Machine vision-based vehicle sunshade curtain control method, system, device and medium

Abstract

The application discloses a vehicle sunshade curtain control method, system and device based on machine vision and a medium. The method comprises the following steps: acquiring first image information of rear seats of a target vehicle, determining human body image information of rear seat passengers according to the first image information, and further determining human body time sequence data of the rear seat passengers according to the human body image information; inputting the human body time sequence data into a pre-trained vehicle-riding state recognition model to obtain a vehicle-riding state recognition result of the rear seat passengers; and controlling the opening and closing of a sunshade curtain of the target vehicle according to the light intensity information of an external environment of the target vehicle and the vehicle-riding state recognition result. The application can recognize the change of the human body posture in the human body time sequence data, thereby accurately recognizing whether the rear seat passengers are in a resting state, further accurately judging whether shading is needed in combination with real-time light intensity, improving the efficiency and accuracy of vehicle sunshade curtain control, improving the vehicle-riding experience of users, and being widely applicable 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 vehicle sunshade curtain control method, system, device and medium based on machine vision. BACKGROUND

[0002] As a common means of transportation, cars have become indispensable in human life. The car window is an important part of the car body and is designed to meet the needs of interior lighting, ventilation and field of view for drivers and passengers. Some models also have additional roof windows, which make it easier for interior lighting and ventilation. When the external light is too strong, the sunlight shining into the car through the rear window and / or roof window will affect the rest of the rear seat passengers. Therefore, a sunshade curtain is usually set up in the car to block the strong light shining into the car from the rear window and / or roof window.

[0003] In the prior art, the sunshade curtain of the vehicle is mostly controlled by the subjective judgment of the driver and passenger whether to need sunshade, which is relatively cumbersome and inefficient. The related technical solution discloses a method of judging whether the passenger is in a sleep state according to the closed-eye time and closed-eye frequency of the face image, and then controlling the sunshade curtain. However, this method needs to accurately collect the face image of the passenger to count the closed-eye time and closed-eye frequency. However, the passenger may be accompanied by actions such as lowering his head and turning his head while sleeping, which makes it impossible to take a complete face image. Even the passenger's eye area may be blocked by a hat brim, glasses or an eye patch, which affects the accuracy of sleep state recognition and further affects the accuracy of sunshade curtain control. SUMMARY

[0004] The present application aims to at least partially solve one of the problems in the prior art.

[0005] To this end, one object of the present application is to provide a vehicle sunshade curtain control method based on machine vision, which improves the efficiency and accuracy of vehicle sunshade curtain control.

[0006] Another object of the present application is to provide a vehicle sunshade curtain control system based on machine vision.

[0007] In order to achieve the above technical purpose, the technical solution adopted by the present application comprises:

[0008] In a first aspect, the present application provides a vehicle sunshade curtain control method based on machine vision, comprising the following steps:

[0009] Obtaining first image information of the rear seat of the target vehicle, determining the human body image information of the rear seat passenger according to the first image information, and further determining the human body time sequence data of the rear seat passenger according to the human body image information.

[0010] inputting the human body time sequence data into a pre-trained vehicle-riding state recognition model to obtain a vehicle-riding state recognition result of the rear seat passenger;

[0011] obtaining illumination intensity information of an external environment of the target vehicle;

[0012] controlling opening and closing of a sunshade curtain of the target vehicle according to the illumination intensity information and the vehicle-riding state recognition result.

[0013] Further, in an embodiment of the present application, the step of obtaining first image information of a rear seat of a target vehicle and determining human body image information of a rear seat passenger according to the first image information specifically comprises:

[0014] obtaining the first image information by a camera device arranged at a back of a front seat of the target vehicle according to a preset sampling frequency;

[0015] obtaining background image information of the rear seat and determining foreground image information of the rear seat according to the first image information and the background image information;

[0016] performing edge detection on the foreground image information to obtain a plurality of continuous contours, and then performing random Hough transformation on the continuous contours to screen out continuous contours meeting a preset threshold condition as object contours;

[0017] performing denoising processing on the foreground image information according to the object contours to obtain the human body image information of the rear seat passenger.

[0018] Further, in an embodiment of the present application, the step of determining human body time sequence data of the rear seat passenger according to the human body image information specifically comprises:

[0019] obtaining a plurality of human body image information in a sampling period;

[0020] sorting the human body image information according to sampling time to obtain the human body time sequence data.

[0021] Further, in an embodiment of the present application, the vehicle sunshade curtain control method further comprises a step of pre-training a vehicle-riding state recognition model, which specifically comprises:

[0022] obtaining a plurality of sample time sequence data of a test passenger in different vehicle-riding states on a vehicle rear seat, and determining sample labels of the sample time sequence data according to vehicle-riding states of the test passenger;

[0023] constructing a training sample set according to the sample time sequence data and the corresponding sample labels;

[0024] inputting the training sample set into a pre-constructed bidirectional recurrent neural network for training, and optimizing model parameters of the bidirectional recurrent neural network to obtain a trained vehicle-riding state recognition model;

[0025] The sample label includes an active state and a rest state, and the bidirectional recurrent neural network includes an input layer, a forward hidden layer, a reverse hidden layer, and an output layer.

[0026] Further, in an embodiment of the present application, the step of inputting the training sample set into a pre-constructed bidirectional recurrent neural network for training, and optimizing model parameters of the bidirectional recurrent neural network to obtain a trained vehicle-riding state recognition model specifically includes:

[0027] inputting the sample time series data into the input layer and calculating a first hidden state vector through the forward hidden layer;

[0028] inputting the sample time series data into the input layer in reverse order and calculating a second hidden state vector through the reverse hidden layer;

[0029] performing reverse processing on the second hidden state vector to obtain a third hidden state vector, and performing splicing processing on the first hidden state vector and the third hidden state vector to obtain a fourth hidden state vector;

[0030] inputting the fourth hidden state vector into the output layer to output a vehicle-riding state prediction result;

[0031] determining a loss value of the bidirectional recurrent neural network according to the vehicle-riding state prediction result and the sample label;

[0032] updating the model parameters of the bidirectional recurrent neural network according to the loss value, and returning to the step of inputting the sample time series data into the input layer;

[0033] stopping training when the loss value reaches a preset first threshold value or the number of iterations reaches a preset second threshold value, to obtain the trained vehicle-riding state recognition model.

[0034] Further, in an embodiment of the present application, the step of obtaining the light intensity information of the external environment of the target vehicle specifically includes:

[0035] obtaining the light intensity information through a light sensor arranged on a roof and / or a window of the target vehicle.

[0036] Furthermore, in one embodiment of the present invention, the step of controlling the opening and closing of the sunshade of the target vehicle based on the light intensity information and the vehicle status recognition result specifically includes:

[0037] When the vehicle status recognition result is a resting state, and the light intensity information is greater than or equal to a preset first threshold, the sunshade is controlled to be activated so that the sunshade blocks the rear window and / or roof window of the target vehicle.

[0038] When the vehicle status recognition result indicates an active state, or the light intensity information is less than the first threshold, the sunshade is controlled to be in a closed state.

[0039] Secondly, embodiments of the present invention provide a vehicle sunshade control system based on machine vision, comprising:

[0040] The human body time series data determination module is used to acquire the first image information of the rear seat of the target vehicle, determine the human body image information of the rear seat personnel based on the first image information, and then determine the human body time series data of the rear seat personnel based on the human body image information.

[0041] The passenger status recognition module is used to input the human body time series data into a pre-trained passenger status recognition model to obtain the passenger status recognition result of the back seat person;

[0042] The light intensity information acquisition module is used to acquire the light intensity information of the external environment of the target vehicle;

[0043] The sunshade opening and closing control module is used to control the opening and closing of the sunshade of the target vehicle based on the light intensity information and the vehicle status recognition result.

[0044] Thirdly, embodiments of the present invention provide a vehicle sunshade control device based on machine vision, comprising:

[0045] At least one processor;

[0046] At least one memory for storing at least one program;

[0047] When the at least one program is executed by the at least one processor, the at least one processor implements the above-described machine vision-based vehicle sunshade control method.

[0048] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to perform the aforementioned machine vision-based vehicle sunshade control method.

[0049] The advantages and beneficial effects of the present invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention:

[0050] This invention acquires first image information of the rear seats of a target vehicle, determines the human image information of the rear seat occupants based on the first image information, and then determines the human body time-series data of the rear seat occupants based on the human body image information. The human body time-series data is then input into a pre-trained vehicle state recognition model to obtain the vehicle state recognition result of the rear seat occupants. Simultaneously, it acquires the light intensity information of the external environment of the target vehicle, thereby controlling the opening and closing of the target vehicle's sunshade based on the light intensity information and the vehicle state recognition result. This invention extracts the human body time-series data of the rear seat occupants, and through a pre-trained vehicle state recognition model, can perceive and recognize changes in human posture in the human body time-series data. This allows for accurate identification of whether the rear seat occupants are in a resting state, and, combined with real-time light intensity, accurately determines whether shading is needed, improving the efficiency and accuracy of vehicle sunshade control and enhancing the user's riding experience. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the embodiments of the present invention are described below. It should be understood that the drawings described below are only for the convenience of clearly describing some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 A flowchart illustrating the steps of a machine vision-based vehicle sunshade control method provided in an embodiment of the present invention;

[0053] Figure 2 A structural block diagram of a vehicle sunshade control system based on machine vision provided in an embodiment of the present invention;

[0054] Figure 3 This is a structural block diagram of a vehicle sunshade control device based on machine vision, provided in an embodiment of the present invention. Detailed Implementation

[0055] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. The step numbers in the following embodiments are set only for ease of explanation, and there is no limitation on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0056] In the description of this invention, "multiple" means two or more. The use of "first" and "second" is for distinguishing technical features only and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or the order of the indicated technical features. Furthermore, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art.

[0057] Reference Figure 1 This invention provides a machine vision-based method for controlling vehicle sunshades, specifically including the following steps:

[0058] S101. Obtain the first image information of the rear seat of the target vehicle, determine the human body image information of the rear seat personnel based on the first image information, and then determine the human body time sequence data of the rear seat personnel based on the human body image information.

[0059] Specifically, in this embodiment of the invention, human images are continuously acquired to form human time-series data, which can reflect changes in the posture of the person in the back seat.

[0060] As a further optional implementation, the step of acquiring first image information of the rear seats of the target vehicle and determining the human image information of the rear seat occupants based on the first image information specifically includes:

[0061] S1011. First image information is acquired by a camera device installed on the back of the front seat of the target vehicle at a preset sampling frequency.

[0062] S1012. Obtain the background image information of the rear seats, and determine the foreground image information of the rear seats based on the first image information and the background image information;

[0063] S1013. Perform edge detection on the foreground image information to obtain several continuous contours, and then perform random Hough transform on the continuous contours to select the continuous contours that meet the preset threshold conditions as object contours.

[0064] S1014. Denoise the foreground image information based on the object contour to obtain the human image information of the person in the back seat.

[0065] Specifically, a camera embedded in the back of the front seat can continuously capture images of the rear seats, with a sampling frequency of 10Hz, or 10 samples per second. After acquiring the first image information, it needs to be differentially processed with the background image information of the rear seats to obtain the foreground image information. This foreground image information includes the human body image of the rear seat occupants, and may also include other objects such as backpacks, toys, etc. Therefore, it is necessary to obtain continuous contours through edge detection and then filter them.

[0066] In this embodiment of the invention, edge detection of the foreground image information can yield several continuous contours (if there is only one person in the back seat and no other objects, only one continuous contour can be obtained; otherwise, multiple continuous contours may be obtained). Image edges are the parts of the image where the brightness changes significantly in a local area. For grayscale images, this means areas where the grayscale value changes significantly, i.e., a rapid change from one grayscale value to another with a large difference in grayscale value within a very small buffer area. This embodiment of the invention uses the Canny operator for edge detection, which improves the sensitivity to the edges of people and objects while suppressing noise. The specific process is as follows:

[0067] 1) Edge detection is performed on the foreground image information using the Canny operator to obtain the second image information, and continuous contours in the second image information are determined;

[0068] 2) Perform a traversal search on the second image information. When a continuous contour is found, perform a random Hough transform on the continuous contour to determine its shape and size.

[0069] 3) When the shape and size of a continuous contour meet the preset threshold conditions, the continuous contour is determined to be an object contour.

[0070] Specifically, the second image information extracted by the Canny operator is traversed and searched, starting from the top left corner of the image, proceeding from top to bottom and from left to right, until the i-th independent continuous contour in the image is found. This continuous contour is then subjected to a Randomized Hough Transform (RHT) to remove straight lines and excessively large contours, and square, circular, and contours with areas below a preset threshold (e.g., 100cm²) are selected. 2 Irregular contours that meet certain conditions can be identified as object contours.

[0071] Understandably, after obtaining the object's outline, it is necessary to denoise the foreground image information based on the object's outline, that is, to remove the pixel area surrounded by the object's outline, and the remaining area is the human image information of the person in the back seat.

[0072] As a further optional implementation, the step of determining the human body time-series data of the person in the back seat based on the human body image information specifically includes:

[0073] S1015. Acquire multiple human image information within a sampling period;

[0074] S1016. Sort the human body image information according to the sampling time to obtain human body time series data.

[0075] Specifically, taking a sampling frequency of 10Hz as an example, the human body image information corresponding to the 10 first image information captured continuously within 1 second is sorted according to the sampling time to obtain a set of human body time series data. This human body time series data can reflect the changes in human posture of the back seat person within the corresponding sampling period, so that the pre-trained riding status recognition model can determine whether the person is in a resting state or an active state.

[0076] S102. Input the human body time series data into the pre-trained vehicle state recognition model to obtain the vehicle state recognition results of the back seat passengers.

[0077] Specifically, the vehicle status recognition model of this invention is trained by pre-collecting multiple sample time-series data of testers in different vehicle statuses as training samples. The training process is described below.

[0078] As a further optional implementation, the vehicle sunshade control method also includes a step of pre-training a passenger state recognition model, which specifically includes:

[0079] A1. Obtain multiple sample time-series data of testers in different riding states in the back seat of the vehicle, and determine the sample label of each sample time-series data according to the riding state of the testers.

[0080] A2. Construct a training sample set based on the sample time series data and the corresponding sample labels;

[0081] A3. Input the training sample set into the pre-built bidirectional recurrent neural network for training, optimize the model parameters of the bidirectional recurrent neural network, and obtain the trained vehicle status recognition model.

[0082] The sample labels include active and resting states, and the bidirectional recurrent neural network includes an input layer, a forward hidden layer, a backward hidden layer, and an output layer.

[0083] Specifically, when constructing the training sample set, the corresponding sample labels are determined based on the actual state of the testers. By collecting a sufficient number of testers of different body types and in different riding situations as training samples, the training sample set can be formed.

[0084] This invention employs a bidirectional recurrent neural network to train a passenger state recognition model. The basic idea of ​​a bidirectional recurrent neural network (BRNN) is to use two recurrent neural networks (RNNs) to calculate the hidden state for each training sequence in both the forward and backward directions, and then integrate and output the calculation results through an output layer.

[0085] As a further optional implementation, the step of inputting the training sample set into a pre-built bidirectional recurrent neural network for training, and optimizing the model parameters of the bidirectional recurrent neural network to obtain a trained passenger status recognition model specifically includes:

[0086] A31. Input the sample time series data into the input layer, and calculate the first hidden state vector through the forward hidden layer;

[0087] A32. Input the sample time series data into the input layer in reverse order, and calculate the second hidden state vector through the reverse hidden layer;

[0088] A33. Reverse the second hidden state vector to obtain the third hidden state vector, and concatenate the first hidden state vector and the third hidden state vector to obtain the fourth hidden state vector.

[0089] A34. Input the fourth hidden state vector into the output layer and output the prediction result of the riding state.

[0090] A35. Determine the loss value of the bidirectional recurrent neural network based on the prediction results of the passenger status and the sample labels;

[0091] A36. Update the model parameters of the bidirectional recurrent neural network based on the loss value and return to the step of inputting the sample time series data into the input layer;

[0092] A37. When the loss value reaches the preset first threshold or the number of iterations reaches the preset second threshold, training is stopped, and the trained vehicle status recognition model is obtained.

[0093] Specifically, after inputting the data from the training dataset into the initialized bidirectional recurrent neural network, the model output recognition result, i.e., the ride status prediction result, can be obtained through the calculation and concatenation of the forward and backward hidden state vectors. The accuracy of the model's recognition can be evaluated based on the ride status prediction result and the aforementioned label information, thereby updating the model's parameters. For the ride status recognition model, the accuracy of the model's recognition result can be measured by a loss function. The loss function is defined on a single training data point and is used to measure the prediction error of that training data point. Specifically, the loss value of that training data point is determined by the label of that individual training data point and the model's prediction result for that training data point. In actual training, a training dataset contains many training data points. Therefore, a cost function is generally used to measure the overall error of the training dataset. The cost function is defined on the entire training dataset and is used to calculate the average prediction error of all training data points, providing a better measure of the model's prediction performance. For general machine learning models, the aforementioned cost function, plus a regularization term to measure model complexity, can serve as the training objective function. Based on this objective function, the loss value of the entire training dataset can be calculated. There are many commonly used loss functions, such as 0-1 loss, squared loss, absolute loss, logarithmic loss, and cross-entropy loss, all of which can be used as loss functions for machine learning models. These will not be elaborated upon here. In this embodiment of the invention, any one of these loss functions can be selected to determine the training loss value. Based on the training loss value, the backpropagation algorithm is used to update the model parameters. After several iterations, a well-trained vehicle status recognition model can be obtained. The specific number of iterations can be preset, or training can be considered complete when the accuracy requirement on the test set is met.

[0094] For the human body time series data to be identified, inputting it into the trained vehicle status recognition model will give the current vehicle status of the passenger in the back seat.

[0095] S103. Obtain the light intensity information of the external environment of the target vehicle.

[0096] As a further optional implementation, the step of obtaining the light intensity information of the external environment of the target vehicle specifically includes:

[0097] Light intensity information is obtained by light sensors installed on the roof and / or windows of the target vehicle.

[0098] S104. Control the opening and closing of the sunshade of the target vehicle based on the light intensity information and the passenger status recognition result.

[0099] As a further optional implementation, the step of controlling the opening and closing of the sunshade of the target vehicle based on light intensity information and vehicle status recognition results specifically includes:

[0100] S1041. When the vehicle status recognition result is a resting state and the light intensity information is greater than or equal to a preset first threshold, control the sunshade to be in the activated state so that the sunshade blocks the rear window and / or roof window of the target vehicle.

[0101] S1042. When the passenger status recognition result is active, or the light intensity information is less than the first threshold, control the sunshade to be closed.

[0102] Specifically, in this embodiment of the invention, the resting state includes states such as sleeping and resting with eyes closed. In this case, when the detected external light intensity is greater than a preset first threshold (which can be preset by the driver or passenger), the sunshade can be activated to provide shade. When the user returns from the resting state to the active state, or when the external light intensity drops below the first threshold, the sunshade is closed to allow light to enter the vehicle.

[0103] The method steps of the embodiments of the present invention have been described above. It can be understood that the embodiments of the present invention extract the human body time-series data of rear-seat occupants. Through a pre-trained vehicle state recognition model, changes in human posture within the human body time-series data can be perceived and recognized, thereby accurately identifying whether rear-seat occupants are in a resting state. Furthermore, combined with real-time light intensity, it accurately determines whether shading is necessary, improving the efficiency and accuracy of vehicle sunshade control and enhancing the user's riding experience.

[0104] Reference Figure 2 This invention provides a machine vision-based vehicle sunshade control system, comprising:

[0105] The human body time series data determination module is used to acquire the first image information of the rear seat of the target vehicle, determine the human body image information of the rear seat personnel based on the first image information, and then determine the human body time series data of the rear seat personnel based on the human body image information.

[0106] The passenger status recognition module is used to input human body time series data into a pre-trained passenger status recognition model to obtain the passenger status recognition results of the back seat passengers.

[0107] The light intensity information acquisition module is used to acquire the light intensity information of the external environment of the target vehicle;

[0108] The sunshade opening and closing control module is used to control the opening and closing of the sunshade of the target vehicle based on light intensity information and passenger status recognition results.

[0109] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0110] Reference Figure 3 This invention provides a machine vision-based vehicle sunshade control device, comprising:

[0111] At least one processor;

[0112] At least one memory for storing at least one program;

[0113] When the above-mentioned at least one program is executed by the above-mentioned at least one processor, the above-mentioned at least one processor implements the above-mentioned machine vision-based vehicle sunshade control method.

[0114] The content of the above method embodiments is applicable to the device embodiments. The specific functions implemented by the device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0115] This invention also provides a computer-readable storage medium storing a processor-executable program that, when executed by a processor, performs the aforementioned machine vision-based vehicle sunshade control method.

[0116] This invention provides a computer-readable storage medium that can execute a machine vision-based vehicle sunshade control method provided in the method embodiments of this invention. It can execute any combination of the implementation steps of the method embodiments and has the corresponding functions and beneficial effects of the method.

[0117] This invention also discloses a computer program product or computer program, which 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, causing the computer device to perform... Figure 1 The method shown.

[0118] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the aforementioned blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0119] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the aforementioned functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0120] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0121] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0122] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the aforementioned program can be printed, because the aforementioned program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or, if necessary, processing in other suitable ways, and then stored in computer memory.

[0123] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0124] In the foregoing description of this specification, references to terms such as "one embodiment," "another embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0125] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0126] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A machine vision-based control method for a vehicle sunshade, characterized by, The method comprises the following steps: obtaining first image information of rear seats of a target vehicle, determining human body image information of a rear seat passenger according to the first image information, and further determining human body time sequence data of the rear seat passenger according to the human body image information; inputting the human body time sequence data into a pre-trained vehicle-riding state recognition model to obtain a vehicle-riding state recognition result of the rear seat passenger; obtaining illumination intensity information of an external environment of the target vehicle; controlling opening and closing of a sunshade curtain of the target vehicle according to the illumination intensity information and the vehicle-riding state recognition result; The vehicle sunshade curtain control method further comprises a step of pre-training a vehicle-riding state recognition model, which specifically comprises: obtaining a plurality of sample time sequence data of a test person in different vehicle-riding states on the rear seats of a vehicle, and determining sample labels of the sample time sequence data according to the vehicle-riding state of the test person; constructing a training sample set according to the sample time sequence data and the corresponding sample labels; inputting the training sample set into a pre-constructed bidirectional recurrent neural network for training, optimizing model parameters of the bidirectional recurrent neural network, and obtaining a trained vehicle-riding state recognition model; wherein the sample labels include an active state and a resting state, and the bidirectional recurrent neural network comprises an input layer, a forward hidden layer, a reverse hidden layer, and an output layer; The step of inputting the training sample set into the pre-constructed bidirectional recurrent neural network for training, optimizing the model parameters of the bidirectional recurrent neural network, and obtaining the trained vehicle-riding state recognition model specifically comprises: inputting the sample time sequence data into the input layer and calculating a first hidden state vector through the forward hidden layer; inputting the sample time sequence data in reverse order into the input layer and calculating a second hidden state vector through the reverse hidden layer; performing reverse order processing on the second hidden state vector to obtain a third hidden state vector, and performing splicing processing on the first hidden state vector and the third hidden state vector to obtain a fourth hidden state vector; inputting the fourth hidden state vector into the output layer to output a vehicle-riding state prediction result; determining a loss value of the bidirectional recurrent neural network according to the vehicle-riding state prediction result and the sample labels; updating the model parameters of the bidirectional recurrent neural network according to the loss value, and returning to the step of inputting the sample time sequence data into the input layer; stopping training when the loss value reaches a pre-set first threshold value or the number of iterations reaches a pre-set second threshold value, and obtaining the trained vehicle-riding state recognition model.

2. The machine vision-based control method of a vehicle sunshade according to claim 1, wherein, The step of obtaining first image information of rear seats of a target vehicle and determining human body image information of a rear seat passenger according to the first image information specifically comprises: obtaining the first image information at a pre-set sampling frequency through a camera device arranged on a back of front seats of the target vehicle; obtaining background image information of the rear seats, and determining foreground image information of the rear seats according to the first image information and the background image information; Edge detection is performed on the foreground image information to obtain a plurality of continuous contours, and then random Hough transformation is performed on the continuous contours to screen out continuous contours meeting preset threshold conditions as object contours; The foreground image information is denoised according to the object contours to obtain human body image information of the rear seat passenger.

3. The machine vision-based control method of a vehicle sunshade according to claim 2, wherein, The step of determining human body time sequence data of the rear seat passenger according to the human body image information specifically includes: Obtaining a plurality of human body image information in a sampling period; The human body time sequence data is obtained by sorting the human body image information according to sampling time.

4. The machine vision-based control method of a vehicle sunshade according to claim 1, wherein, The step of obtaining the light intensity information of the external environment of the target vehicle specifically includes: The light intensity information is obtained by a light sensor arranged on the roof and / or window of the target vehicle.

5. The machine vision-based control method of a vehicle sunshade according to any one of claims 1 to 4, characterized in that, The step of controlling the opening and closing of the sunshade curtain of the target vehicle according to the light intensity information and the riding state recognition result specifically includes: When the riding state recognition result is the resting state and the light intensity information is greater than or equal to a preset first threshold, the sunshade curtain is controlled to be in the starting state so that the sunshade curtain shields the rear window and / or roof window of the target vehicle; When the riding state recognition result is the active state or the light intensity information is less than the first threshold, the sunshade curtain is controlled to be in the closed state.

6. A machine vision-based control system for a vehicle sunshade, the system comprising: A vehicle sunshade curtain control method based on machine vision is implemented, as claimed in any one of claims 1 to 5, comprising: A human body time sequence data determination module is configured to obtain first image information of rear seats of a target vehicle, determine human body image information of a rear seat passenger according to the first image information, and further determine human body time sequence data of the rear seat passenger according to the human body image information; A riding state recognition module is configured to input the human body time sequence data into a pre-trained riding state recognition model to obtain a riding state recognition result of the rear seat passenger; A light intensity information acquisition module is configured to acquire light intensity information of an external environment of the target vehicle; A sunshade curtain opening and closing control module is configured to control the opening and closing of the sunshade curtain of the target vehicle according to the light intensity information and the riding state recognition result.

7. A machine vision-based control apparatus for a vehicle sunshade, characterized by comprising: 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 implements a vehicle sunshade curtain control method based on machine vision as claimed in any one of claims 1 to 5.

8. 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 perform a vehicle sunshade curtain control method based on machine vision as claimed in any one of claims 1 to 5.

Images