Traffic status detection method and device, gate system and electronic equipment

Abstract

The present invention discloses a traffic status detection method and device, a gate system and electronic equipment, in which an established neural network model is used to analyze the light spot signals of each light source sensor fed back by a light curtain detection device, determine the traffic status of the passerby, and then control the opening and closing time of the gate according to the traffic status of the passerby. The traffic status detection method provided by the embodiment of the present invention can accurately identify the traffic status of the current passerby, and then determine the traffic time of the gate according to the traffic status of the passerby, so that the passerby can pass through the gate smoothly, avoiding the problem of pinching the passerby or pinching the luggage. At the same time, it can also prohibit the situation of multiple people passing, ensuring the safety and rationality of the passers-by passing through the gate.

Description

Technical Field

[0001] The present invention relates to the field of gate technology, and in particular to a traffic status detection method and device, a gate system and electronic equipment. Background Art

[0002] With the development of urban rail transit, more and more people choose to travel by subway, high-speed rail and other means of transportation. In order to enable people to board the train in an orderly manner, a corresponding gate system is generally set up at the entrance of the station. The boarding passengers enter the station in an orderly manner through the gates of the gate system. The gates are generally opened one by one for each passenger, thereby ensuring the order and stability of entering the station.

[0003] When passengers pass through the gate system, they will present a variety of different passing states. For example, the passing state can be a single passenger passing, a passenger passing with luggage, a passenger passing with a child, a passenger passing in a wheelchair, and other reasonable passing states. At the same time, there are sometimes unreasonable passing states where multiple adults pass through the gate system together. In the traditional gate system, when the detection device detects the passing state of the passing passengers, it is impossible to accurately identify the different passing states of the passengers, which can easily lead to the passengers' accompanying items or followers being pinched in the reasonable passing state, or the passengers are still released in the unreasonable passing state, resulting in fare evasion. Summary of the invention

[0004] In view of the above problems, the present invention provides a traffic status detection method and device, a gate system and an electronic device that overcome the above problems or at least partially solve the above problems. The technical solution is as follows:

[0005] A method for detecting a traffic state, the method being applied to a gate system, the gate system comprising a gate and a light curtain detection device, the light curtain detection device being arranged on one side of the gate, the light curtain detection device being arranged with a plurality of light source sensors, the plurality of light source sensors being distributed in a preset geometric shape; the method comprising:

[0006] When a passer-by passes through the gate, light spot signals of each light source sensor in each time slice within a preset time interval are obtained; the preset time interval is composed of N time slices, where N is a positive integer;

[0007] According to the light spot signals of each of the light source sensors on each of the time slices, a light spot matrix corresponding to the passer-by is determined, wherein the column data of the light spot matrix are the light spot signals of each of the light source sensors arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each of the light source sensors on each time slice arranged in a time order of each of the time slices;

[0008] The light spot matrix is ​​input into the established first neural network model. After being processed by the first neural network model, the passage status of the passerby corresponding to the light spot matrix is ​​output, and then the passage time of the gate is controlled according to the passage status of the passerby.

[0009] In the above method, optionally, the preset geometric shape is a π shape.

[0010] In the above method, optionally, the preset identification order is an order in which the identification numbers of the light source sensors are arranged in sequence.

[0011] In the above method, optionally, the process of establishing the first neural network model includes:

[0012] Determine the initial network model;

[0013] Determine a data set for establishing the neural network model, the data set comprising: a training data set, a verification data set and a test data set; the training data set comprises a plurality of training light spot matrices and a first pass state corresponding to each of the training light spot matrices, the verification data set comprises a plurality of verification light spot matrices and a second pass state corresponding to each of the verification light spot matrices, the test data set comprises a plurality of test light spot matrices and a third pass state corresponding to each of the test light spot matrices; the data structures of the training light spot matrices, the verification light spot matrices and the test light spot matrices are the same;

[0014] Taking each training light spot matrix in the training data set as the input of the initial network model, and taking the first traffic state corresponding to each training light spot matrix as the output of the initial network model, to train the initial network model;

[0015] During the training process of the initial network model, when the loss function of the initial network model is less than a preset threshold, the training of the initial network model is terminated, and a second neural network model corresponding to the initial network model is generated;

[0016] Using each verification light spot matrix in the verification data set and the second traffic state corresponding to each verification light spot matrix, the second neural network model is verified until the qualified rate of the traffic state output by the second neural network model is greater than a preset qualified threshold value, and a third neural network model corresponding to the second neural network model is generated;

[0017] The third neural network model is tested using each test light spot matrix in the test data set and the third traffic state corresponding to each test light spot matrix. If the third neural network model is not in an overfitting state, the third neural network model is used as the neural network model for traffic state detection.

[0018] The above method may optionally further include:

[0019] If the third neural network model is in an overfitting state, the model parameters of the third neural network model and the function parameters of the loss function are adjusted until the state of the adjusted current network model is no longer overfitting, and the adjusted current network model is used as the neural network model for traffic status detection.

[0020] In the above method, optionally, controlling the passage time of the gate according to the passage status of the passer-by includes:

[0021] Determine the status type of the passerby's pass status;

[0022] Determining a preset travel time corresponding to the state type;

[0023] The gate is controlled to open and close according to the passage time.

[0024] A traffic status detection device, the device is applied to a gate system, the gate system includes a gate and a light curtain detection device, the light curtain detection device is arranged on one side of the gate, a plurality of light source sensors are arranged in the light curtain detection device, and the plurality of light source sensors are distributed in a preset geometric shape; the device includes:

[0025] An acquisition unit, used for acquiring light spot signals of each of the light source sensors in each time slice within a preset time interval when a passerby passes through the gate; the preset time interval is composed of N time slices, where N is a positive integer;

[0026] A determination unit, used to determine the light spot matrix corresponding to the passerby according to the light spot signals of each of the light source sensors on each of the time slices, wherein the column data of the light spot matrix are the light spot signals of each of the light source sensors arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each of the light source sensors on each time slice arranged in a time order of each of the time slices;

[0027] An output unit is used to input the light spot matrix into the established first neural network model, and after being processed by the first neural network model, output the passage status of the passerby corresponding to the light spot matrix, and then control the passage time of the gate according to the passage status of the passerby.

[0028] In the above device, optionally, the preset geometric shape is a π shape.

[0029] A gate system, comprising: a gate, a light curtain detection device and a gate processor;

[0030] The light curtain detection device is arranged on one side of the gate, and a plurality of light source sensors are arranged in the light curtain detection device, and the plurality of light source sensors are distributed in a π shape;

[0031] The gate processor is used to execute the above-mentioned traffic status detection method.

[0032] An electronic device includes at least one processor, and at least one memory and a bus connected to the processor; wherein the processor and the memory communicate with each other via the bus; the processor is used to call program instructions in the memory to execute the above-mentioned traffic status detection method.

[0033] Compared with the prior art, the present invention has the following advantages: a traffic status detection method, the method is applied to a gate system, the gate system includes a gate and a light curtain detection device, the light curtain detection device is arranged on one side of the gate, a plurality of light source sensors are arranged in the light curtain detection device, and the plurality of light source sensors are distributed in a preset geometric shape; the method comprises:

[0034] When a passerby passes through the gate, the light spot signals of each of the light source sensors on each time slice within a preset time interval are obtained; the preset time interval is composed of N time slices, and N is a positive integer; based on the light spot signals of each of the light source sensors on each of the time slices, the light spot matrix corresponding to the passerby is determined, the column data of the light spot matrix are the light spot signals of each of the light source sensors arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each of the light source sensors on each time slice arranged in the time order of each of the time slices; the light spot matrix is ​​input into the established first neural network model, and after being processed by the first neural network model, the passage status of the passerby corresponding to the light spot matrix is ​​output, and then the passage time of the gate is controlled according to the passage status of the passerby.

[0035] The traffic status detection method provided in the embodiment of the present invention can accurately identify the traffic status of the current passerby, and then determine the traffic time of the gate according to the traffic status of the passerby, so that the passerby can pass through the gate smoothly, avoiding the problem of pinching the passerby or pinching the luggage. At the same time, it can also prohibit the passage of multiple people, ensuring the safety and rationality of the passers-by passing through the gate.

[0036] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0038] Figure 1 A method flow chart of a traffic status detection method provided by an embodiment of the present invention;

[0039] Figure 2 A data structure diagram of a light spot matrix in a traffic status detection method provided by an embodiment of the present invention;

[0040] Figure 3 A flow chart of another method for detecting a traffic status provided by an embodiment of the present invention;

[0041] Figure 4 A flow chart of another method for detecting a traffic status provided by an embodiment of the present invention;

[0042] Figure 5 A structural diagram of a traffic status detection device provided by an embodiment of the present invention;

[0043] Figure 6 A structural diagram of a gate system provided by an embodiment of the present invention;

[0044] Figure 7 A structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0046] refer to Figure 1 FIG. 2 shows a flow chart of a method for detecting a traffic status provided by an embodiment of the present invention. Figure 1 The method execution process shown is a feasible implementation scheme of the traffic status detection method provided by the embodiment of the present invention. The method can be applied in various gate systems, and its execution subject is a processor set in the gate system.

[0047] The traffic status detection method provided in an embodiment of the present invention is applied to a gate system, wherein the gate system includes a gate and a light curtain detection device, wherein the light curtain detection device is arranged on one side of the gate, and wherein a plurality of light source sensors are arranged in the light curtain detection device, and wherein the plurality of light source sensors are distributed in a preset geometric shape.

[0048] In the method provided in an embodiment of the present invention, a light curtain detection device is arranged on one side of the gate. Those skilled in the art can understand that, in order to realize the status detection of each passerby passing through the gate, the light curtain detection device can be arranged on one side of the gate in the passing direction, and the setting position can be adjusted according to actual needs, and can be the left front or right front of the side where the passersby enter the gate in the passing direction.

[0049] In the method provided by the embodiment of the present invention, multiple light source sensors are distributed in a preset geometric shape. Preferably, the geometric shape may be a π shape, that is, the shape in which the multiple light source sensors are arranged is π. In the method provided by the embodiment of the present invention, various other geometric shapes that can implement the method can also be set according to actual needs.

[0050] The method comprises:

[0051] S101: When a passerby passes through the gate, light spot signals of each light source sensor in each time slice within a preset time interval are obtained; the preset time interval is composed of N time slices, where N is a positive integer;

[0052] In the method provided by an embodiment of the present invention, a passerby who is currently preparing to pass through the gate will pass by the position where the light curtain detection device is located. When the passerby passes through this position, the light source sensor in the light curtain detection device will generate a light source signal at a certain point in time, and the processor in the gate system can obtain the light source signal generated by the light source sensor.

[0053] When a passerby passes the location of the light curtain detection device, the time of his passing will form a time period. In the method provided in the embodiment of the present invention, a time interval is set according to the time when the passerby passes the light curtain detection device in combination with actual needs, that is, when executing the method, only the light point signals of each light source sensor in the time interval obtained by the processor are obtained.

[0054] In the method provided by the embodiment of the present invention, the preset time interval is composed of N time slices, each time slice can be understood as a time point, and the light point signals of each light source sensor obtained on the time slice are combined into a signal sequence.

[0055] S102: Determine a light spot matrix corresponding to the passerby according to the light spot signals of each light source sensor on each time slice, wherein the column data of the light spot matrix are the light spot signals of each light source sensor arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each light source sensor on each time slice arranged in a time order of each time slice;

[0056] In the method provided by an embodiment of the present invention, each light source sensor in the light curtain detection device is provided with a corresponding identification number. In this method, the identification numbers of each light source sensor can be sorted, that is, the sequence of the identification numbers of each light source sensor can be set, and the light source signals of each light source sensor obtained in each time slice are arranged in the sequence of the identification numbers, that is, a signal sequence of a light spot signal is obtained.

[0057] In the method provided by the embodiment of the present invention, the signal sequence obtained on each time slice is arranged in a vertical manner as a column of the light spot matrix, and a two-dimensional light spot matrix is ​​obtained by arranging them in the time sequence of each time slice. The data structure diagram of the light spot matrix provided by the embodiment of the present invention is shown in FIG. Figure 2 shown.

[0058] In the method provided by an embodiment of the present invention, taking a π-type light curtain detection device as an example, if the light curtain detection device has M light source sensors, M is a positive integer, then M light source signals are obtained in each time slice. Among the M light source signals, the light source signal generated by the light source sensor that the passerby passes through in the current time slice has an actual data value, and the value of the light source signal of the light source sensor that the passerby has not passed through in the current time slice can be recorded as 0.

[0059] exist Figure 2 It can be understood that the vertical column data are light source signals of M light source sensors, and the arrangement order from top to bottom is the order of the identification number of each light source sensor.

[0060] The horizontal row data is the column data arranged in the time sequence of N time slices.

[0061] S103: Input the light spot matrix into the established first neural network model, and after being processed by the first neural network model, output the passage status of the passerby corresponding to the light spot matrix, and then control the passage time of the gate according to the passage status of the passerby.

[0062] In the method provided in the embodiment of the present invention, a first neural network model is pre-established, which can identify the current passerby's pass status based on the arrangement of light spot signals in the light spot matrix, and then the processor can control the pass time of the gate according to the pass status of the passerby, and allocate different pass times to passers-by in the blocked state, thereby ensuring that every passer-by can pass through the gate smoothly, and at the same time, access control is set for the passage of unreasonable passers-by.

[0063] In the method provided in the embodiment of the present invention, the identified traffic status can be multiple modes, for example, the traffic status of the passer-by is single person passing, passing with luggage, passing with children, or multiple people passing simultaneously, etc. Different modes represent different traffic modes of the passer-by.

[0064] The traffic status detection method provided in the embodiment of the present invention can accurately identify the traffic status of the current passerby, and then determine the traffic time of the gate according to the traffic status of the passerby, so that the passerby can pass through the gate smoothly, avoiding the problem of pinching the passerby or pinching the luggage. At the same time, it can also prohibit the passage of multiple people, ensuring the safety and rationality of the passers-by passing through the gate.

[0065] In the method provided by the embodiment of the present invention, the preset identification order is the order in which the identification numbers of the light source sensors are arranged in sequence.

[0066] refer to Figure 3 , shows a flow chart of the method for establishing the first neural network model in an embodiment of the present invention, which specifically includes:

[0067] S201: Determine an initial network model;

[0068] In the method provided by the embodiment of the present invention, an initial network model is selected for establishing a first neural network model, and the initial network model may be a convolutional neural network model.

[0069] S202: Determine a data set for establishing the neural network model, the data set comprising: a training data set, a verification data set and a test data set; the training data set comprises a plurality of training light spot matrices and a first pass state corresponding to each of the training light spot matrices, the verification data set comprises a plurality of verification light spot matrices and a second pass state corresponding to each of the verification light spot matrices, the test data set comprises a plurality of test light spot matrices and a third pass state corresponding to each of the test light spot matrices; the data structures of the training light spot matrices, the verification light spot matrices and the test light spot matrices are the same;

[0070] In the method provided by the embodiment of the present invention, the data in the training data set, the verification data set and the test data set can all be obtained in advance by relevant passers-by through the light curtain detection device according to the predetermined pass status. And they are classified according to different uses, thereby obtaining the training data set, the verification data set and the test data set in the data set.

[0071] S203: using each training light spot matrix in the training data set as an input of the initial network model, and using the first traffic state corresponding to each training light spot matrix as an output of the initial network model, to train the initial network model;

[0072] S204: during the training process of the initial network model, when the loss function of the initial network model is less than a preset threshold, the training of the initial network model is terminated, and a second neural network model corresponding to the initial network model is generated;

[0073] In the method provided by an embodiment of the present invention, the training light spot matrix in the training data set is continuously input into the initial network model, and after the parameters of the initial network model are continuously adjusted, when the loss function of the initial network model is less than a certain threshold, the training of the initial network model is terminated and a corresponding second neural network model is generated. The second neural network model is an intermediate network model in the process of establishing the first neural network model.

[0074] In the method provided in the embodiment of the present invention, a loss function is set, and the loss function runs through the entire model building process.

[0075] S205: applying each verification light spot matrix in the verification data set and the second traffic state corresponding to each verification light spot matrix to verify the second neural network model until the qualified rate of the traffic state output by the second neural network model is greater than a preset qualified threshold, and generating a third neural network model corresponding to the second neural network model;

[0076] In the method provided by the embodiment of the present invention, the second neural network model is verified by using the verification light spot matrix in the verification data set. If the second neural network model passes the verification, the second neural network model is directly used as the third neural network model to enter the test process in S206. Otherwise, the model parameters of the second neural network model and the set loss function are adjusted until the qualified rate of the traffic state output by the adjusted second neural network model is greater than the preset qualified threshold value, and the adjusted second neural network model is determined as the third neural network model.

[0077] S206: Apply each test light spot matrix in the test data set and the third traffic state corresponding to each test light spot matrix to test the third neural network model. If the third neural network model is not in an overfitting state, the third neural network model is used as the neural network model for traffic state detection.

[0078] In the method provided by the embodiment of the present invention, during the process of testing the third neural network model, it is necessary to determine whether the third neural network model is in an overfitting state during the test process and is unable to effectively identify the traffic state.

[0079] When the network model of the third application is not overfitting, the final network model of the third application is used as the first neural network model for traffic status detection.

[0080] The method provided in the embodiment of the present invention further includes:

[0081] If the third neural network model is in an overfitting state, the model parameters of the third neural network model and the function parameters of the loss function are adjusted until the state of the adjusted current network model is no longer overfitting, and the adjusted current network model is used as the neural network model for traffic status detection.

[0082] refer to Figure 4 A flowchart showing a detailed resume process of the first neural network model provided by an embodiment of the present invention is shown. Figure 4 The initial network model can be a convolutional neural network CNN model.

[0083] exist Figure 4 In the execution process shown, during the verification process and the testing process, if the current network model does not meet the current processing conditions, the model parameters of the current network model and the function parameters of the loss function are adjusted until the readjusted network model meets the processing conditions and then proceeds to the next step of processing.

[0084] In the method provided in the embodiment of the present invention, the controlling of the passage time of the gate according to the passage status of the passer-by includes:

[0085] Determine the status type of the passerby's pass status;

[0086] Determining a preset travel time corresponding to the state type;

[0087] The gate is controlled to open and close according to the passage time.

[0088] In the method provided in the embodiment of the present invention, the passage status corresponds to a certain status type, that is, the status mode mentioned above, such as single person passage, passage with children, etc. The corresponding passage duration is set according to different status types, and then the opening and closing time of the gate can be controlled to allow the passers-by to pass smoothly.

[0089] In the method provided in the embodiment of the present invention, the passage status of the passers-by passing through the gate is pattern recognized through the deep learning artificial neural network model, which can accurately judge the passage status of the passers-by and distinguish whether the passers-by are in different states such as normal passage state, reasonable following state, unreasonable following state, etc.

[0090] In the method provided by an embodiment of the present invention, based on the spatially distributed light spot signal, the recorded data of a single time slice is a one-dimensional sequence of length M. The sampling frequency is set according to the actual situation. According to the set pass / fail judgment interval, the first N time slices are taken to form a two-dimensional matrix, in which the vertical direction is the light spot data of a single time slice, and the horizontal direction is represented as a time series, which is used as training and test data.

[0091] In the method provided in the embodiment of the present invention, the selection of the network model can be based on mobilenet, and its basic unit is depthwise separable convolution, that is, different convolution kernels are used for each input channel to perform calculations separately (depthwise convolution), and then a 1*1 convolution kernel is used for integration (pointwise convolution), which can greatly reduce the amount of calculation and model parameters.

[0092] According to the actual test accuracy and speed, the number of channels and the scale of the feature map of Mobilenet are reduced in a certain proportion, so as to ensure timeliness without reducing the accuracy.

[0093] In the method provided by the embodiment of the present invention, in order to ensure the output accuracy of the established neural network model, a corresponding loss function is set. This function is used to express the degree of difference between the model prediction value and the true value. Arc-SoftmaxLoss modified based on softmaxLoss (cross entropy loss function) is adopted here. The latter introduces the angle between classes, so that there is a certain angle interval between each class, and at the same time changes the interval to the comparison of the angle value, that is, the classification is affected by directly increasing the angle; the angle in the formula is the angle θ corresponding to the cross product of the weight and feature normalization, and the normalization makes the points on the feature mapped to the unit hypersphere, and introduces hyperparameters such as the spherical radius.

[0094] The specific formula is as follows:

[0095]

[0096] The embodiment of the present invention also provides a traffic state detection device corresponding to a traffic state detection method, and the traffic state detection device is used to implement the application of the traffic state detection method in practice. The structure diagram of the traffic state detection device can be referred to Figure 5 The device is applied to a gate system, which includes a gate and a light curtain detection device. The light curtain detection device is arranged on one side of the gate. A plurality of light source sensors are arranged in the light curtain detection device, and the plurality of light source sensors are distributed in a preset geometric shape.

[0097] The device comprises:

[0098] An acquisition unit 301 is used to acquire light spot signals of each of the light source sensors in each time slice within a preset time interval when a passerby passes through the gate; the preset time interval is composed of N time slices, where N is a positive integer;

[0099] A determination unit 302 is used to determine a light spot matrix corresponding to the passerby according to the light spot signals of each of the light source sensors on each of the time slices, wherein the column data of the light spot matrix are the light spot signals of each of the light source sensors arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each of the light source sensors on each time slice arranged in a time order of each of the time slices;

[0100] The output unit 303 is used to input the light spot matrix into the established first neural network model, and after being processed by the first neural network model, output the passage status of the passerby corresponding to the light spot matrix, and then control the passage time of the gate according to the passage status of the passerby.

[0101] The traffic status detection device provided by the embodiment of the present invention can accurately identify the traffic status of the current passerby, and then determine the traffic time of the gate according to the traffic status of the passerby, so that the passerby can pass through the gate smoothly, avoiding the problem of pinching the passerby or pinching the luggage. At the same time, it can also prohibit the passage of multiple people, ensuring the safety and rationality of the passers-by passing through the gate.

[0102] In the device provided by the embodiment of the present invention, the preset geometric shape is a π shape.

[0103] The traffic status detection device provided in the embodiment of the present invention includes a processor and a memory. The acquisition unit, the determination unit and the output unit are all stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.

[0104] The processor includes a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and the detection process of the traffic status can be dynamically executed by adjusting the kernel parameters to improve the update efficiency.

[0105] An embodiment of the present invention provides a storage medium on which a program is stored. When the program is executed by a processor, the traffic status detection method is implemented.

[0106] An embodiment of the present invention provides a processor, which is used to run a program, wherein the traffic status detection method is executed when the program is running.

[0107] refer to Figure 6 , shows a gate system provided by an embodiment of the present invention, the gate system comprising: a gate, a light curtain detection device and a gate processor;

[0108] The light curtain detection device is arranged on one side of the gate, and a plurality of light source sensors are arranged in the light curtain detection device, and the plurality of light source sensors are distributed in a π shape;

[0109] The gate processor acquires the light spot signals of each of the light source sensors in each time slice within a preset time interval when a passerby passes through the gate; the preset time interval is composed of N time slices, where N is a positive integer;

[0110] The gate processor determines the light spot matrix corresponding to the passer-by according to the light spot signals of each light source sensor on each time slice, wherein the column data of the light spot matrix are the light spot signals of each light source sensor arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each light source sensor on each time slice arranged in a time order of each time slice;

[0111] The gate processor inputs the light spot matrix into the established first neural network model, and after being processed by the first neural network model, outputs the passage status of the passerby corresponding to the light spot matrix, and then controls the passage time of the gate according to the passage status of the passerby.

[0112] like Figure 7 As shown, an embodiment of the present invention provides an electronic device, the electronic device 40 includes at least one processor 401, and at least one memory 402 and a bus 403 connected to the processor 401; wherein the processor 401 and the memory 402 communicate with each other through the bus 403; the processor 401 is used to call the program instructions in the memory 402 to execute the above-mentioned traffic status detection method. The device in this article can be a server, a PC, a PAD, a mobile phone, etc.

[0113] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that is initialized with the following method steps, the method being applied to a gate system, the gate system comprising a gate and a light curtain detection device, the light curtain detection device being arranged on one side of the gate, the light curtain detection device being arranged with a plurality of light source sensors, the plurality of light source sensors being distributed in a preset geometric shape; the method comprising:

[0114] When a passer-by passes through the gate, light spot signals of each light source sensor in each time slice within a preset time interval are obtained; the preset time interval is composed of N time slices, where N is a positive integer;

[0115] According to the light spot signals of each of the light source sensors on each of the time slices, a light spot matrix corresponding to the passer-by is determined, wherein the column data of the light spot matrix are the light spot signals of each of the light source sensors arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each of the light source sensors on each time slice arranged in a time order of each of the time slices;

[0116] The light spot matrix is ​​input into the established first neural network model. After being processed by the first neural network model, the passage status of the passerby corresponding to the light spot matrix is ​​output, and then the passage time of the gate is controlled according to the passage status of the passerby.

[0117] In the above method, optionally, the preset geometric shape is a π shape.

[0118] In the above method, optionally, the preset identification order is an order in which the identification numbers of the light source sensors are arranged in sequence.

[0119] In the above method, optionally, the process of establishing the first neural network model includes:

[0120] Determine the initial network model;

[0121] Determine a data set for establishing the neural network model, the data set comprising: a training data set, a verification data set and a test data set; the training data set comprises a plurality of training light spot matrices and a first pass state corresponding to each of the training light spot matrices, the verification data set comprises a plurality of verification light spot matrices and a second pass state corresponding to each of the verification light spot matrices, the test data set comprises a plurality of test light spot matrices and a third pass state corresponding to each of the test light spot matrices; the data structures of the training light spot matrices, the verification light spot matrices and the test light spot matrices are the same;

[0122] Taking each training light spot matrix in the training data set as the input of the initial network model, and taking the first traffic state corresponding to each training light spot matrix as the output of the initial network model, to train the initial network model;

[0123] During the training process of the initial network model, when the loss function of the initial network model is less than a preset threshold, the training of the initial network model is terminated, and a second neural network model corresponding to the initial network model is generated;

[0124] Using each verification light spot matrix in the verification data set and the second traffic state corresponding to each verification light spot matrix, the second neural network model is verified until the qualified rate of the traffic state output by the second neural network model is greater than a preset qualified threshold value, and a third neural network model corresponding to the second neural network model is generated;

[0125] The third neural network model is tested using each test light spot matrix in the test data set and the third traffic state corresponding to each test light spot matrix. If the third neural network model is not in an overfitting state, the third neural network model is used as the neural network model for traffic state detection.

[0126] The above method may optionally further include:

[0127] If the third neural network model is in an overfitting state, the model parameters of the third neural network model and the function parameters of the loss function are adjusted until the state of the adjusted current network model is no longer overfitting, and the adjusted current network model is used as the neural network model for traffic status detection.

[0128] In the above method, optionally, controlling the passage time of the gate according to the passage status of the passer-by includes:

[0129] Determine the status type of the passerby's pass status;

[0130] Determining a preset travel time corresponding to the state type;

[0131] The gate is controlled to open and close according to the passage time.

[0132] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0133] In a typical configuration, the device includes one or more processors (CPU), memory and bus. The device may also include input / output interface, network interface and the like.

[0134] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip. The memory is an example of a computer-readable medium.

[0135] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0136] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0137] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0138] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A traffic status detection method, characterized in that: The method is applied to a gate system, wherein the gate system includes a gate and a light curtain detection device, wherein the light curtain detection device is arranged on one side of the gate, wherein a plurality of light source sensors are arranged in the light curtain detection device, and wherein the plurality of light source sensors are distributed in a preset geometric shape; the method includes: When a passer-by passes through the gate, light spot signals of each light source sensor in each time slice within a preset time interval are obtained; the preset time interval is composed of N time slices, where N is a positive integer; According to the light spot signals of each of the light source sensors on each of the time slices, a light spot matrix corresponding to the passer-by is determined, wherein the column data of the light spot matrix are the light spot signals of each of the light source sensors arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each of the light source sensors on each time slice arranged in a time order of each of the time slices; The light spot matrix is ​​input into the established first neural network model. After being processed by the first neural network model, the passage status of the passerby corresponding to the light spot matrix is ​​output, and then the passage time of the gate is controlled according to the passage status of the passerby.

2. The method according to claim 1, characterized in that The preset geometric shape is a π shape.

3. The method according to claim 1, characterized in that The preset identification sequence is the sequence in which the identification numbers of the light source sensors are arranged in sequence.

4. The method according to claim 1, characterized in that: The process of establishing the first neural network model includes: Determine the initial network model; Determine a data set for establishing the neural network model, the data set comprising: a training data set, a verification data set and a test data set; the training data set comprises a plurality of training light spot matrices and a first pass state corresponding to each of the training light spot matrices, the verification data set comprises a plurality of verification light spot matrices and a second pass state corresponding to each of the verification light spot matrices, the test data set comprises a plurality of test light spot matrices and a third pass state corresponding to each of the test light spot matrices; the data structures of the training light spot matrices, the verification light spot matrices and the test light spot matrices are the same; Taking each training light spot matrix in the training data set as the input of the initial network model, and taking the first traffic state corresponding to each training light spot matrix as the output of the initial network model, to train the initial network model; During the training process of the initial network model, when the loss function of the initial network model is less than a preset threshold, the training of the initial network model is terminated, and a second neural network model corresponding to the initial network model is generated; Using each verification light spot matrix in the verification data set and the second traffic state corresponding to each verification light spot matrix, the second neural network model is verified until the qualified rate of the traffic state output by the second neural network model is greater than a preset qualified threshold value, and a third neural network model corresponding to the second neural network model is generated; The third neural network model is tested using each test light spot matrix in the test data set and the third traffic state corresponding to each test light spot matrix. If the third neural network model is not in an overfitting state, the third neural network model is used as the neural network model for traffic state detection.

5. The method according to claim 4, characterized in that Also includes: If the third neural network model is in an overfitting state, the model parameters of the third neural network model and the function parameters of the loss function are adjusted until the state of the adjusted current network model is no longer overfitting, and the adjusted current network model is used as the neural network model for traffic status detection.

6. The method according to claim 1, characterized in that The controlling of the passage time of the gate according to the passage status of the passer-by includes: Determine the status type of the passerby's pass status; Determining a preset travel time corresponding to the state type; The gate is controlled to open and close according to the passage time.

7. A traffic status detection device, characterized in that: The device is applied to a gate system, wherein the gate system includes a gate and a light curtain detection device, wherein the light curtain detection device is arranged on one side of the gate, wherein a plurality of light source sensors are arranged in the light curtain detection device, and wherein the plurality of light source sensors are distributed in a preset geometric shape; the device includes: An acquisition unit, used for acquiring light spot signals of each of the light source sensors in each time slice within a preset time interval when a passerby passes through the gate; the preset time interval is composed of N time slices, where N is a positive integer; A determination unit, used to determine the light spot matrix corresponding to the passerby according to the light spot signals of each of the light source sensors on each of the time slices, wherein the column data of the light spot matrix are the light spot signals of each of the light source sensors arranged in a preset identification order on each time slice, and the row data of the light spot matrix are the light spot signals of each of the light source sensors on each time slice arranged in a time order of each of the time slices; An output unit is used to input the light spot matrix into the established first neural network model, and after being processed by the first neural network model, output the passage status of the passerby corresponding to the light spot matrix, and then control the passage time of the gate according to the passage status of the passerby.

8. The device according to claim 7, characterized in that The preset geometric shape is a π shape.

9. A gate system, characterized in that: include: Gate, light curtain detection device and gate processor; The light curtain detection device is arranged on one side of the gate, and a plurality of light source sensors are arranged in the light curtain detection device, and the plurality of light source sensors are distributed in a π shape; The gate processor is used to execute the traffic status detection method as described in any one of claims 1-6.

10. An electronic device, characterized in that: It includes at least one processor, and at least one memory and a bus connected to the processor; wherein the processor and the memory communicate with each other through the bus; the processor is used to call the program instructions in the memory to execute the traffic status detection method as described in any one of claims 1 to 6.

Images