Child safety belt intelligent adjustment method and system based on machine vision, and medium

Abstract

The application discloses a child safety belt intelligent adjusting method and system based on machine vision and a medium. The method comprises the following steps: acquiring face image information, body posture information, pressure distribution information and electrocardiogram information of a child passenger; extracting expression features of the child passenger according to the face image information, extracting posture features of the child passenger according to the body posture information and the pressure distribution information, and extracting physiological features of the child passenger according to the electrocardiogram information; inputting the expression features, the posture features and the physiological features into a pre-trained comfort recognition model to obtain the ride comfort of the child passenger; acquiring real-time vehicle states of a current vehicle, and adjusting the locking force of the safety belt of the child passenger according to the ride comfort and the real-time vehicle states. The application can adjust the locking force of the child safety belt in real time and adaptively, reduces the discomfort of the child wearing the safety belt, improves the ride experience and the ride safety of the child, and can be widely applied to the technical field of vehicle control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle control, and in particular to a child safety belt intelligent adjustment method and system based on machine vision and a medium. BACKGROUND

[0002] A safety belt is a vehicle active safety device that reduces collision damage by limiting the movement range of the occupant, is made of high-strength materials such as polyester fibers, contains components such as a belt body, a buckle, and a tension mechanism, and reduces the risk of secondary impact by restraining the body of the occupant, thereby reducing the probability of injury to key parts such as the ribs and head. Therefore, in a driving scenario, it is crucial to use the safety belt correctly.

[0003] Because the safety belt has a certain restraint on the human body, children are prone to discomfort after wearing the safety belt for a long time, and during driving, parents often need to focus on driving and cannot help children adjust the locking force of the safety belt, and children self-adjusting may result in incorrect wearing of the safety belt, affecting the safety of the ride.

[0004] The above problems need to be solved. SUMMARY

[0005] The present application aims to at least partially solve one of the problems in the prior art.

[0006] To this end, one object of the present application is to provide a child safety belt intelligent adjustment method based on machine vision, which can adjust the locking force of the child safety belt in real time and adaptively, reduce the discomfort of the child wearing the safety belt, and improve the ride experience and safety of the child.

[0007] Another object of the present application is to provide a child safety belt intelligent adjustment system based on machine vision.

[0008] In order to achieve the above technical purpose, the technical solution adopted by the present application comprises:

[0009] In a first aspect, the present application provides a child safety belt intelligent adjustment method based on machine vision, comprising the following steps:

[0010] Obtaining facial image information, body posture information, pressure distribution information, and electrocardiogram information of a child occupant;

[0011] Extracting facial expression features of the child occupant according to the facial image information, extracting posture features of the child occupant according to the body posture information and the pressure distribution information, and extracting physiological features of the child occupant according to the electrocardiogram information;

[0012] input the expression feature, the posture feature and the physiological feature into a pre-trained comfort level recognition model to obtain the ride comfort level of the child passenger;

[0013] obtain real-time vehicle status of the current vehicle, and adjust the locking force of the safety belt of the child passenger according to the ride comfort level and the real-time vehicle status.

[0014] Further, in an embodiment of the present application, the obtaining of the face image information, the body posture information, the pressure distribution information and the electrocardiogram information of the child passenger specifically comprises:

[0015] obtaining the face image information through a vehicle-mounted camera;

[0016] obtaining the body posture information of the child passenger by means of a seat-embedded millimeter wave radar to obtain the coordinates of multiple body key points of the child passenger;

[0017] obtaining the pressure distribution information of the back and legs of the child passenger by means of a flexible piezoelectric sensor matrix integrated on the surface of the seat;

[0018] obtaining the electrocardiogram information by means of a non-contact electrocardiogram sensor.

[0019] Further, in an embodiment of the present application, the extracting of the expression feature of the child passenger according to the face image information specifically comprises:

[0020] performing face detection and alignment on the face image information by means of a pre-set face detection model to obtain a face feature image;

[0021] performing expression recognition on the face feature image to obtain a pain level and / or a restlessness level of the child passenger;

[0022] determining the expression feature according to the pain level and / or the restlessness level.

[0023] Further, in an embodiment of the present application, the extracting of the posture feature of the child passenger according to the body posture information and the pressure distribution information specifically comprises:

[0024] determining a trunk inclination and a head offset of the child passenger according to the body posture information;

[0025] determining a center of gravity offset of the child passenger according to the pressure distribution information;

[0026] determining the posture feature according to the trunk inclination, the head offset and the center of gravity offset.

[0027] Further, in an embodiment of the present application, the physiological characteristics of the child passenger are extracted according to the electrocardiogram information, which specifically includes:

[0028] R-wave detection is performed on the electrocardiogram information to obtain an RR interval sequence;

[0029] Frequency domain analysis is performed on the RR interval sequence to obtain a low-frequency part and a high-frequency part;

[0030] The HRV low / high frequency ratio is determined according to the low-frequency part and the high-frequency part, and the HRV low / high frequency ratio is taken as the physiological characteristic.

[0031] Further, in an embodiment of the present application, the comfort recognition model is obtained by training through the following steps:

[0032] Sample expression characteristics, sample posture characteristics and sample physiological characteristics of a test child passenger when riding a test vehicle are obtained, and corresponding comfort labels are determined by manual labeling;

[0033] Training samples are determined according to the sample expression characteristics, the sample posture characteristics and the sample physiological characteristics;

[0034] The training samples are input into a pre-constructed deep learning neural network to obtain comfort recognition results;

[0035] A loss value is determined according to the comfort recognition results and the comfort labels;

[0036] The parameters of the deep learning neural network are updated according to the loss value to obtain the trained comfort recognition model.

[0037] Further, in an embodiment of the present application, the real-time vehicle state includes vehicle speed and vehicle acceleration, and the adjustment of the locking force of the safety belt of the child passenger according to the ride comfort and the real-time vehicle state specifically includes:

[0038] When the absolute value of the vehicle acceleration is greater than or equal to a preset first threshold value, a preset first locking force adjustment range and a first locking force median value corresponding to the first locking force adjustment range are obtained, a target adjustment degree is determined according to the difference ratio between the ride comfort and a preset comfort threshold value, and a corresponding first target locking force is determined in the first locking force adjustment range according to the target adjustment degree and the first locking force median value, and then the locking force of the safety belt of the child passenger is adjusted according to the first target locking force;

[0039] When the absolute value of the vehicle acceleration is less than the first threshold value and the vehicle speed is greater than or equal to a preset second threshold value, a preset second locking force adjustment range and a second locking force median value corresponding to the second locking force adjustment range are obtained, a target adjustment degree is determined according to a difference ratio of the ride comfort and the comfort threshold value, and then a corresponding second target locking force is determined in the second locking force adjustment range according to the target adjustment degree and the second locking force median value, and then the locking force of the safety belt of the child occupant is adjusted according to the second target locking force.

[0040] When the absolute value of the vehicle acceleration is less than the first threshold value and the vehicle speed is less than the second threshold value, a preset third locking force adjustment range and a third locking force median value corresponding to the third locking force adjustment range are obtained, a target adjustment degree is determined according to a difference ratio of the ride comfort and the comfort threshold value, and then a corresponding third target locking force is determined in the third locking force adjustment range according to the target adjustment degree and the third locking force median value, and then the locking force of the safety belt of the child occupant is adjusted according to the third target locking force.

[0041] In a second aspect, an embodiment of the present application provides a child safety belt intelligent adjustment system based on machine vision, comprising:

[0042] An information acquisition module is configured to acquire face image information, body posture information, pressure distribution information, and electrocardiogram information of a child occupant.

[0043] A feature extraction module is configured to extract expression features of the child occupant according to the face image information, extract posture features of the child occupant according to the body posture information and the pressure distribution information, and extract physiological features of the child occupant according to the electrocardiogram information.

[0044] A comfort degree identification module is configured to input the expression features, the posture features, and the physiological features into a pre-trained comfort degree identification model to obtain a ride comfort degree of the child occupant.

[0045] A locking force adjustment module is configured to acquire a real-time vehicle state of a current vehicle, and adjust a locking force of a safety belt of the child occupant according to the ride comfort degree and the real-time vehicle state.

[0046] In a third aspect, an embodiment of the present application provides a child safety belt intelligent adjustment device based on machine vision, comprising:

[0047] At least one processor;

[0048] At least one memory configured to store at least one program;

[0049] When the at least one program is executed by the at least one processor, the at least one processor implements the machine vision-based child safety belt intelligent adjustment method.

[0050] In a fourth aspect, an embodiment of the present application further provides a computer readable storage medium, wherein a processor executable program is stored, and the processor executable program is used for executing the machine vision-based child safety belt intelligent adjustment method when executed by a processor.

[0051] The advantages and beneficial effects of the present application will be partially given in the following description, partially will become obvious from the following description, or will be understood through the practice of the present application:

[0052] In the present application, the face image information, the body posture information, the pressure distribution information and the electrocardiogram information of the child passenger are acquired, the expression feature of the child passenger is extracted according to the face image information, the posture feature of the child passenger is extracted according to the body posture information and the pressure distribution information, and the physiological feature of the child passenger is extracted according to the electrocardiogram information, the expression feature, the posture feature and the physiological feature are input into the pre-trained comfort recognition model to obtain the ride comfort of the child passenger, the real-time vehicle state of the current vehicle is acquired, and the locking force of the safety belt of the child passenger is adjusted according to the ride comfort and the real-time vehicle state. In the present application, the expression feature, the posture feature and the physiological feature of the child passenger are extracted in real time based on the face image information, the body posture information, the pressure distribution information and the electrocardiogram information, and then input into the pre-trained comfort recognition model to obtain the ride comfort of the child passenger, and the locking force of the safety belt of the child passenger is adjusted according to the ride comfort and the real-time vehicle state, so that the locking force of the child safety belt can be adjusted in real time and adaptively, the discomfort of the child wearing the safety belt is reduced, and the ride experience and the ride safety of the child are improved. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following introduces the drawings needed to be used in the embodiments of the present application as follows. It should be understood that the drawings introduced in the following merely for the convenience of clearly describing part of the embodiments of the technical solutions of the present application, and for those skilled in the art, other drawings can also be obtained without paying creative labor on the premise.

[0054] Figure 1 A step flow chart of a machine vision-based child safety belt intelligent adjustment method provided by an embodiment of the present application is provided.

[0055] Figure 2 A structural block diagram of a machine vision-based child safety belt intelligent adjustment system provided by an embodiment of the present application is provided.

[0056] Figure 3 A structural block diagram of a child safety belt intelligent adjusting device based on machine vision is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0057] Embodiments of the present application are described in detail below with reference to examples shown in the attached drawings, wherein the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are only used to explain the present application and cannot be understood as a limitation of the present application. For the step numbers in the following embodiments, they are only set for the convenience of explanation and description, and the order between the steps is not limited in any way, and the execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0058] In the description of the present application, the meaning of multiple is two or more, and if the first, the second is described, it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features or implicitly indicating the sequence of indicated technical features. In addition, unless otherwise defined, all technical and scientific terms used herein have the same meaning as generally understood by those skilled in the art.

[0059] Reference Figure 1 , the embodiment of the present application provides a child safety belt intelligent adjusting method based on machine vision, which specifically comprises the following steps:

[0060] S101, obtaining face image information, body posture information, pressure distribution information and electrocardiogram information of a child passenger;

[0061] S102, extracting expression features of the child passenger according to the face image information, extracting posture features of the child passenger according to the body posture information and the pressure distribution information, and extracting physiological features of the child passenger according to the electrocardiogram information;

[0062] S103, inputting the expression features, the posture features and the physiological features into a pre-trained comfort recognition model to obtain the ride comfort of the child passenger;

[0063] S103, obtaining real-time vehicle state of the current vehicle, and adjusting the locking force of the safety belt of the child passenger according to the ride comfort and the real-time vehicle state.

[0064] The embodiment of the present application extracts the expression feature, the posture feature and the physiological feature of the child passenger in real time based on the face image information, the body posture information, the pressure distribution information and the electrocardiogram information, and then inputs them into a pre-trained comfort recognition model to obtain the ride comfort of the child passenger, and adjusts the locking force of the safety belt of the child passenger according to the ride comfort and the real-time vehicle state, so that the locking force of the child safety belt can be adjusted in real time and adaptively, the discomfort of the child wearing the safety belt is reduced, and the ride experience and the ride safety of the child are improved.

[0065] Further, as an optional implementation, the face image information, the body posture information, the pressure distribution information and the electrocardiogram information of the child passenger are obtained, which specifically include:

[0066] S1011, acquiring face image information through a vehicle-mounted camera;

[0067] S1012, acquiring multiple human body key point coordinates of the child passenger through a seat built-in millimeter wave radar to obtain body posture information;

[0068] S1013, acquiring the pressure distribution condition of the back and legs of the child passenger through a flexible piezoelectric sensor matrix integrated on the surface of the seat to obtain pressure distribution information;

[0069] S1014, acquiring electrocardiogram information through a non-contact electrocardiograph sensor.

[0070] Specifically, the face image information of the child passenger is collected in real time through an in-vehicle wide-angle camera (such as a fisheye lens); the posture data of the child passenger is collected in real time through a seat built-in millimeter wave radar, and 17 key point coordinates are extracted by using a posture estimation algorithm to obtain body posture information; a flexible piezoelectric sensor matrix (100x100 grid) is integrated on the surface of the seat to monitor the pressure distribution of the back and legs in real time to obtain pressure distribution information; and the electrocardiogram information of the child passenger is monitored through a non-contact electrocardiograph sensor.

[0071] Further, as an optional implementation, the expression feature of the child passenger is extracted according to the face image information, which specifically includes:

[0072] S1021, face detection and alignment of the face image information are performed through a pre-set face detection model to obtain a face feature image;

[0073] S1022, expression recognition is performed on the face feature image to obtain a pain level and / or a restlessness level of the child passenger;

[0074] S1023, the expression feature is determined according to the pain level and / or the restlessness level.

[0075] Specifically, a lightweight model (such as MobileNetV3) is used to detect and align the face in the image information collected in the foregoing steps, excluding light, and shielding interference, to obtain a facial feature image; an expression recognition model is used to recognize the facial feature image, and in combination with the Wong-Baker pain scale (6-level classification) and a child emotion database, to recognize discomfort expression features such as frowning and drooping corners of the mouth, and to quantify the levels of “pain” and / or “irritation”; the pain level and / or the irritation level are standardized to obtain expression features, such as a pain level of 3 and an irritation level of 1, which can be represented as a two-dimensional vector (0.3, 0.1).

[0076] Further as an optional implementation, the posture feature of the child occupant is extracted according to the human posture information and the pressure distribution information, which specifically includes:

[0077] S1024, determining a trunk inclination degree and a head offset degree of the child occupant according to the human posture information;

[0078] S1025, determining a gravity center offset of the child occupant;

[0079] S1026, determining the posture feature according to the trunk inclination degree, the head offset degree, and the gravity center offset.

[0080] Specifically, the trunk inclination degree, the head offset, and the like are calculated according to the human posture information, which can be used to judge the abnormal sitting posture (such as curling up and leaning to one side); the gravity center offset is recognized according to the pressure distribution information; the trunk inclination degree, the head offset, and the gravity center offset are standardized to obtain the posture feature, such as a trunk inclination degree of 10°, a head offset of 5°, and a gravity center offset of 10 cm, which can be represented as a three-dimensional vector (π / 18, π / 36, 0.1).

[0081] Further as an optional implementation, the physiological feature of the child occupant is extracted according to the electrocardiogram information, which specifically includes:

[0082] S1027, performing R-wave detection on the electrocardiogram information to obtain an RR interval sequence;

[0083] S1028, performing frequency domain analysis on the RR interval sequence to obtain a low-frequency part and a high-frequency part;

[0084] S1029, determining an HRV low / high frequency ratio according to the low-frequency part and the high-frequency part, and taking the HRV low / high frequency ratio as the physiological feature.

[0085] Specifically, the R-wave peak (the highest point of the QRS complex) in the electrocardiogram is automatically identified by an algorithm to generate a RR interval sequence (the time interval of adjacent R waves), and the error needs to be controlled within ±5 ms; the RR interval sequence is analyzed in the frequency domain, and the RR interval sequence is decomposed into low-frequency components (LF, 0.04-0.15 Hz) and high-frequency components (HF, 0.15-0.4 Hz) by Fourier transform, wherein the low-frequency components reflect the combined action of sympathetic nerves and vagus nerves (normal value 1170±416 ms 2 ), and the high-frequency components represent the independent regulation of vagus nerves (normal value 975±203 ms 2 ); the HRV low / high frequency ratio is determined according to the low-frequency part and the high-frequency part, and the HRV low / high frequency ratio is taken as a physiological feature, and the HRV low / high frequency ratio can reflect the activity degree of sympathetic nerves.

[0086] Further as an optional implementation manner, the comfort recognition model is obtained by the following steps:

[0087] S201, sample expression features, sample posture features and sample physiological features of a test child when riding a test vehicle are obtained, and corresponding comfort labels are determined by manual labeling;

[0088] S202, training samples are determined according to the sample expression features, sample posture features and sample physiological features;

[0089] S203, the training samples are input into a pre-constructed deep learning neural network to obtain a comfort recognition result;

[0090] S204, a loss value is determined according to the comfort recognition result and the comfort label;

[0091] S205, the parameters of the deep learning neural network are updated according to the loss value to obtain a trained comfort recognition model.

[0092] Specifically, the input data of the comfort recognition model includes sample expression features, sample posture features and sample physiological features of a test child when riding a test vehicle, the model architecture can adopt a deep learning neural network, and the model output is a ride comfort level. The sample expression features, sample posture features and sample physiological features of the test child when riding the test vehicle are obtained, and the corresponding comfort labels are determined by manual labeling, and then the sample expression features, sample posture features and sample physiological features are input into the deep learning neural network as training samples for training.

[0093] After the training sample is input into the initialized deep learning neural network, a recognition result of the model output, i.e., a comfort level recognition result, can be obtained, and the accuracy of the model recognition can be evaluated according to the comfort level recognition result and the aforementioned comfort level label, so as to update the parameters of the model. For the comfort level recognition model, the accuracy of the model recognition result can be measured by a loss function, which is defined on a single training data and is used to measure the prediction error of a training data. Specifically, the loss value of the training data is determined by the label of the single training data and the prediction result of the model for the training data. In actual training, there are many training data in a training data set, so a cost function is generally used to measure the overall error of the training data set. The cost function is defined on the entire training data set and is used to calculate the average value of the prediction errors of all training data, which can better measure the prediction effect of the model. For a general machine learning model, based on the aforementioned cost function, plus a regular term that measures the complexity of the model, the loss value of the entire training data set can be obtained based on the objective function. There are many commonly used loss functions, such as 0-1 loss function, square loss function, absolute loss function, logarithmic loss function, cross-entropy loss function, etc., which can be used as the loss function of the machine learning model. Here, they will not be described one by one. In the embodiment of the present application, any one of the loss functions can be selected to determine the loss value of the training. Based on the loss value of the training, the parameters of the model are updated by using the back propagation algorithm, and after several iterations, a trained comfort level recognition model can be obtained. Specifically, the number of iterations can be preset, or the training can be considered to be completed when the test set reaches the accuracy requirement.

[0094] The facial feature, the posture feature and the physiological feature of the child occupant are input into the trained comfort level recognition model, so as to obtain the ride comfort level of the child occupant inferred by the model.

[0095] Further, as an optional implementation, the real-time vehicle state includes vehicle speed and vehicle acceleration, and the locking force of the safety belt of the child occupant is adjusted according to the ride comfort level and the real-time vehicle state, which specifically includes:

[0096] In S1041, when the absolute value of the vehicle acceleration is greater than or equal to a preset first threshold value, a preset first locking force adjustment range and a first locking force median value corresponding to the first locking force adjustment range are obtained, a target adjustment degree is determined according to the difference ratio of the ride comfort level and the preset comfort level threshold value, a corresponding first target locking force is determined in the first locking force adjustment range according to the target adjustment degree and the first locking force median value, and then the locking force of the safety belt of the child occupant is adjusted according to the first target locking force.

[0097] S1042, when the absolute value of the vehicle acceleration is less than the first threshold value and the vehicle speed is greater than or equal to a preset second threshold value, a preset second locking force adjustment range and a second locking force median value corresponding to the second locking force adjustment range are obtained, a target adjustment degree is determined according to the difference ratio of the ride comfort and the comfort threshold value, then a corresponding second target locking force is determined in the second locking force adjustment range according to the target adjustment degree and the second locking force median value, and then the locking force of the safety belt of the child occupant is adjusted according to the second target locking force.

[0098] S1043, when the absolute value of the vehicle acceleration is less than the first threshold value and the vehicle speed is less than the second threshold value, a preset third locking force adjustment range and a third locking force median value corresponding to the third locking force adjustment range are obtained, a target adjustment degree is determined according to the difference ratio of the ride comfort and the comfort threshold value, then a corresponding third target locking force is determined in the third locking force adjustment range according to the target adjustment degree and the third locking force median value, and then the locking force of the safety belt of the child occupant is adjusted according to the third target locking force.

[0099] Specifically, the embodiment of the present application divides the first locking force adjustment range, the second locking force adjustment range and the third locking force adjustment range in advance for different vehicle accelerations and vehicle speeds, for example, when the vehicle acceleration is greater than or equal to 0.4g (the first threshold value), it indicates that the vehicle is in the state of rapid acceleration / braking, at this time, in order to suppress the inertial displacement, the corresponding first locking force adjustment range is set to 30N to 50N, and the corresponding first locking force median value is 40N, when the vehicle acceleration is less than 0.4g and the vehicle speed is greater than or equal to 60km / h (the second threshold value), it indicates that the vehicle is in the state of high-speed driving, at this time, in order to protect the safety of the occupant, the corresponding second locking force adjustment range is set to 10N to 30N, and the corresponding second locking force median value is 20N, when the vehicle acceleration is less than 0.4g and the vehicle speed is less than 60km / h, it indicates that the vehicle is in the state of medium-low speed driving, at this time, the corresponding third locking force adjustment range is set to 5N to 10N, and the corresponding third locking force median value is 7.5N.

[0100] After determining the current adjustable locking force adjustment range and the corresponding locking force median value, the target adjustment degree is determined according to the difference ratio of the ride comfort and the comfort threshold value, then the corresponding target locking force is determined in the locking force adjustment range according to the target adjustment degree and the locking force median value, and the locking force of the safety belt of the child occupant is adjusted according to the determined target locking force.

[0101] Taking the first locking force adjustment range 30N to 50N, the first locking force median value 40N as an example, if the riding comfort score of the identified child occupant is 40, and the comfort threshold is 50, then the difference ratio of the riding comfort and the comfort threshold is -1 / 5, that is, the target adjustment degree is -1 / 5 (positive or negative indicates upward or downward adjustment), so the target locking force is 38N obtained by downward adjustment of 1 / 5 based on the reference value 40N in the range of 30N to 50N.

[0102] The method steps of the embodiments of the present application are described above. It can be understood that the embodiments of the present application extract the expression features, posture features and physiological features of the child occupant in real time based on the face image information, body posture information, pressure distribution information and electrocardiogram information, and then input them into the pre-trained comfort recognition model to obtain the riding comfort of the child occupant. The locking force of the child safety belt is adjusted according to the riding comfort and the real-time vehicle state, which can adjust the locking force of the child safety belt in real time and adaptively, reduce the discomfort of the child wearing the safety belt, and improve the riding experience and safety of the child.

[0103] Referring to Figure 2 The embodiments of the present application provide a child safety belt intelligent adjustment system based on machine vision, comprising:

[0104] An information acquisition module is configured to acquire face image information, body posture information, pressure distribution information and electrocardiogram information of a child occupant.

[0105] A feature extraction module is configured to extract expression features of the child occupant according to the face image information, extract posture features of the child occupant according to the body posture information and the pressure distribution information, and extract physiological features of the child occupant according to the electrocardiogram information.

[0106] A comfort recognition module is configured to input the expression features, posture features and physiological features into a pre-trained comfort recognition model to obtain the riding comfort of the child occupant.

[0107] A locking force adjustment module is configured to acquire a real-time vehicle state of a current vehicle, and adjust the locking force of the safety belt of the child occupant according to the riding comfort and the real-time vehicle state.

[0108] The contents in the above method embodiments are applicable to the system embodiments. The system embodiments specifically realize the same functions as the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0109] Referring to Figure 3 The embodiments of the present application provide a child safety belt intelligent adjustment device based on machine vision, comprising:

[0110] at least one processor;

[0111] at least one memory for storing at least one program;

[0112] The at least one program, when executed by the at least one processor, causes the at least one processor to implement the machine vision-based child safety belt intelligent adjustment method.

[0113] The contents in the method embodiments are applicable to the device embodiments, the device embodiments specifically implement the functions same as the method embodiments, and achieve the beneficial effects same as the method embodiments.

[0114] The embodiment of the present application further provides a computer readable storage medium, wherein a program executable by a processor is stored, and the program executable by the processor is used for executing the machine vision-based child safety belt intelligent adjustment method when executed by the processor.

[0115] The computer readable storage medium of the embodiment of the present application can execute the machine vision-based child safety belt intelligent adjustment method provided by the method embodiment of the present application, execute the steps of any combination of the method embodiments, and has the corresponding functions and beneficial effects of the method.

[0116] The embodiment of the present application further discloses a computer program product or a computer program, and the computer program product or the computer program includes computer instructions stored in a computer readable storage medium. A processor of a computer device can read the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method shown in the embodiment of the present application. Figure 1 The method shown in the embodiment of the present application.

[0117] In some alternative embodiments, the functions / operations mentioned in the block diagram can not occur in the order mentioned in the operation diagram. For example, two blocks shown in succession can actually be executed substantially simultaneously or the blocks mentioned above can be executed in reverse order depending on the functions / operations involved. In addition, the embodiments presented and described in the flowcharts of the present application are provided by way of example, and the purpose is to provide a more comprehensive understanding of the technology. The disclosed method is not limited to the operations and logical flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and in which sub-operations described as part of larger operations are independently executed.

[0118] Furthermore, although the present application is described in the context of functional modules, it is to be understood that one or more of the functions and / or features described above can be integrated in a single physical device and / or software module, or one or more functions and / or features can be implemented in separate physical devices or software modules. It will also be appreciated that detailed discussion of the actual implementation of each module is unnecessary to an understanding of the present application. Rather, the actual implementation of the modules, in conjunction with their attributes, functions, and internal relationships, are to be understood within the context of the devices disclosed herein. Thus, those skilled in the art with access to the teachings presented herein will be able to devise suitable implementations of the present application without undue experimentation. It is also to be understood that the particular concepts disclosed are merely illustrative and are not intended to limit the scope of the present application, which is defined by the appended claims and equivalents thereof.

[0119] If the above functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0120] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a list of executable instructions for implementing logic functions, which can be specifically embodied in any computer readable medium for use by an instruction execution system, device or apparatus, such as a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, device or apparatus, or in conjunction with these instructions execution system, device or apparatus. For the purpose of this specification, "computer readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by an instruction execution system, device or apparatus, or in conjunction with these instruction execution system, device or apparatus.

[0121] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example, via optical scanning of the paper or other medium, then compiled, interpreted, or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.

[0122] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the above described embodiments, various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any of the following technologies, or combinations thereof, can be used with the necessary modifications: a discrete logic circuit(s) having logic gates for implementing logic functions upon an application of data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array(s) (PGA), a field programmable gate array (FPGA), and / or the like.

[0123] In the above description of the present specification, reference to the description of the terms "one embodiment", "another embodiment", or "certain embodiments" or the like means that a particular feature, structure, material or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. The illustrative appearances of the above described terms in the present specification are not necessarily referred to the same embodiment or example. Also, the particular features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0124] Although embodiments of the present application have been shown and described, it would be recognized by those of ordinary skill in the art that various changes, modifications, alternatives, and variations can be made to the embodiments without departing from the spirit and scope of the application, which is defined by the claims and their equivalents.

[0125] The above is a specific description of the preferred embodiments of the present application, but the present application is not limited to the above-described embodiments, and those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present application, and these equivalent modifications or substitutions are included in the scope defined by the claims of the present application.

Claims

1. A method for intelligent adjustment of a child safety belt based on machine vision, characterized in that, The method comprises the following steps: obtaining facial image information, body posture information, pressure distribution information and electrocardiogram information of a child passenger; extracting expression features of the child passenger according to the facial image information, extracting posture features of the child passenger according to the body posture information and the pressure distribution information, and extracting physiological features of the child passenger according to the electrocardiogram information; inputting the expression features, the posture features and the physiological features into a pre-trained comfort level recognition model to obtain a ride comfort level of the child passenger; obtaining a real-time vehicle state of a current vehicle, and adjusting a locking force of a safety belt of the child passenger according to the ride comfort level and the real-time vehicle state; the real-time vehicle state comprises a vehicle speed and a vehicle acceleration, and the adjusting of the locking force of the safety belt of the child passenger according to the ride comfort level and the real-time vehicle state specifically comprises: when the absolute value of the vehicle acceleration is greater than or equal to a preset first threshold value, obtaining a preset first locking force adjustment range and a first locking force median value corresponding to the first locking force adjustment range, determining a target adjustment degree according to a difference ratio of the ride comfort level and a preset comfort level threshold value, and determining a corresponding first target locking force in the first locking force adjustment range according to the target adjustment degree and the first locking force median value, and then adjusting the locking force of the safety belt of the child passenger according to the first target locking force; when the absolute value of the vehicle acceleration is less than the first threshold value, and the vehicle speed is greater than or equal to a preset second threshold value, obtaining a preset second locking force adjustment range and a second locking force median value corresponding to the second locking force adjustment range, determining a target adjustment degree according to a difference ratio of the ride comfort level and the comfort level threshold value, and then determining a corresponding second target locking force in the second locking force adjustment range according to the target adjustment degree and the second locking force median value, and then adjusting the locking force of the safety belt of the child passenger according to the second target locking force; when the absolute value of the vehicle acceleration is less than the first threshold value, and the vehicle speed is less than the second threshold value, obtaining a preset third locking force adjustment range and a third locking force median value corresponding to the third locking force adjustment range, determining a target adjustment degree according to a difference ratio of the ride comfort level and the comfort level threshold value, and then determining a corresponding third target locking force in the third locking force adjustment range according to the target adjustment degree and the third locking force median value, and then adjusting the locking force of the safety belt of the child passenger according to the third target locking force.

2. The method of claim 1, wherein, The obtaining of the facial image information, the body posture information, the pressure distribution information and the electrocardiogram information of the child passenger specifically comprises: obtaining the facial image information through a vehicle-mounted camera; obtaining multiple body key point coordinates of the child passenger through a seat built-in millimeter wave radar to obtain the body posture information; obtaining pressure distribution conditions of the back and legs of the child passenger through a flexible piezoelectric sensor matrix integrated on a seat surface to obtain the pressure distribution information; and obtaining the electrocardiogram information through a seat built-in electrocardiogram sensor. The electrocardiogram information is acquired by a non-contact electrocardiogram sensor.

3. The method of claim 1, wherein, The expression feature of the child passenger is extracted according to the face image information, and specifically includes: Face detection and alignment are performed on the face image information by a preset face detection model to obtain a face feature image; Expression recognition is performed on the face feature image to obtain a pain level and / or a restlessness level of the child passenger; The expression feature is determined according to the pain level and / or the restlessness level.

4. The method of claim 1, wherein, The posture feature of the child passenger is extracted according to the body posture information and the pressure distribution information, and specifically includes: The trunk inclination and the head offset of the child passenger are determined according to the body posture information; The center of gravity offset of the child passenger is determined according to the pressure distribution information; The posture feature is determined according to the trunk inclination, the head offset and the center of gravity offset.

5. The method of claim 1, wherein, The physiological feature of the child passenger is extracted according to the electrocardiogram information, and specifically includes: R-wave detection is performed on the electrocardiogram information to obtain an RR interval sequence; Frequency domain analysis is performed on the RR interval sequence to obtain a low-frequency part and a high-frequency part; An HRV low-frequency / high-frequency ratio is determined according to the low-frequency part and the high-frequency part, and the HRV low-frequency / high-frequency ratio is taken as the physiological feature.

6. The method of claim 1, wherein, The comfort recognition model is obtained by training through the following steps: Sample expression features, sample posture features and sample physiological features of a test child passenger in a test vehicle are acquired, and corresponding comfort labels are determined by manual labeling; Training samples are determined according to the sample expression features, the sample posture features and the sample physiological features; The training samples are input into a pre-constructed deep learning neural network to obtain comfort recognition results; A loss value is determined according to the comfort recognition results and the comfort labels; Parameters of the deep learning neural network are updated according to the loss value to obtain the trained comfort recognition model.

7. A machine vision based child seat belt smart adjustment system, characterized in that, A machine vision-based child seatbelt intelligent adjustment method according to any one of claims 1-6, comprising: An information acquisition module for acquiring face image information, body posture information, pressure distribution information and electrocardiogram information of a child passenger; A feature extraction module for extracting an expression feature of the child passenger according to the face image information, a posture feature of the child passenger according to the body posture information and the pressure distribution information, and a physiological feature of the child passenger according to the electrocardiogram information; A comfort recognition module for inputting the expression feature, the posture feature and the physiological feature into a pre-trained comfort recognition model to obtain a ride comfort of the child passenger; A locking force adjustment module for acquiring real-time vehicle status of a current vehicle, and adjusting a locking force of a seatbelt of the child passenger according to the ride comfort and the real-time vehicle status.

8. A machine vision based child seat belt intelligent adjustment device, characterized in that, Comprise: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor is caused to realize a machine vision-based child seat belt intelligent adjustment method according to any one of claims 1 to 6.

9. A computer readable storage medium having stored therein a program which is executable by a processor, characterized in that, The program executable by the processor, when executed by the processor, is used to execute a machine vision-based child seat belt intelligent adjustment method according to any one of claims 1 to 6.

Images