Walking detection method and computer equipment based on intelligent walker

Abstract

The embodiments of the present application relate to the field of walking detection, and specifically to a walking detection method and computer device based on an intelligent walker. A specific implementation of the method includes: performing signal conversion processing on the target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal collected in advance within a preset time period to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence; integrating and processing to obtain walking muscle activity data; inputting into a pre-trained muscle feature extraction model to obtain walking muscle feature information; determining force difference values and muscle abnormality information; and switching the intelligent walker operating mode in response to determining that a preset abnormality condition is met. This implementation improves the safety of the intelligent walker during use.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of walking detection, and specifically to a walking detection method and computer device based on an intelligent walker. Background Art

[0002] Elderly patients often face numerous health issues after surgery, including common lower limb blood clots and lung infections, and even more common muscle weakness or loss of muscle mass. In fact, muscle loss in elderly patients is often systemic, due to the close connection between muscle, protein, and the immune system. This increases the risk of sarcopenia after surgery. Elderly patients differ significantly from younger patients in terms of their tolerance for physical issues, pain tolerance, and postoperative recovery. Consequently, the aforementioned postoperative issues are more common in elderly patients, and the subsequent consequences are significant. Given the high prevalence of sarcopenia and its potential adverse consequences, preventing its occurrence is crucial. Current prevention strategies primarily encourage elderly patients to engage in physical exercise and attempt to ambulate as soon as possible. However, mobility is already extremely difficult for elderly patients after surgery, and muscle loss is significantly accelerated, especially after major surgery.

[0003] The current solution to this problem is to use a walker to assist elderly patients with postoperative exercise. During this exercise, patients undergo a series of tests. These tests include a body composition analyzer, which provides detailed measurements of body composition, including body fat, weight, BMI (body mass index), and fat-free mass; a handgrip dynamometer, which measures grip strength; and a gait analysis device, which assesses patients' walking posture and other aspects. The combined use of these instruments allows for a more systematic monitoring of muscle loss. This monitoring typically begins on the third day after surgery, when patients are able to resume activity. A similar monitoring session is also conducted preoperatively to collect relevant data, which is then compared with postoperative data for a more comprehensive understanding of changes in the patient's muscle condition.

[0004] However, existing rehabilitation prevention and treatment programs have many shortcomings. On the one hand, their operation is not convenient enough. Whether it is the use of the walker or the operation of various testing instruments, it requires a certain amount of manpower and material resources, which places a considerable burden on medical staff and patients. On the other hand, the program lacks systematicity and cannot conduct timely and complete all-round monitoring of the patient's muscle condition. This may result in the omission of some important information, which in turn affects the accurate diagnosis and effective intervention of the patient's condition. As a result, the safety of the smart walker is reduced when in use. Summary of the Invention

[0005] The content of this application is used to briefly introduce concepts that will be described in detail in the detailed description section below. The content of this application is not intended to identify key features or essential features of the technical solution for which protection is sought, nor is it intended to limit the scope of the technical solution for which protection is sought.

[0006] Some embodiments of the present application propose a walking detection method, apparatus, computer device, and computer-readable storage medium based on an intelligent walker to solve one or more of the technical problems mentioned in the above background technology section.

[0007] In a first aspect, some embodiments of the present application provide a walking detection method based on an intelligent walker, the method comprising: performing signal conversion processing on an initial walking pressure signal, an initial walking electromyographic signal, and an initial walking acceleration signal of a target user collected in advance within a preset time period to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence, wherein the above-mentioned initial walking pressure signal is collected by a control pressure sensor, the above-mentioned initial walking electromyographic signal is collected by a control electromyographic sensor, and the above-mentioned initial walking acceleration signal is collected by a control acceleration sensor; integrating the above-mentioned walking muscle pressure value sequence, the above-mentioned walking muscle activity state information sequence, and the above-mentioned walking motion parameter information sequence to obtain walking muscle activity data; inputting the above-mentioned walking muscle activity data into a pre-trained muscle feature extraction model to obtain walking muscle feature information; determining a force difference value and muscle abnormality information corresponding to the above-mentioned walking muscle feature information and the pre-acquired initial muscle feature information; in response to determining that the above-mentioned force difference value is greater than a target threshold or the above-mentioned muscle abnormality information meets a preset abnormality condition, sending a preset mode switching signal to a mobile controller to switch the operation mode of the intelligent walker.

[0008] In the second aspect, some embodiments of the present disclosure provide a walking detection device based on an intelligent walker, the device comprising: a signal conversion unit, configured to perform signal conversion processing on the initial walking pressure signal, initial walking electromyographic signal and initial walking acceleration signal of the target user collected in advance within a preset time period, to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence and a walking motion parameter information sequence, wherein the above-mentioned initial walking pressure signal is collected by a control pressure sensor, the above-mentioned initial walking electromyographic signal is collected by a control electromyographic sensor, and the above-mentioned initial walking acceleration signal is collected by a control acceleration sensor; an integration unit, configured to perform signal conversion processing on the above-mentioned walking muscle The muscle pressure value sequence, the walking muscle activity state information sequence and the walking motion parameter information sequence are integrated and processed to obtain walking muscle activity data; the input unit is configured to input the walking muscle activity data into a pre-trained muscle feature extraction model to obtain walking muscle feature information; the determination unit is configured to determine the force difference value and muscle abnormality information corresponding to the walking muscle feature information and the pre-acquired initial muscle feature information; the sending unit is configured to send a preset mode switching signal to the mobile controller to switch the intelligent walker operation mode in response to determining that the force difference value is greater than the target threshold or the muscle abnormality information meets the preset abnormality condition.

[0009] In a third aspect, the present application also provides a computer device, comprising a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, it implements the method described in any implementation of the first aspect.

[0010] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, wherein when the computer program is executed by a processor, the method described in any implementation of the first aspect is implemented.

[0011] The above-mentioned embodiments of the present application have the following beneficial effects: Through the walking detection method based on an intelligent walker in some embodiments of the present application, the safety of the intelligent walker during use can be improved. Specifically, the reasons for the reduced safety of the intelligent walker during use are: on the one hand, its operation is not convenient enough. Whether it is the use of the walker or the operation of various detection instruments, it requires a certain amount of manpower and material resources, which places a considerable burden on medical staff and patients; on the other hand, the system is also lacking in systematization, and it is unable to timely and completely monitor the patient's muscle state in all directions, resulting in the possibility of missing some important information, which in turn affects the accurate diagnosis and effective intervention of the patient's condition. Based on this, the walking detection method based on an intelligent walker in some embodiments of the present application first converts the initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal of the target user collected within a preset time period into a signal sequence, thereby obtaining a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence. In this way, it is possible to collect signals representing the user's muscle state and convert the collected electrical signals into analyzable data information. Then, the walking muscle pressure value sequence, the walking muscle activity state information sequence and the walking motion parameter information sequence are integrated and processed to obtain walking muscle activity data. Thus, the information after conversion of each signal can be integrated to extract muscle feature information. Next, the walking muscle activity data is input into a pre-trained muscle feature extraction model to obtain walking muscle feature information. Thus, the user's current muscle state characteristics can be obtained. Then, the force difference value and muscle abnormality information corresponding to the walking muscle feature information and the pre-acquired initial muscle feature information are determined. Thus, the user's muscle state change information can be obtained. Finally, in response to determining that the force difference value is greater than the target threshold or the muscle abnormality information meets the preset abnormality condition, the preset mode switching signal is sent to the mobile controller to switch the intelligent walker operation mode. Thus, the operation state of the intelligent walker can be adjusted according to the user's muscle state change information. Therefore, some walking detection methods based on smart walkers in this application can embed the smart walker with monitoring test data acquisition software. During the rehabilitation training process, the walker structure is equipped with sensors and other electronic equipment to conduct real-time testing of the patient's condition, timely detect abnormal muscle loss in the patient, and reduce the number of times the patient falls due to muscle weakness, thereby improving the safety of the smart walker when in use. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other features, advantages, and aspects of the various embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that the elements and components are not necessarily drawn to scale.

[0013] Figure 1 is a flowchart of some embodiments of the walking detection method based on the intelligent walker according to the present application;

[0014] Figure 2 is a schematic diagram of application scenarios of some embodiments of the walking detection method based on an intelligent walker according to the present application;

[0015] Figure 3 is a schematic diagram of application scenarios of other embodiments of the walking detection method based on an intelligent walker according to the present application;

[0016] Figure 4 is a schematic structural diagram of some embodiments of a walking detection device based on an intelligent walker according to the present application;

[0017] Figure 5 is a schematic diagram of the structure of a computer device suitable for implementing some embodiments of the present application;

[0018] Figure 6 Schematic diagram of the structure of some embodiments of the intelligent walking aid according to the present application. DETAILED DESCRIPTION

[0019] The following will describe embodiments of the present application in more detail with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are for illustrative purposes only and are not intended to limit the scope of protection of the present application.

[0020] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of this application can be combined with each other.

[0021] It should be noted that the concepts of "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0022] It should be noted that the modifications of "one" and "multiple" mentioned in this application are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more".

[0023] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0024] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0025] Figure 1 The flowchart 100 of some embodiments of the walking detection method based on an intelligent walker according to the present application is shown. The walking detection method based on an intelligent walker is applied to an intelligent walker, which includes: a pressure sensor, an electromyographic sensor, an acceleration sensor, a movement controller, and a camera assembly, and includes the following steps:

[0026] Step 101 , performing signal conversion processing on the target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal collected in advance within a preset time period, respectively, to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence.

[0027] In some embodiments, the execution entity may perform signal conversion processing on the target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal, collected pre-collected over a preset time period, to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence. The initial walking pressure signal may be collected by a control pressure sensor. The initial walking electromyographic signal may be collected by a control electromyographic sensor. The initial walking acceleration signal may be collected by a control acceleration sensor. Specifically, the target user may be a user walking with an intelligent walker. The pressure sensor may be a sensor for collecting the target user's pressure signal. The electromyographic sensor may be a sensor for collecting the target user's electromyographic signal. The acceleration sensor may be a sensor for measuring the target user's acceleration. The movement controller may be a controller for controlling the movement of the intelligent walker. The camera in the camera assembly may be a camera for capturing scene images. The camera assembly may be divided into a first camera assembly and a second camera assembly. The first camera in the first camera assembly may be a camera located on a leg of the intelligent walker in the forward direction of travel. The second camera assembly may be any camera in the camera assembly except the first camera assembly.

[0028] As an example, the preset time length may be, but is not limited to, at least one of the following: 1 second, 30 milliseconds, or 10 milliseconds. The execution subject may be a central processing unit. The specific structure of the intelligent walking aid may refer to Figure 6 FIG. 6 shows a schematic structural diagram 600 of some embodiments of the intelligent walking aid according to the present application. Figure 6 As shown, the intelligent walker 600 may include, but is not limited to, a pressure sensor 601, an electromyographic sensor 602, an acceleration sensor 603, a central processing unit 604, a motion controller 605, a camera assembly 606, an armrest 607, a foot assembly 608, a seat 609, and a wheel assembly 610. The pressure sensor 601 and electromyographic sensor 602 may be mounted on the armrest 607. The acceleration sensor 603 and camera assembly 606 may be mounted on the foot 608. The central processing unit 604 and motion controller 605 may be mounted in the seat 609.

[0029] In some optional implementations of some embodiments, the execution entity performs signal conversion processing on the target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal collected in advance within a preset time period to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence, which may include the following steps:

[0030] In the first step, the target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal, which are collected in advance within a preset time period, are preprocessed to obtain a walking pressure filter signal, a walking electromyographic filter signal, and a walking acceleration filter signal. The target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal, which are collected in advance within a preset time period, can be preprocessed in advance using a preset signal processing algorithm to obtain the walking pressure filter signal, the walking electromyographic filter signal, and the walking acceleration filter signal.

[0031] As an example, the above-mentioned preset signal processing algorithm may include but is not limited to at least one of the following: a median filter algorithm, a wavelet transform algorithm, a Fourier transform algorithm, and a polyphase filter algorithm.

[0032] The second step is to perform pressure conversion processing on the walking pressure filter signal to obtain a walking muscle pressure value sequence. The walking pressure filter signal can be pressure-converted using a preset pressure conversion algorithm to obtain the walking muscle pressure value sequence.

[0033] As an example, the preset pressure conversion algorithm may be, but is not limited to, at least one of the following: a linear regression algorithm or a polynomial regression algorithm.

[0034] The third step is to perform feature extraction processing on the walking electromyography filter signal to obtain a walking muscle activity state information sequence. The feature extraction processing can be performed on the walking electromyography filter signal using a preset feature extraction algorithm to obtain the walking muscle activity state information sequence.

[0035] As an example, the preset feature extraction algorithm may be, but is not limited to, at least one of the following: a principal component analysis algorithm, a genetic algorithm, and a heterogeneous feature mapping algorithm.

[0036] The fourth step is to split the walking acceleration filter signal to obtain a walking acceleration sub-signal sequence. The walking acceleration filter signal can be evenly split at preset time intervals to obtain the walking acceleration sub-signal sequence.

[0037] As an example, the preset time interval may be 5 milliseconds.

[0038] Step 5: For each walking acceleration sub-signal sequence in the above walking acceleration sub-signal sequence, perform the following detection sub-steps:

[0039] The first sub-step is to perform gait detection processing on the walking acceleration sub-signal to obtain a walking speed value, a walking cadence value, and a walking stride value. The gait detection processing can be performed on the walking acceleration sub-signal using a preset gait detection algorithm to obtain the walking speed value, the walking cadence value, and the walking stride value.

[0040] As an example, the above-mentioned preset gait detection algorithm may include but is not limited to at least one of the following: an acceleration amplitude formula, a cadence formula, and a Weinberg nonlinear stride estimation model algorithm.

[0041] The second sub-step is to fuse the walking speed value, the walking frequency value, and the walking stride value to obtain walking motion parameter information. The fusing of the walking speed value, the walking frequency value, and the walking stride value to obtain the walking motion parameter information may include determining the walking speed value, the walking frequency value, and the walking stride value as the walking speed value, the walking frequency value, and the walking stride value included in the walking motion parameter information.

[0042] The sixth step is to determine the generated walking motion parameter information as a walking motion parameter information sequence.

[0043] Step 102 : Integrate the walking muscle pressure value sequence, the walking muscle activity state information sequence, and the walking motion parameter information sequence to obtain walking muscle activity data.

[0044] In some embodiments, the execution entity may integrate the walking muscle pressure value sequence, the walking muscle activity state information sequence, and the walking motion parameter information sequence to obtain the walking muscle activity data. The walking muscle activity data may be obtained by integrating the walking muscle pressure value sequence, the walking muscle activity state information sequence, and the walking motion parameter information sequence using a preset integration algorithm.

[0045] As an example, the aforementioned preset integration algorithm may be, but is not limited to, at least one of the following: a Kalman filter algorithm, a sliding window algorithm, or a dynamic time warping algorithm.

[0046] Step 103: Input the walking muscle activity data into a pre-trained muscle feature extraction model to obtain walking muscle feature information.

[0047] In some embodiments, the execution entity may input the walking muscle activity data into a pre-trained muscle feature extraction model to obtain walking muscle feature information. The pre-trained muscle feature extraction model may be a pre-trained neural network model that takes walking muscle activity data as input and takes walking muscle feature information as output. The walking muscle feature information may include: walking muscle strength value and walking muscle content information. The walking muscle strength value may represent the muscle strength of the target user at the current moment. The walking muscle content information may represent the muscle content of the target user at the current moment.

[0048] Optionally, before inputting the walking muscle activity data into the pre-trained muscle feature extraction model to obtain walking muscle feature information, the execution subject may further perform the following steps:

[0049] The first step is to obtain a sample walking muscle activity data set. The sample walking muscle activity data set can be obtained from a storage terminal. The sample walking muscle activity data in the sample walking muscle activity data set can be pre-collected data representing muscle characteristics.

[0050] The second step is to select target sample walking muscle activity data from the above sample walking muscle activity dataset and perform the following first training sub-step:

[0051] In the first sub-step, the target sample walking muscle activity data is input into the three-dimensional feature extraction sub-model included in the initial muscle feature extraction model to obtain the initial muscle three-dimensional feature information. The initial muscle feature extraction model also includes: a three-dimensional attention sub-model, an aggregation sub-model, a comprehensive feature extraction sub-model, a feature fusion sub-model and a feature mapping sub-model. The sample walking muscle activity data can be randomly selected from the above-mentioned sample walking muscle activity data set as the target sample walking muscle activity data. The above-mentioned initial muscle feature extraction model can be an untrained neural network model that takes the target sample walking muscle activity data as input and the initial muscle feature information as output.

[0052] Specifically, the 3D feature extraction sub-model can be a neural network model that takes the target sample walking muscle activity data as input and outputs the initial muscle 3D feature information. For example, the 3D feature extraction sub-model can be a 3D convolutional model.

[0053] The 3D attention sub-model can be a neural network model that takes initial muscle 3D feature information as input and outputs initial muscle 3D attention information. For example, the 3D attention sub-model can be a 3D-CBAM (3D-Convolutional Block Attention Module) model.

[0054] The aggregation sub-model may be a neural network model that takes initial muscle three-dimensional attention information as input and outputs initial muscle aggregation information. For example, the aggregation sub-model may be an LTA (Local Temporal Aggregation) model.

[0055] The comprehensive feature extraction sub-model can be a neural network model that takes initial muscle aggregation information as input and outputs initial muscle comprehensive feature information. For example, the comprehensive feature extraction sub-model can be an EGLFE (enhanced global and local feature extractor) model.

[0056] The feature fusion sub-model may be a neural network model that takes the initial muscle comprehensive feature information as input and outputs the initial muscle feature fusion information. For example, the feature fusion sub-model may be a residual convolution model.

[0057] The feature mapping sub-model may be a neural network model that takes the initial muscle feature fusion information as input and outputs the initial muscle feature fusion information. For example, the feature mapping sub-model may be a maximum pooling layer.

[0058] In the second sub-step, the initial muscle three-dimensional feature information is input into the three-dimensional attention sub-model included in the initial muscle feature extraction model to obtain the initial muscle three-dimensional attention information.

[0059] In the third sub-step, the initial muscle three-dimensional attention information is input into the aggregation sub-model included in the initial muscle feature extraction model to obtain the initial muscle aggregation information.

[0060] In the fourth sub-step, the initial muscle aggregation information is input into the comprehensive feature extraction sub-model included in the initial muscle feature extraction model to obtain the initial muscle comprehensive feature information.

[0061] In the fifth sub-step, the initial muscle comprehensive feature information is input into the feature fusion sub-model included in the initial muscle feature extraction model to obtain the initial muscle feature fusion information.

[0062] In the sixth sub-step, the initial muscle feature fusion information is input into the feature mapping sub-model included in the initial muscle feature extraction model to obtain the initial muscle feature information.

[0063] The seventh sub-step is to determine a feature extraction loss value corresponding to the initial muscle feature information based on at least one preset feature extraction loss function.

[0064] As an example, the at least one preset feature extraction loss function may include but is not limited to at least one of the following: a triple loss function and an upper loss function.

[0065] In an eighth sub-step, in response to determining that the feature extraction loss value is less than the first target threshold, the initial muscle feature extraction model is determined as the muscle feature extraction model.

[0066] As an example, the first target threshold may be 0.01.

[0067] Optionally, the execution entity may further adjust relevant parameters in the initial muscle feature extraction model in response to determining that the feature extraction loss value is greater than or equal to the first target threshold, determine the adjusted initial muscle feature extraction model as the initial muscle feature extraction model, and select target sample walking muscle activity data from each unselected sample walking muscle activity data in the sample walking muscle activity dataset for re-executing the first training step. The relevant parameters in the initial muscle feature extraction model may be adjusted using a preset adjustment algorithm.

[0068] As an example, the preset adjustment algorithm may be, but is not limited to, at least one of the following: a back propagation algorithm or a stochastic gradient algorithm.

[0069] Therefore, a 3D feature extraction model and a 3D attention model can be used to extract the spatial and temporal features of walking muscle activity data. By fusing these multi-scale spatial and temporal features, the dimensionality and accuracy of feature extraction are increased. This improves the accuracy of muscle feature information extracted by the model (for example, the accuracy of a patient's current muscle strength value), enabling the timely detection of abnormal muscle loss in patients.

[0070] Step 104 : determining the strength difference value and muscle abnormality information corresponding to the walking muscle characteristic information and the pre-acquired initial muscle characteristic information.

[0071] In some embodiments, the execution entity may determine the strength difference value and muscle abnormality information corresponding to the walking muscle characteristic information and the pre-acquired initial muscle characteristic information. The initial muscle characteristic information may be obtained from a storage terminal. The storage terminal may be a terminal for storing initial muscle characteristic information. The initial muscle characteristic information may characterize the muscle characteristics of the target user at a moment before the preset time interval of the current moment. The initial muscle characteristic information may include: an initial walking muscle strength value and an initial walking muscle content information. The initial walking muscle strength value may characterize the muscle strength of the target user at a moment before the preset time interval of the current moment. The initial walking muscle content information may characterize the muscle content of the target user at a moment before the preset time interval of the current moment. Specifically, the determination of the strength difference value and muscle abnormality information corresponding to the walking muscle characteristic information and the initial muscle characteristic information may be as follows: first, determining the difference between the walking muscle strength value included in the walking muscle characteristic information and the initial walking muscle strength value included in the initial muscle characteristic information as the strength difference value. Then, a preset abnormality identification algorithm is used to determine muscle abnormality information corresponding to the walking muscle content information included in the walking muscle characteristic information and the initial walking muscle content information included in the initial muscle characteristic information. The muscle abnormality information can be information indicating "abnormal muscle loss" or information indicating "no abnormal muscle loss."

[0072] As an example, the above-mentioned preset anomaly recognition algorithm can be, but is not limited to, at least one of the following: an LSTM (Long Short-Term Memory) algorithm, a dynamic clustering algorithm, or an exponential smoothing algorithm.

[0073] Step 105 : In response to determining that the force difference value is greater than the target threshold or the muscle abnormality information satisfies a preset abnormality condition, a preset mode switching signal is sent to the mobile controller to switch the operation mode of the intelligent walker.

[0074] In some embodiments, the execution entity may send a preset mode switching signal to the mobility controller to switch the intelligent walker's operating mode in response to determining that the force difference value is greater than a target threshold or that the muscle abnormality information satisfies a preset abnormality condition. The preset abnormality condition may be that the muscle abnormality information indicates "abnormal muscle loss."

[0075] As an example, the target threshold may be 10. The preset mode switching signal may be a signal indicating "increasing the walking assistance level of the mobile controller."

[0076] Optionally, the above execution entity may further perform the following steps:

[0077] The first step is to synthesize the individual walking scene images in a pre-captured walking scene image set of the target user to obtain a target walking scene image. The walking scene image set may be captured by controlling a camera assembly. Here, each camera in the camera assembly may be controlled to capture at least one walking scene image to obtain the walking scene image set. The walking scene images in the walking scene image set may represent the characteristics of the scene surrounding the target user while using the smart walker.

[0078] The second step is to perform scene recognition processing on the target walking scene image to obtain walking scene feature information. The scene recognition processing can be performed on the target walking scene image using a preset scene recognition algorithm to obtain the walking scene feature information.

[0079] As an example, the preset scene recognition algorithm may be, but is not limited to, at least one of the following: a naive Bayes algorithm, a support vector machine algorithm, or a random forest algorithm. The walking scene feature information may be, but is not limited to, at least one of the following: information indicating steady movement, information indicating stationary movement, or information indicating falling.

[0080] In the third step, in response to determining that the walking scene characteristic information meets the preset dangerous scene condition, a preset mode switching signal is sent to the mobile controller to switch the intelligent walker operation mode.

[0081] As an example, the preset dangerous scene condition may be that the walking scene characteristic information is information indicating "falling down." The preset mode switching signal may be a signal indicating that "the mobile controller turns on the balance mode."

[0082] In some optional implementations of some embodiments, the execution subject synthesizes each walking scene image in a set of pre-taken walking scene images of a target user to obtain a target walking scene image, which may include the following steps:

[0083] In the first step, a set of pre-captured walking scene images of a target user is grouped to obtain a first scene image set and a second scene image set. The first scene image set may include images of the walking scene captured by the first camera assembly in the camera assembly, and the second scene image set may include images of the walking scene captured by the second camera assembly in the camera assembly.

[0084] The second step is to construct an image generation model based on the first scene image set. The image generation model can be constructed based on the first scene image set by using a preset construction algorithm.

[0085] As an example, the aforementioned preset construction algorithm may be a NeRF (Neural Radiance Fields) algorithm.

[0086] The third step is to generate a first scene generated image set and a second scene generated image set corresponding to the first scene image set and the second scene image set respectively based on the image generation model.

[0087] In a fourth step, each first scene generated image in the first scene generated image set is input into a pre-trained image mapping model to generate a first scene mapping image, thereby obtaining a first scene mapping image set. The pre-trained image mapping model may be a pre-trained neural network model that takes the first scene generated image as input and outputs the first scene mapping image.

[0088] In the fifth step, each second scene generated image in the above second scene generated image set is input into a pre-trained image mapping model to generate a second scene mapping image, thereby obtaining a second scene mapping image set.

[0089] In step 6, the first scene mapping image set and the second scene mapping image set are subjected to 3D reconstruction processing to obtain a target walking scene image. The target walking scene image can be obtained by performing 3D reconstruction processing on the first scene mapping image set and the second scene mapping image set using a preset 3D reconstruction algorithm.

[0090] As an example, the above-mentioned preset 3D reconstruction algorithm can be but is not limited to at least one of the following: SfM (Structure from Motion) algorithm, MVS (Multi-View Stereo) algorithm or Monodepth2 (monocular depth estimation) algorithm.

[0091] In some optional implementations of some embodiments, the execution entity generates, based on the image generation model, a first scene generated image set and a second scene generated image set corresponding to the first scene image set and the second scene image set, respectively, which may include the following steps:

[0092] In the first step, a pose extraction process is performed on each first scene image in the first scene image set to generate first scene pose information, thereby obtaining a first scene pose information set. The pose extraction process can be performed on each first scene image in the first scene image set to generate first scene pose information, thereby obtaining a first scene pose information set, using a preset pose extraction algorithm.

[0093] As an example, the above-mentioned preset pose extraction algorithm can be but is not limited to at least one of the following: an epipolar geometry algorithm, a direct linear transformation algorithm or a PoseNet (pose network) model algorithm.

[0094] In the second step, each first scene pose information in the above-mentioned first scene pose information set is input into the above-mentioned image generation model to generate a first scene generation image, thereby obtaining a first scene generation image set.

[0095] In a third step, a pose extraction process is performed on each second scene image in the second scene image set to generate second scene pose information, thereby obtaining a second scene pose information set. The pose extraction process can be performed on each second scene image in the second scene image set to generate second scene pose information, thereby obtaining a second scene pose information set, using the preset pose extraction algorithm.

[0096] The fourth step is to input each second scene pose information in the above second scene pose information set into the above image generation model to generate a second scene generated image, thereby obtaining a second scene generated image set.

[0097] Optionally, before inputting each first scene generation image in the first scene generation image set into a pre-trained image mapping model to generate a first scene mapping image, the execution entity may further perform the following steps:

[0098] The first step is to obtain a sample collection image set. The sample collection image set can be obtained from a storage terminal. The sample collection images in the sample collection image set can be pre-captured scene images.

[0099] In the second step, based on the image generation model, a sample generation image corresponding to each sample acquisition image in the sample acquisition image set is generated to obtain the sample generation image set. The specific implementation method for generating the sample generation images and the resulting technical effects can be found in step 105 of the above embodiment and will not be further described here.

[0100] The third step is to select a target sample generated image from the above sample generated image set and perform the following second training sub-step:

[0101] In the first sub-step, the target sample generated image is input into the initial image mapping model to obtain an initial mapping image. A sample generated image can be randomly selected from the sample generated image set to serve as the target sample generated image. The initial image mapping model can be an untrained neural network model that takes the target sample generated image as input and outputs the initial mapping image.

[0102] As an example, the initial image mapping model may be a codec network model.

[0103] In the second sub-step, based on a preset mapping loss function, a mapping loss value between the sample acquisition image corresponding to the initial mapping image in the sample acquisition image set and the initial mapping image is determined.

[0104] As an example, the preset mapping loss function may be, but is not limited to, one of the following: a cross entropy loss function, a least squares function, or a classification cross entropy loss function.

[0105] In a third sub-step, in response to determining that the mapping loss value is less than a second target threshold, the initial image mapping model is determined as the image mapping model.

[0106] As an example, the second target threshold may be 0.01.

[0107] Optionally, the execution entity may further adjust relevant parameters in the initial image mapping model in response to determining that the mapping loss value is greater than or equal to a second target threshold, determine the adjusted initial image mapping model as the initial image mapping model, and select a target sample generation image from each of the unselected sample generation images in the sample generation image set for re-performing the second training step. The relevant parameters in the initial image mapping model may be adjusted using the preset adjustment algorithm.

[0108] Therefore, the image generation model can be used to initially generate a 3D feature image. Then, the image mapping model can be used to improve the image quality (peak signal-to-noise ratio). This improves the quality of the synthesized image and, in turn, the accuracy of scene recognition. This allows for timely detection of dangerous situations (like those about to fall) for patients, thereby enhancing the safety of intelligent walkers.

[0109] The above-mentioned embodiments of the present application have the following beneficial effects: Through the walking detection method based on an intelligent walker in some embodiments of the present application, the safety of the intelligent walker during use can be improved. Specifically, the reasons for the reduced safety of the intelligent walker during use are: on the one hand, its operation is not convenient enough. Whether it is the use of the walker or the operation of various detection instruments, it requires a certain amount of manpower and material resources, which places a considerable burden on medical staff and patients; on the other hand, the system is also lacking in systematization, and it is unable to timely and completely monitor the patient's muscle state in all directions, resulting in the possibility of missing some important information, which in turn affects the accurate diagnosis and effective intervention of the patient's condition. Based on this, the walking detection method based on an intelligent walker in some embodiments of the present application first converts the initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal of the target user collected within a preset time period into a signal sequence, thereby obtaining a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence. In this way, it is possible to collect signals representing the user's muscle state and convert the collected electrical signals into analyzable data information. Then, the walking muscle pressure value sequence, the walking muscle activity state information sequence and the walking motion parameter information sequence are integrated and processed to obtain walking muscle activity data. Thus, the information after conversion of each signal can be integrated to extract muscle feature information. Next, the walking muscle activity data is input into a pre-trained muscle feature extraction model to obtain walking muscle feature information. Thus, the user's current muscle state characteristics can be obtained. Then, the force difference value and muscle abnormality information corresponding to the walking muscle feature information and the pre-acquired initial muscle feature information are determined. Thus, the user's muscle state change information can be obtained. Finally, in response to determining that the force difference value is greater than the target threshold or the muscle abnormality information meets the preset abnormality condition, the preset mode switching signal is sent to the mobile controller to switch the intelligent walker operation mode. Thus, the operation state of the intelligent walker can be adjusted according to the user's muscle state change information. Therefore, some walking detection methods based on smart walkers in this application can embed the smart walker with monitoring test data acquisition software. During the rehabilitation training process, the walker structure is equipped with sensors and other electronic equipment to conduct real-time testing of the patient's condition, timely detect abnormal muscle loss in the patient, and reduce the number of times the patient falls due to muscle weakness, thereby improving the safety of the smart walker when in use.

[0110] Figure 2 A schematic diagram showing an application scenario of some embodiments of walking detection based on an intelligent walker according to the present application is shown. Figure 3 A schematic diagram illustrating an application scenario of other embodiments of walking detection based on an intelligent walker according to the present application is shown.

[0111] like Figure 2As shown, the execution body of the walking detection method based on the intelligent walker first performs signal conversion processing on the target user's initial walking pressure signal 202, initial walking electromyographic signal 204, and initial walking acceleration signal 206, which were previously collected within a preset time period, to obtain a walking muscle pressure value sequence 207, a walking muscle activity state information sequence 208, and a walking motion parameter information sequence 209. The initial walking pressure signal 202 is collected by the control pressure sensor 201, the initial walking electromyographic signal 204 is collected by the control electromyographic sensor 203, and the initial walking acceleration signal 206 is collected by the control acceleration sensor 205. Here, the initial walking pressure signal 202 can be "Pressure Signal 1." The initial walking electromyographic signal 204 can be "Electromyographic Signal 1." The initial walking acceleration signal 206 can be "Acceleration Signal 1." The walking muscle pressure values in the walking muscle pressure value sequence 207 can be "Muscle Pressure Value: 19.5 kg," "Muscle Pressure Value: 19.8 kg," or "Muscle Pressure Value: 20 kg." The walking muscle activity state information in the walking muscle activity state information sequence 208 may be "muscle fatigue information: fatigue occurs," "muscle fatigue information: fatigue occurs," or "muscle fatigue information: fatigue occurs." The walking motion parameter information in the walking motion parameter information sequence 209 may be "walking speed: 1 meter per second, cadence: 60, step size: 40 centimeters," "walking speed: 1 meter per second, cadence: 58, step size: 38 centimeters," or "walking speed: 1 meter per second, cadence: 59, step size: 41 centimeters." Then, the walking muscle pressure value sequence 207, the walking muscle activity state information sequence 208, and the walking motion parameter information sequence 209 are integrated to obtain walking muscle activity data 210. Here, the walking muscle activity data 210 may be "muscle pressure value: 20 kilograms, muscle fatigue information: fatigue occurs, walking speed: 1 meter per second, cadence: 60, step size: 40 centimeters." Next, the walking muscle activity data 210 is input into a pre-trained muscle feature extraction model to obtain walking muscle feature information 211. Here, the walking muscle characteristic information 211 may be "muscle characteristic B". Then, the strength difference value 213 and muscle abnormality information 214 corresponding to the above-mentioned walking muscle characteristic information 211 and the pre-acquired initial muscle characteristic information 212 are determined. Here, the initial muscle characteristic information 212 may be "initial muscle characteristic A". The strength difference value 213 may be "10". The muscle abnormality information 214 may be "muscle loss phenomenon". Finally, in response to determining that the above-mentioned strength difference value 213 is greater than the target threshold or the above-mentioned muscle abnormality information 214 meets the preset abnormal condition, the preset mode switching signal 215 is sent to the mobile controller to switch the intelligent walker operation mode. Here, the preset mode switching signal 215 may be "increasing the walking assistance level of the mobile controller".

[0112] Optionally, the execution entity may first synthesize individual walking scene images from a pre-captured walking scene image set of the target user to obtain a target walking scene image 307. The walking scene image set is captured by controlling a camera assembly. The camera assembly may include "Camera A" 301, "Camera B" 302, and "Camera C" 303. The walking scene images from the walking scene image set may include "Image A" 304, "Image B" 305, and "Image C" 306. The target walking scene image 307 may be a "synthesized image." Scene recognition processing is then performed on the target walking scene image 307 to obtain walking scene feature information 308. Here, walking scene feature information 308 may include "Identified scenario: user is falling." Finally, in response to determining that the walking scene feature information meets a preset dangerous scenario condition, a preset mode switching signal 309 is sent to the mobile controller to switch the intelligent walker's operating mode. Here, the preset mode switching signal 309 may include "mobile controller activates balance mode."

[0113] Further references Figure 4 As an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of a walking detection device based on an intelligent walker. These device embodiments are similar to Figure 1 Corresponding to the method embodiments shown, the walking detection device based on the intelligent walker can be specifically applied to various electronic devices.

[0114] like Figure 4As shown, some embodiments of the walking detection device 400 based on the intelligent walker include: a signal conversion unit 401, an integration unit 402, an input unit 403, a determination unit 404 and a sending unit 405. The signal conversion unit 401 is configured to perform signal conversion processing on the initial walking pressure signal, the initial walking electromyographic signal and the initial walking acceleration signal of the target user collected in advance within a preset time period, respectively, to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence and a walking motion parameter information sequence, wherein the above-mentioned initial walking pressure signal is collected by the control pressure sensor, the above-mentioned initial walking electromyographic signal is collected by the control electromyographic sensor, and the above-mentioned initial walking acceleration signal is collected by the control acceleration sensor; the integration unit 402 is configured to process the above-mentioned walking muscle pressure value sequence, the above-mentioned walking muscle activity state information sequence, and the above-mentioned walking muscle pressure value sequence. The column and the above-mentioned walking motion parameter information sequence are integrated and processed to obtain walking muscle activity data; the input unit 403 is configured to input the above-mentioned walking muscle activity data into a pre-trained muscle feature extraction model to obtain walking muscle feature information; the determination unit 404 is configured to determine the force difference value and muscle abnormality information corresponding to the above-mentioned walking muscle feature information and the pre-acquired initial muscle feature information; the sending unit 405 is configured to send a preset mode switching signal to the mobile controller to switch the intelligent walker operation mode in response to determining that the above-mentioned force difference value is greater than the target threshold or the above-mentioned muscle abnormality information meets the preset abnormality condition.

[0115] It is understandable that the units described in the walking detection device 400 based on the intelligent walker are similar to those described in the reference Figure 1 Therefore, the operations, features and beneficial effects described above for the method are also applicable to the walking detection device 400 based on the intelligent walker and the units included therein, and will not be repeated here.

[0116] This application also provides a computer device 500. Figure 5 As shown, computer device 500 includes: bus 501, processor 502, memory 503, and communication interface 504. Processor 502, memory 503, and communication interface 504 communicate with each other via bus 501. Computer device 500 can be a server or a terminal device. It should be understood that this application does not limit the number of processors and memories in computer device 500.

[0117] The bus 501 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 The bus 501 may include a path for transmitting information between various components of the computer device 500 (eg, memory 503, processor 502, and communication interface 504).

[0118] The processor 502 may include any one or more processors such as a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0119] The memory 503 may include a volatile memory, such as a random access memory (RAM). The memory 503 may also include a non-volatile memory, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

[0120] The memory 503 stores executable program code, which the processor 502 executes to implement the functions of the aforementioned signal conversion unit, integration unit, input unit, determination unit, and transmission unit, thereby implementing the aforementioned method for detecting walking using an intelligent walker. In other words, the memory 503 stores instructions for executing the aforementioned method for detecting walking using an intelligent walker.

[0121] The communication interface 504 uses a transceiver module such as, but not limited to, a network interface card or a transceiver to implement communication between the computer device 500 and other devices or a communication network.

[0122] An embodiment of the present application also provides a chip, which includes a processor and a data interface. The processor reads instructions stored in a memory through the data interface to execute the above-mentioned walking detection method based on an intelligent walker.

[0123] Embodiments of the present application also provide a computer-readable storage medium. This computer-readable storage medium can be any available medium capable of being stored by a computing device, or a data storage device such as a data center that contains one or more available media. This available medium can be a magnetic medium (e.g., a floppy disk, hard disk, or magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive). This computer-readable storage medium includes instructions that instruct the computing device to execute the aforementioned intelligent walker-based walking detection method.

[0124] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0125] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the protection scope of the technical solutions of the embodiments of the present application.

Claims

1. A walking detection method based on an intelligent walker, applied to the intelligent walker, the intelligent walker comprising: Pressure sensor, myoelectric sensor, acceleration sensor, motion controller, the method includes: Performing signal conversion processing on the target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal collected in advance within a preset time period to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence, wherein the initial walking pressure signal is collected by the control pressure sensor, the initial walking electromyographic signal is collected by the control electromyographic sensor, and the initial walking acceleration signal is collected by the control acceleration sensor; Integrating the walking muscle pressure value sequence, the walking muscle activity state information sequence, and the walking motion parameter information sequence to obtain walking muscle activity data; Inputting the walking muscle activity data into a pre-trained muscle feature extraction model to obtain walking muscle feature information, wherein the walking muscle feature information includes: walking muscle strength value and walking muscle content information, the walking muscle strength value represents the muscle strength of the target user at the current moment, and the walking muscle content information represents the muscle content of the target user at the current moment, and the muscle feature extraction model includes: a three-dimensional attention sub-model, an aggregation sub-model, a comprehensive feature extraction sub-model, a feature fusion sub-model, and a feature mapping sub-model; Determining a force difference value and muscle abnormality information corresponding to the walking muscle characteristic information and the pre-acquired initial muscle characteristic information; In response to determining that the force difference value is greater than a target threshold or the muscle abnormality information satisfies a preset abnormality condition, sending a preset mode switching signal to the mobile controller to switch the intelligent walker operation mode; The intelligent walking aid further comprises: a camera component, and the method further comprises: synthesizing each walking scene image in a pre-taken walking scene image set of a target user to obtain a target walking scene image, wherein the walking scene image set is taken by controlling the camera assembly; Performing scene recognition processing on the target walking scene image to obtain walking scene feature information; In response to determining that the walking scene characteristic information meets the preset dangerous scene condition, a preset mode switching signal is sent to the mobile controller to switch the intelligent walker operation mode.

2. The walking detection method based on the intelligent walker according to claim 1, wherein: The signal conversion processing is performed on the target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal collected in advance within a preset time period to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence, including: Preprocessing the target user's initial walking pressure signal, initial walking electromyography signal, and initial walking acceleration signal collected in advance within a preset time period to obtain a walking pressure filter signal, a walking electromyography filter signal, and a walking acceleration filter signal; performing pressure conversion processing on the walking pressure filter signal to obtain a walking muscle pressure value sequence; Performing feature extraction processing on the walking electromyographic filter signal to obtain a walking muscle activity state information sequence; Splitting the walking acceleration filter signal to obtain a walking acceleration sub-signal sequence; For each walking acceleration sub-signal sequence in the walking acceleration sub-signal sequence, the following detection steps are performed: Performing gait detection processing on the walking acceleration sub-signal to obtain a walking speed value, a walking frequency value, and a walking stride value; fusing the walking speed value, the walking frequency value, and the walking stride value to obtain walking motion parameter information; The generated pieces of walking motion parameter information are determined as a walking motion parameter information sequence.

3. The walking detection method based on the intelligent walker according to claim 1, wherein: Before inputting the walking muscle activity data into a pre-trained muscle feature extraction model to obtain walking muscle feature information, the method further includes: Obtain a sample walking muscle activity dataset; Select target sample walking muscle activity data from the sample walking muscle activity data set and perform the following first training step: Inputting the target sample walking muscle activity data into the three-dimensional feature extraction sub-model included in the initial muscle feature extraction model to obtain initial muscle three-dimensional feature information, wherein the initial muscle feature extraction model also includes: a three-dimensional attention sub-model, an aggregation sub-model, a comprehensive feature extraction sub-model, a feature fusion sub-model and a feature mapping sub-model; Inputting the initial muscle three-dimensional feature information into the three-dimensional attention sub-model included in the initial muscle feature extraction model to obtain the initial muscle three-dimensional attention information; Inputting the initial muscle three-dimensional attention information into the aggregation sub-model included in the initial muscle feature extraction model to obtain initial muscle aggregation information; Inputting the initial muscle aggregation information into the comprehensive feature extraction sub-model included in the initial muscle feature extraction model to obtain the initial muscle comprehensive feature information; Inputting the initial muscle comprehensive feature information into the feature fusion sub-model included in the initial muscle feature extraction model to obtain initial muscle feature fusion information; Inputting the initial muscle feature fusion information into the feature mapping sub-model included in the initial muscle feature extraction model to obtain initial muscle feature information; Determining a feature extraction loss value corresponding to the initial muscle feature information based on at least one preset feature extraction loss function; In response to determining that the feature extraction loss value is less than the first target threshold, the initial muscle feature extraction model is determined as the muscle feature extraction model.

4. The walking detection method based on the intelligent walker according to claim 3, wherein: The method further comprises: In response to determining that the feature extraction loss value is greater than or equal to the first target threshold, the relevant parameters in the initial muscle feature extraction model are adjusted, the adjusted initial muscle feature extraction model is determined as the initial muscle feature extraction model, and the target sample walking muscle activity data is selected from the unselected sample walking muscle activity data in the sample walking muscle activity data set for re-execution of the first training step.

5. A walking detection device based on an intelligent walker, applied to the intelligent walker, the intelligent walker comprising: Pressure sensor, myoelectric sensor, acceleration sensor, motion controller, camera assembly, the device includes: a signal conversion unit configured to perform signal conversion processing on the target user's initial walking pressure signal, initial walking electromyographic signal, and initial walking acceleration signal collected in advance within a preset time period, respectively, to obtain a walking muscle pressure value sequence, a walking muscle activity state information sequence, and a walking motion parameter information sequence, wherein the initial walking pressure signal is collected by the control pressure sensor, the initial walking electromyographic signal is collected by the control electromyographic sensor, and the initial walking acceleration signal is collected by the control acceleration sensor; an integration unit configured to integrate the walking muscle pressure value sequence, the walking muscle activity state information sequence, and the walking motion parameter information sequence to obtain walking muscle activity data; an input unit configured to input the walking muscle activity data into a pre-trained muscle feature extraction model to obtain walking muscle feature information, wherein the walking muscle feature information includes: a walking muscle strength value and walking muscle content information, the walking muscle strength value represents the muscle strength of the target user at the current moment, and the walking muscle content information represents the muscle content of the target user at the current moment, and the muscle feature extraction model includes: a three-dimensional attention sub-model, an aggregation sub-model, a comprehensive feature extraction sub-model, a feature fusion sub-model, and a feature mapping sub-model; a determining unit configured to determine a strength difference value and muscle abnormality information corresponding to the walking muscle characteristic information and the pre-acquired initial muscle characteristic information; a sending unit configured to send a preset mode switching signal to the mobile controller to switch the operation mode of the intelligent walker in response to determining that the force difference value is greater than a target threshold or the muscle abnormality information satisfies a preset abnormality condition; a synthesis unit configured to synthesize each walking scene image in a set of walking scene images of a target user captured in advance, to obtain a target walking scene image, wherein the walking scene image set is captured by controlling the camera assembly; a recognition unit configured to perform scene recognition processing on the target walking scene image to obtain walking scene feature information; The signal sending unit is configured to send a preset mode switching signal to the mobile controller to switch the operation mode of the intelligent walker in response to determining that the walking scene characteristic information meets the preset dangerous scene condition.

6. A computer device, wherein: The computer device comprises a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium, wherein: The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images