO2020033380A1 (en) Methods and apparatuses for determining and displaying locations on images of body portions based on ultrasound data
Suligoj et al. RobUSt–an autonomous robotic ultrasound system for medical imaging
CN111544037B (zh) 一种基于双目视觉的超声定位方法及系统
CN106512239B (zh) 一种穿戴式医学超声治疗装置及其声发射阵列定位与声束聚焦方法
EP3843636A1 (en) Methods and apparatuses for collection of ultrasound data
CN109760107A (zh) 一种基于单目视觉的机器人定位精度评价方法
CN115661862A (zh) 一种基于压力视觉卷积模型的坐姿样本集自动标注方法
US11250592B2 (en) Information processing apparatus
CN107527316B (zh) 二维超声影像序列上的任意点构建点云数据的方法及系统
CN107993227B (zh) 一种获取3d腹腔镜手眼矩阵的方法和装置
WO2023112728A1 (ja) 医療支援方法、医療支援ロボット装置、およびコンピュータプログラム
CN111563880A (zh) 一种基于目标检测与聚类的横突棘突检测定位方法
Jin et al. Human-robot interaction for assisted object grasping by a wearable robotic object manipulation aid for the blind
CN113813168A (zh) 经络穴位随动识别方法和装置、设备及存储介质

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE