CN114495169A - Training data processing method, device and equipment for human body posture recognition - Google Patents
Training data processing method, device and equipment for human body posture recognition Download PDFInfo
- Publication number
- CN114495169A CN114495169A CN202210096397.3A CN202210096397A CN114495169A CN 114495169 A CN114495169 A CN 114495169A CN 202210096397 A CN202210096397 A CN 202210096397A CN 114495169 A CN114495169 A CN 114495169A
- Authority
- CN
- China
- Prior art keywords
- training
- human body
- personnel
- body posture
- action
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012549 training Methods 0.000 title claims abstract description 338
- 238000003672 processing method Methods 0.000 title claims abstract description 30
- 230000009471 action Effects 0.000 claims abstract description 79
- 238000000034 method Methods 0.000 claims abstract description 23
- 230000000007 visual effect Effects 0.000 claims abstract description 10
- 238000012545 processing Methods 0.000 claims description 29
- 238000013527 convolutional neural network Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000033001 locomotion Effects 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 8
- 230000009466 transformation Effects 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 abstract description 6
- 210000000988 bone and bone Anatomy 0.000 description 7
- 238000004891 communication Methods 0.000 description 7
- 238000007726 management method Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005553 drilling Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention relates to a training data processing method, a device and equipment for human body posture recognition, wherein the method comprises the following steps: acquiring training video images of different visual angles of training personnel in real time through camera equipment; adopting human body posture estimation to recognize the training video image, and recognizing to obtain posture action data of the training personnel; and comparing the posture action data of the training personnel with the training standard action of the model base to obtain the training result of the training personnel. According to the method, the training video images of the training personnel are obtained through the camera equipment, the training video images are analyzed through human body posture estimation, posture action data which can be compared with training standard actions of a model base are obtained, training results of the training personnel are obtained through comparison and analysis, all standard and unqualified training action data of the training personnel can be accurately obtained, a training report is generated, a training plan can be conveniently made for the training personnel in a targeted mode, and training data are provided for scientific and effective training.
Description
Technical Field
The invention relates to the technical field of training, in particular to a training data processing method, a training data processing device and training data processing equipment for human body posture recognition.
Background
The training business is used as a normal work task in various fields, such as military training of troops, emergency rescue training and the like. With the progress of the times and the development of technologies, the training management is assisted by the modern high-technology means, and the improvement of the training quality and efficiency becomes a necessary trend.
With the scientific progress, the digital informatization is widely applied, an emergency training intelligent management platform is constructed, the interconnection and intercommunication of rescue training management information are realized, a scientific basis is provided for simulation training and rescue actual combat, and the method becomes an important means for accelerating the informatization development and construction of intelligent emergency rescue training. At present, the technical content of emergency rescue teams is not high, the emergency rescue information guarantee technology is not enough, more importance is placed on the construction of a corresponding emergency rescue system, the daily emergency training quality is improved, and the smooth proceeding of emergency rescue work is ensured. Advanced information means are fully utilized by all levels of emergency management units, information rescue training modes of rescue teams are actively explored, reference schemes for training and rescue are timely, accurately and effectively provided for the rescue teams, various systems integrated by various devices based on the Internet of things are preliminarily constructed, perfect intelligent information training and command management platforms are realized to different degrees, and daily training and rescue actual combat efficiency is improved.
The current various emergency training systems better solve the problems of emergency plan library management, a drilling plan and plan design module, drilling process tracking and cooperation, drilling effect evaluation analysis and report and the like basically in the aspects of data management and emergency command cooperation, and utilize an informatization means to enable emergency training. However, the training action of the training personnel cannot be accurately acquired by the training system, the data acquired by the training system is acquired by wearing equipment by the training personnel, and the equipment is easy to loosen or fall off in the training process, so that the training data acquired by the training system is not accurate enough.
Disclosure of Invention
The embodiment of the invention provides a training data processing method, a training data processing device and training data processing equipment for human body posture recognition, which are used for solving the technical problem that the existing training system cannot accurately acquire the actions and the training data of a training person.
In order to achieve the above object, the embodiments of the present invention provide the following technical solutions:
a training data processing method for human body posture recognition comprises the following steps:
acquiring training video images of different visual angles of training personnel in real time through camera equipment;
adopting human body posture estimation to recognize the training video image, and recognizing to obtain posture action data of the training personnel;
and comparing the posture action data of the training personnel with the training standard action of the model base to obtain the training result of the training personnel.
Preferably, the method for recognizing the training video image by using human body posture estimation comprises the following steps of:
according to the human body skeleton joint model, joints of training personnel are extracted from the training video images at different visual angles by adopting human body posture estimation, and a plurality of groups of joint information are obtained;
performing iterative optimization on the plurality of groups of joint information by adopting a CNN convolutional neural network to obtain a group of key skeletal joint point information of the training personnel;
and connecting all joints in a group of key skeleton joint point information through a human body skeleton joint model to obtain the posture action data of the training personnel.
Preferably, the iterative optimization of the plurality of groups of joint information by using the CNN convolutional neural network to obtain a group of key skeletal joint information of the training personnel comprises: and matching and overlapping the same joints in the multiple groups of joint information by adopting a CNN convolutional neural network to obtain key nodes of the corresponding joints.
Preferably, matching and overlapping the same joints in the multiple groups of joint information by using a CNN convolutional neural network to obtain key nodes corresponding to the joints includes: in the matching and overlapping process, 3DPose reconstruction is carried out on a plurality of same joints by adopting Direct Linear Transformation (DLT) to obtain key nodes of the corresponding joints.
Preferably, the training data processing method for human body posture recognition includes: and judging whether the training action of the training personnel is qualified or not according to the training result, and marking the unqualified training action of the training personnel.
Preferably, the training results comprise standard actions, non-standard actions and error actions.
The invention also provides a training data processing device for human body posture recognition, which comprises: the system comprises a plurality of camera devices, an edge computing module and a cloud server;
the camera equipment is arranged on a training field and used for acquiring training video images of training personnel in real time and transmitting the training video images to the edge calculation module;
the edge computing module is used for executing according to the training data processing method for human body posture recognition, and transmitting a training result of a training person to the cloud server;
and the cloud server is used for storing the training results and the training video images of the training personnel.
Preferably, the training data processing device for human body posture recognition comprises a display module, and the display module is used for displaying the posture and action data of the training personnel obtained in the edge calculation module and the training result.
Preferably, the camera device is a 105-degree binocular 3D camera with an ultra-large field angle.
The invention also provides training data processing equipment for human body posture recognition, which comprises a processor and a memory;
the memory is used for storing program codes and transmitting the program codes to the processor;
the processor is used for executing the training data processing method for human body posture recognition according to the instructions in the program codes.
According to the technical scheme, the embodiment of the invention has the following advantages: the training data processing method, the device and the equipment for human body posture recognition comprise the following steps: acquiring training video images of different visual angles of training personnel in real time through camera equipment; adopting human body posture estimation to recognize the training video image, and recognizing to obtain posture action data of the training personnel; and comparing the posture action data of the training personnel with the training standard action of the model base to obtain the training result of the training personnel. According to the training data processing method for human body posture recognition, training video images of training personnel are obtained through the camera equipment, the training video images are analyzed through human body posture estimation, posture action data capable of being compared with training standard actions of a model base are obtained, training results of the training personnel are obtained through comparison and analysis, all standards and unqualified training action data of the training personnel can be accurately obtained, a training report is generated, a training plan can be conveniently formulated for the training personnel in a targeted mode, training data are also provided for scientific and effective training, and the technical problem that the action and the training data of the training personnel cannot be accurately obtained in an existing training system is solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without inventive exercise.
FIG. 1 is a flowchart illustrating steps of a training data processing method for human body posture recognition according to an embodiment of the present invention;
fig. 2 is a block diagram of a training data processing apparatus for human body posture recognition according to an embodiment of the present invention.
Detailed Description
In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the embodiments described below are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Interpretation of terms:
human body posture recognition belongs to an AI algorithm, all AI algorithms relate to learning training of the AI algorithm, the training data enable the algorithm to learn, the algorithm is issued after the detection reaches the standard, the training data can be understood as training learning data of posture recognition by people, but in the embodiment, the training data refers to real-time application data in various business training scenes.
The embodiment of the application provides a training data processing method, a training data processing device and training data processing equipment for human body posture recognition, and aims to solve the technical problem that the existing training system cannot accurately acquire actions and training data of training personnel.
The first embodiment is as follows:
fig. 1 is a flowchart illustrating steps of a training data processing method for human body posture recognition according to an embodiment of the present invention.
As shown in fig. 1, a training data processing method for human body posture recognition provided by the embodiment of the present invention includes the following steps:
s1, acquiring training video images of different visual angles of training personnel in real time through camera equipment.
It should be noted that, several cameras and several wireless hotspots are installed in the training field of the training personnel, and the several cameras are respectively installed at different positions of the training field, so that the cameras can conveniently acquire the training video images of the training personnel from different perspectives. In this embodiment, the training field may be an outdoor or indoor large-scale training field, and a plurality of wireless hotspots and camera devices are deployed around the training field according to requirements of training subjects of training personnel and the number of personnel participating in training, and the training personnel confirm the identity through face recognition before entering the field. The camera device can be a 105-degree binocular 3D camera with an ultra-large field angle or a camera, and can also be a device with a camera function.
And S2, recognizing the training video image by adopting human body posture estimation to obtain posture action data of the training personnel.
In step S2, the training video images of the motions of the trainee obtained in step S1 are mainly processed to obtain posture motion data corresponding to each motion, so that the trainee can be analyzed in step S3 according to the posture motion data.
In the embodiment of the invention, the human body posture estimation mainly comprises the steps of establishing a model according to human body skeleton joint information, obtaining all joint points forming human body skeletons, connecting corresponding joint points according to all joints of the human body skeletons, and realizing the human body posture estimation.
And S3, comparing the posture action data of the training personnel with the training standard action of the model base to obtain the training result of the training personnel.
It should be noted that the model library refers to a human skeleton node information model created according to the requirements of training subjects, and the model library includes various training standard actions. In this embodiment, the posture and motion data obtained in step S2 and trained by the trainee are mainly compared and analyzed with the corresponding normative motion in the model library to obtain the training result of the trainee. Wherein the training result comprises standard action, non-standard action and error action. For example: when the posture action data trained by the trainee is compared with the standard action and the similarity reaches more than 85 percent, the training result of the trainee is the standard action; when the posture action data trained by the trainee is compared with the standard action and the similarity reaches between 60% and 85%, the training result of the trainee is a non-standard action; when the posture action data trained by the trainer is compared with the standard action and the similarity is less than 60%, the training result of the trainer is a wrong action. Wherein, the similarity threshold value can be reasonably configured according to different training subjects.
The invention provides a training data processing method for human body posture recognition, which comprises the following steps: acquiring training video images of different visual angles of training personnel in real time through camera equipment; adopting human body posture estimation to recognize the training video image, and recognizing to obtain posture action data of the training personnel; and comparing the posture action data of the training personnel with the training standard action of the model base to obtain the training result of the training personnel. According to the training data processing method for human body posture recognition, training video images of training personnel are obtained through the camera equipment, the training video images are analyzed through human body posture estimation, posture action data capable of being compared with training standard actions of a model base are obtained, training results of the training personnel are obtained through comparison and analysis, all standards and unqualified training action data of the training personnel can be accurately obtained, a training report is generated, a training plan can be conveniently formulated for the training personnel in a targeted mode, training data are also provided for scientific and effective training, and the technical problem that the action and the training data of the training personnel cannot be accurately obtained in an existing training system is solved.
In an embodiment of the present invention, the step of recognizing the training video image by using human body posture estimation to obtain the posture and motion data of the training personnel includes:
extracting joints of training personnel from training video images with different visual angles by adopting human body posture estimation according to a human body skeleton joint model to obtain a plurality of groups of joint information;
performing iterative optimization on a plurality of groups of joint information by adopting a CNN convolutional neural network to obtain a group of key skeletal joint point information of a training person;
and mapping the key bone joint point information to the ground, and connecting all joints in the key bone joint point information mapped to the ground according to the human body skeleton joint model to obtain the posture and action data of the training personnel.
In the embodiment of the invention, the iterative optimization of a plurality of groups of joint information by adopting a CNN convolutional neural network to obtain a group of key skeletal joint information of a training person comprises the following steps: and matching and overlapping the same joints in the multiple groups of joint information by adopting a CNN convolutional neural network to obtain key nodes of the corresponding joints. And in the matching and overlapping process, performing 3DPose reconstruction on a plurality of same joints by adopting Direct Linear Transformation (DLT) to obtain key nodes of the corresponding joints.
It is noted that the training data processing method for human body posture recognition mainly obtains a plurality of groups of joint information according to the existing human body skeleton joint model and the human body posture estimation by acquiring all the joints of the human body of a training person from each training video image at the same moment, namely, a multi-target similar matrix is constructed by the obtained plurality of groups of joint information and the geographic position information of the corresponding camera equipment, and iterative optimization is carried out on the plurality of groups of joint information by adopting a CNN convolutional neural network technology to obtain a group of key skeleton joint point information of the training person; the method comprises the steps of obtaining a corresponding target matching matrix by iterative optimization solution of a multi-target similarity matrix, carrying out corresponding fine adjustment on key points (human body bone joints) of a group of training personnel matched with different visual angles to achieve sub-pixel matching, obtaining coordinate points (namely mapping processing) of all key bone joint point information of a human body skeleton of the training personnel relative to the ground, and then connecting all the key nodes to obtain posture action data of the training personnel. The method comprises the steps of carrying out 3DPose reconstruction based on Direct Linear Transformation (DLT) in the matching process, using a human skeleton joint model manufactured in the early stage to carry out comparison calculation, finally adjusting and outputting bone key nodes with higher precision, connecting all key nodes of a training person in sequence according to the human skeleton node model to obtain the bone key node model of the training person at a certain moment, wherein each bone key node model represents gesture action data of the training person.
In the embodiment of the invention, the training data processing method for human body posture recognition carries out recognition processing on each training person in a training video image through human body posture estimation, obtains posture action data of a single training person or a plurality of training persons through recognition, and realizes behavior recognition of single-person or multi-person training movement.
In an embodiment of the present invention, the training data processing method for human body posture recognition includes: and judging whether the training action of the training personnel is qualified or not according to the training result, and marking the unqualified training action of the training personnel.
If the training result is a standard action, the training action of the training personnel is judged to be qualified; and if the training result is a non-standard action and a wrong action, judging that the training action of the training personnel is unqualified. In this embodiment, the training data processing method for human body posture recognition further includes counting the judgment results of the training personnel each time, so that the training results of the training personnel can be conveniently checked by the relevant personnel.
Example two:
fig. 2 is a block diagram of a training data processing apparatus for human body posture recognition according to an embodiment of the present invention.
As shown in fig. 2, an embodiment of the present invention further provides a training data processing apparatus for human body posture recognition, including: a plurality of image pickup apparatuses 10, an edge calculation module 20, and a cloud server 30;
the camera device 10 is arranged on a training field and used for acquiring training video images of a trainer in real time and transmitting the training video images to the edge calculation module;
the edge computing module 20 is configured to execute the training data processing method according to the human body posture recognition, and transmit a training result of the obtained training person to the cloud server 30;
and the cloud server 30 is used for storing the training results of the training personnel.
It should be noted that the contents of the training data processing method for human body posture recognition in the device according to the second embodiment are already described in detail in the first embodiment, and the contents of the training data processing method for human body posture recognition in the device according to the second embodiment are not described in detail in this second embodiment.
In the embodiment of the present invention, the training data processing apparatus for human body posture recognition includes a display module 40, and the display module 40 is configured to display the posture and motion data of the trainee obtained in the edge calculation module 20 and the training result.
The display module 40 may be a display screen or a mobile terminal having a display function.
In the embodiment of the present invention, the image capturing apparatus 10 may be a binocular 3D camera or a camera with a 105 ° ultra-large field angle, or may be an apparatus having an image capturing function. Each camera device 10 collects training video images of training personnel and transmits the training video images to the edge calculation module 20 through video streams, and data communication between each camera device 10 and the edge calculation module 20 can adopt a wireless communication module or wireless communication modes such as WiFi.
In the embodiment of the present invention, the edge calculation module 20 mainly obtains posture action data of various trainings of each trainee in a training field according to a training data processing method for human posture recognition, and performs contrastive analysis with the standard actions of the current training subject, automatically obtains training results of standard actions, non-standard actions or error actions, and obtains a total accumulated number of training results, the training results and the total accumulated number of training results are referred to as training data, the training data are timely returned to the cloud server 30 for analysis and statistics, a training file is established for each trainee, an individual training scheme is formulated, and targeted training is performed for unqualified persons according to a standard action achievement rate specified by the training subject, different levels are distinguished, different training contents and training strengths are adopted, so that the training is more scientific and effective. For example, for indoor less-scaled daily physical training programs such as rotational stability, single-leg squat, passing-over deep squat, stability push-ups, and the like.
In the embodiment of the present invention, the cloud server 30 is mainly used for storing data. Wherein data communication between the cloud server 30 and the edge computing module 20 may be transmitted through the 5G communication module.
In the embodiment of the invention, the training data processing device for human body posture recognition displays the training video image and the training result of the real-time training personnel through the display module 40, adopts the cloud server 30 to establish the personal training file and make the personal progress scheme according to the training result of the training personnel, organizes the unqualified training personnel to carry out the targeted training, distinguishes different levels, adopts different training contents and training intensity, and makes the training more scientific and effective. For example: taking the training of the push-up body as an example, the posture recognition effect in the training process is shown. The human body posture recognition real-time acquisition training personnel human body skeleton joint model, according to the skeleton node model base of various standard actions of the push-up project, whether each posture action data accords with the standard action is automatically judged, corresponding correction prompts are given in time for non-standard actions, and an individual training report is automatically generated after the specified training time is finished, a current training result can be provided for the training personnel through the training report and training personnel files, a data basis is provided for accurate problem diagnosis of the training personnel, and personalized guidance training is realized.
Example three:
the embodiment of the invention provides training data processing equipment for human body posture recognition, which comprises a processor and a memory, wherein the processor is used for processing training data;
a memory for storing the program code and transmitting the program code to the processor;
and the processor is used for executing the training data processing method for human body posture recognition according to the instructions in the program codes.
It should be noted that the processor is configured to execute the steps in the above-mentioned embodiment of the training data processing method for human body posture recognition according to the instructions in the program code. Alternatively, the processor, when executing the computer program, implements the functions of each module/unit in each system/apparatus embodiment described above.
Illustratively, a computer program may be partitioned into one or more modules/units, which are stored in a memory and executed by a processor to accomplish the present application. One or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution of a computer program in a terminal device.
The terminal device may be a desktop computer, a notebook, a palm computer, a cloud server, or other computing devices. The terminal device may include, but is not limited to, a processor, a memory. Those skilled in the art will appreciate that the terminal device is not limited and may include more or fewer components than those shown, or some components may be combined, or different components, e.g., the terminal device may also include input output devices, network access devices, buses, etc.
The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The storage may be an internal storage unit of the terminal device, such as a hard disk or a memory of the terminal device. The memory may also be an external storage device of the terminal device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided on the terminal device. Further, the memory may also include both an internal storage unit of the terminal device and an external storage device. The memory is used for storing computer programs and other programs and data required by the terminal device. The memory may also be used to temporarily store data that has been output or is to be output.
It can be clearly understood by those skilled in the art that, for convenience and simplicity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A training data processing method for human body posture recognition is characterized by comprising the following steps:
acquiring training video images of different visual angles of training personnel in real time through camera equipment;
adopting human body posture estimation to recognize the training video image, and recognizing to obtain posture action data of the training personnel;
and comparing the posture action data of the training personnel with the training standard action of the model base to obtain the training result of the training personnel.
2. The method for processing training data of human body posture recognition according to claim 1, wherein the training video image is recognized by human body posture estimation, and the step of recognizing posture motion data of a training person comprises:
according to the human body skeleton joint model, joints of training personnel are extracted from the training video images at different visual angles by adopting human body posture estimation, and a plurality of groups of joint information are obtained;
performing iterative optimization on the plurality of groups of joint information by adopting a CNN convolutional neural network to obtain a group of key skeletal joint point information of the training personnel;
and connecting all joints in a group of key skeleton joint point information through a human body skeleton joint model to obtain the posture action data of the training personnel.
3. The method for processing training data of human body posture recognition according to claim 2, wherein performing iterative optimization on a plurality of groups of joint information by using a CNN convolutional neural network to obtain a group of key skeletal joint information of a training person comprises: and matching and overlapping the same joints in the multiple groups of joint information by adopting a CNN convolutional neural network to obtain key nodes of the corresponding joints.
4. The method for processing training data of human body posture recognition according to claim 3, wherein matching and overlapping the same joints in the plurality of groups of joint information by using a CNN convolutional neural network to obtain key nodes of the corresponding joints comprises: in the matching and overlapping process, 3DPose reconstruction is carried out on a plurality of same joints by adopting Direct Linear Transformation (DLT) to obtain key nodes of the corresponding joints.
5. The method for processing training data of human body posture recognition according to claim 1, comprising: and judging whether the training action of the training personnel is qualified or not according to the training result, and marking the unqualified training action of the training personnel.
6. The method for processing training data of human body posture recognition according to claim 1, wherein the training result includes standard action, non-standard action and error action.
7. A training data processing device for human body posture recognition is characterized by comprising: the system comprises a plurality of camera devices, an edge computing module and a cloud server;
the camera equipment is arranged on a training field and used for acquiring training video images of training personnel in real time and transmitting the training video images to the edge calculation module;
the edge calculation module is used for executing according to the training data processing method for human body posture recognition as claimed in any one of claims 1 to 6, and transmitting the obtained training result of the training personnel to the cloud server;
and the cloud server is used for storing the training results and the training video images of the training personnel.
8. The human body posture recognition training data processing device according to claim 7, comprising a display module, wherein the display module is used for displaying the posture and motion data of the training personnel obtained in the edge calculation module and the training result.
9. The human body posture recognition training data processing device as claimed in claim 7, wherein the camera device is a 105 ° binocular 3D camera with an ultra-large field angle or a camera.
10. The training data processing equipment for human body posture recognition is characterized by comprising a processor and a memory;
the memory is used for storing program codes and transmitting the program codes to the processor;
the processor is used for executing the training data processing method for human body posture recognition according to any one of claims 1-6 according to instructions in the program code.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210096397.3A CN114495169A (en) | 2022-01-26 | 2022-01-26 | Training data processing method, device and equipment for human body posture recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210096397.3A CN114495169A (en) | 2022-01-26 | 2022-01-26 | Training data processing method, device and equipment for human body posture recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114495169A true CN114495169A (en) | 2022-05-13 |
Family
ID=81476646
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210096397.3A Pending CN114495169A (en) | 2022-01-26 | 2022-01-26 | Training data processing method, device and equipment for human body posture recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114495169A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114863750A (en) * | 2022-06-17 | 2022-08-05 | 中国人民解放军32128部队 | Bolt dismounting training device and training method |
| CN115937894A (en) * | 2022-08-11 | 2023-04-07 | 北京中微盛鼎科技有限公司 | Military training method and system based on human body posture recognition |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111652078A (en) * | 2020-05-11 | 2020-09-11 | 浙江大学 | Yoga action guidance system and method based on computer vision |
| CN112560665A (en) * | 2020-12-13 | 2021-03-26 | 同济大学 | Professional dance evaluation method for realizing human body posture detection based on deep migration learning |
| CN113516064A (en) * | 2021-07-02 | 2021-10-19 | 深圳市悦动天下科技有限公司 | Method, device, equipment and storage medium for judging sports motion |
| CN113658211A (en) * | 2021-07-06 | 2021-11-16 | 江汉大学 | User posture evaluation method and device and processing equipment |
-
2022
- 2022-01-26 CN CN202210096397.3A patent/CN114495169A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111652078A (en) * | 2020-05-11 | 2020-09-11 | 浙江大学 | Yoga action guidance system and method based on computer vision |
| CN112560665A (en) * | 2020-12-13 | 2021-03-26 | 同济大学 | Professional dance evaluation method for realizing human body posture detection based on deep migration learning |
| CN113516064A (en) * | 2021-07-02 | 2021-10-19 | 深圳市悦动天下科技有限公司 | Method, device, equipment and storage medium for judging sports motion |
| CN113658211A (en) * | 2021-07-06 | 2021-11-16 | 江汉大学 | User posture evaluation method and device and processing equipment |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114863750A (en) * | 2022-06-17 | 2022-08-05 | 中国人民解放军32128部队 | Bolt dismounting training device and training method |
| CN115937894A (en) * | 2022-08-11 | 2023-04-07 | 北京中微盛鼎科技有限公司 | Military training method and system based on human body posture recognition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114097248B (en) | Video stream processing method, device, equipment and medium | |
| CN114495169A (en) | Training data processing method, device and equipment for human body posture recognition | |
| CN107241572B (en) | Training video tracking evaluation system for students | |
| WO2014187223A1 (en) | Method and apparatus for identifying facial features | |
| CN110472613B (en) | Object behavior identification method and device | |
| CN110544301A (en) | Three-dimensional human body action reconstruction system, method and action training system | |
| TWI769787B (en) | Target tracking method and apparatus, storage medium | |
| CN108681700A (en) | A kind of complex behavior recognition methods | |
| CN113239797B (en) | Human body action recognition method, device and system | |
| CN112528812A (en) | Pedestrian tracking method, pedestrian tracking device and pedestrian tracking system | |
| CN111582240B (en) | Method, device, equipment and medium for identifying number of objects | |
| WO2023097967A1 (en) | Action detection method and apparatus, device, storage medium, and computer program product | |
| CN113947742A (en) | Person trajectory tracking method and device based on face recognition | |
| CN111967407B (en) | Action evaluation method, electronic device, and computer-readable storage medium | |
| CN111507301B (en) | Video processing method, video processing device, computer equipment and storage medium | |
| CN111241926A (en) | Attendance checking and learning condition analysis method, system, equipment and readable storage medium | |
| CN112001209A (en) | Student classroom learning behavior monitoring system based on artificial intelligence | |
| CN111626212B (en) | Method and device for identifying object in picture, storage medium and electronic device | |
| CN115205750B (en) | Motion real-time counting method and system based on deep learning model | |
| CN115346640A (en) | Intelligent monitoring method and system for closed-loop feedback of functional rehabilitation training | |
| CN112099330B (en) | Holographic human body reconstruction method based on external camera and wearable display control equipment | |
| CN114494427A (en) | Method, system and terminal for detecting illegal behavior of person standing under suspension arm | |
| CN113724176A (en) | Multi-camera motion capture seamless connection method, device, terminal and medium | |
| CN112200698A (en) | Campus social relationship big data analysis system based on artificial intelligence | |
| CN111967406A (en) | Method, system, equipment and storage medium for generating human body key point detection model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220513 |
|
| RJ01 | Rejection of invention patent application after publication |