Target screening method and system based on machine learning, vehicle and equipment

Abstract

The invention discloses a target screening method and system based on machine learning, a vehicle and equipment. The target screening method based on machine learning comprises the following steps: obtaining road perception data; inputting the road perception data into a machine learning model to obtain a target obstacle screened from the road perception data by the machine learning model, the machine learning model being pre-trained based on an actual road obstacle data set with a label, the label being used for representing the target obstacle; judging whether the target obstacle reaches a preset threat level or not according to the driving track of the vehicle and the information of the screened target obstacle; and if yes, outputting the target obstacle as a final screening target. The target obstacle screening method and device have the advantage of high target screening accuracy, target obstacle screening can be more accurately and efficiently carried out, and the driving comfort is effectively improved.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and in particular to a target screening method, system, vehicle and equipment based on machine learning. Background Art

[0002] With the vigorous development of the automobile industry, intelligent driving has gradually entered people's lives. As an important part of assisted intelligent driving, Adaptive Cruise Control (ACC) technology has received widespread attention. So far, the ACC system has been fully proven to have the function of reducing driver fatigue and improving vehicle driving safety, and has been widely used. However, due to the inaccuracy of the camera's visual perception results and the irrationality of target screening, ghost braking and brake failure may occur when the driver follows the vehicle in front. In some cases, the target vehicle is about to enter or exit the trajectory range of the vehicle. Since only the existing state is considered during screening, it is easy to cause braking or acceleration that contradicts the habits of human drivers. That is, ghost braking and brake failure may occur when the driver follows the vehicle in front. For example, when the vehicle in front of you on the left suddenly squeezes into the lane you are driving, the ACC system will change the target to follow from the vehicle in front to the vehicle in front of you on the left, causing your car to decelerate suddenly. Or when the vehicle in front of you on the left suddenly does not want to squeeze into the lane you are driving, the ACC system will change the target to follow from the vehicle in front of you on the left back to the vehicle in front of you on the left, causing your car to accelerate suddenly. Sudden acceleration or deceleration will make the driver feel uncomfortable or even highly nervous. For another example, when the vehicle in front of you that you are following suddenly turns, the ACC system will lose the target to follow, causing it to accelerate suddenly, which will also make the driver feel uncomfortable. Summary of the invention

[0003] Based on this, it is necessary to provide a target screening method, system, vehicle and equipment based on machine learning to address the above technical problems. By screening targets through a trained model, it can well handle the screening problems of different targets under different working conditions. It has the advantage of high target screening accuracy. After the model screens the target, it verifies whether the target is a threatening target obstacle through the vehicle's driving trajectory. Therefore, it can screen target obstacles more accurately and efficiently while ensuring safety, effectively improving driving comfort.

[0004] In a first aspect, a target screening method based on machine learning is provided, comprising:

[0005] Obtain road perception data;

[0006] Inputting the road perception data into a machine learning model to obtain a target obstacle screened out by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles;

[0007] According to the driving trajectory of the vehicle and the filtered information of the target obstacle, determining whether the target obstacle reaches a predetermined threat level;

[0008] If yes, the target obstacle is output as the final screening target.

[0009] In some examples, before inputting the road perception data into a machine learning model to obtain a target obstacle screened out by the machine learning model from the road perception data, the method includes:

[0010] Obtaining actual road obstacle data, and performing labeling processing on the actual road obstacle data to obtain the labeled actual road obstacle data set;

[0011] An initial machine learning model is trained based on the labeled actual road obstacle data set to obtain the trained machine learning model.

[0012] In some examples, training an initial machine learning model according to the labeled actual road obstacle dataset to obtain the trained machine learning model includes:

[0013] Sequentially performing normalization processing, data filling processing, and data time serialization processing on the data in the actual road obstacle data set;

[0014] Perform data feature extraction and association operations on the processed labeled actual road obstacle data set;

[0015] After the data feature extraction and association operation is completed, the data is input into the initial machine learning model to train the initial machine learning model until the trained machine learning model is obtained.

[0016] In some examples, the training of the initial machine learning model until the trained machine learning model is obtained includes:

[0017] According to the output of the initial machine learning model and the loss before the label, the parameters of the initial machine learning model are iteratively updated until the loss reaches the desired target, thereby obtaining the trained machine learning model.

[0018] In some examples, after obtaining the machine learning model after training, the method further includes:

[0019] Obtain test data, and test the machine learning model based on the test data.

[0020] In some examples, judging whether the target obstacle reaches a predetermined threat level based on the driving trajectory of the vehicle and the filtered information of the target obstacle includes:

[0021] Obtaining the horizontal coordinate and the vertical coordinate of the target obstacle according to the information of the target obstacle;

[0022] Obtaining a threshold range according to the driving trajectory of the vehicle;

[0023] Determining whether the horizontal coordinate and the vertical coordinate of the target obstacle are within the threshold range;

[0024] If yes, it is determined that the target obstacle reaches the predetermined threat level.

[0025] In some examples, the machine learning model is a machine learning model of a comparison target XGBoost based on a shared LSTM network.

[0026] In a second aspect, a target screening system based on machine learning is provided, comprising:

[0027] An acquisition module, used to obtain road perception data;

[0028] a screening module, configured to input the road perception data into a machine learning model to obtain a target obstacle screened by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles;

[0029] The output module is used to determine whether the target obstacle reaches a predetermined threat level according to the driving trajectory of the vehicle and the information of the screened target obstacle, and output the target obstacle as a final screening target when the target obstacle reaches the predetermined threat level.

[0030] In a third aspect, a vehicle is provided, comprising: a target screening system based on machine learning according to the second aspect described above.

[0031] In a fourth aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the target screening method based on machine learning in the above-mentioned first aspect and any possible implementation method of the first aspect are implemented.

[0032] In a fifth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the steps of the target screening method based on machine learning in the above-mentioned first aspect and any possible implementation method of the first aspect are implemented.

[0033] In a sixth aspect, a computer program product is provided, on which a computer program is stored. When the program is executed by a processor, the steps of the target screening method based on machine learning in the above-mentioned first aspect and any possible implementation method of the first aspect are implemented.

[0034] By adopting the embodiment of the present application, road perception data is obtained, and the road perception data is input into the machine learning model to obtain the target obstacle screened out by the machine learning model from the road perception data, and according to the driving trajectory of the vehicle and the information of the screened target obstacle, it is determined that when the target obstacle reaches a predetermined threat level, the target obstacle is output as the final screening target. By screening targets with a trained model, it is possible to handle the screening problem of different targets under different working conditions well, and it has the advantage of high target screening accuracy. After the model screens out the target, it verifies whether the target is a threatening target obstacle through the driving trajectory of the vehicle. Therefore, it is possible to screen target obstacles more accurately and efficiently while ensuring safety, effectively improving driving comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0036] Figure 1 A flow chart of a target screening method based on machine learning provided in an embodiment of the present application;

[0037] Figure 2 A schematic diagram of a machine learning model screening target provided in an embodiment of the present application;

[0038] Figure 3 A training and verification flow chart of a machine learning model provided in an embodiment of the present application;

[0039] Figure 4 A schematic diagram of the model structure of the machine learning model provided in the embodiment of the present application;

[0040] Figure 5 A schematic diagram of the final target determination process of the target screening method based on machine learning provided in an embodiment of the present application;

[0041] Figure 6 A structural block diagram of a target screening system based on machine learning provided in an embodiment of the present application;

[0042] Figure 7 A structural block diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] The present application is further described in detail below in conjunction with the embodiments and drawings. It is to be understood that the specific embodiments described herein are only used to explain the relevant application, rather than to limit the application. It is also necessary to explain that, for ease of description, only the parts related to the application are shown in the drawings.

[0044] It should be noted that, in the absence of conflict, the embodiments of the present application, that is, the features of the embodiments, can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0045] The following describes in detail the target screening method, system, device and medium based on machine learning according to the embodiments of the present application in conjunction with the accompanying drawings.

[0046] Figure 1 : is a flow chart of a target screening method based on machine learning according to an embodiment of the present application. Figure 1 As shown, the target screening method based on machine learning according to the embodiment of the present application includes the following steps:

[0047] S101: Obtain road perception data, for example, road environment images during driving.

[0048] S102: Input the road perception data into a machine learning model to obtain a target obstacle screened out by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles.

[0049] It should be noted that the machine learning model is pre-trained. Therefore, before the road perception data is input into the machine learning model and the target obstacles are screened out by the machine learning model from the road perception data, it includes: obtaining actual road obstacle data and labeling the actual road obstacle data to obtain the labeled actual road obstacle data set; training the initial machine learning model according to the labeled actual road obstacle data set to obtain the trained machine learning model.

[0050] In this example, an initial machine learning model is trained according to the labeled actual road obstacle data set to obtain the trained machine learning model, including: normalizing, filling and time serializing the data in the actual road obstacle data set in sequence; performing data feature extraction and association operations on the processed labeled actual road obstacle data set; and inputting the data after the data feature extraction and association operations into the initial machine learning model to train the initial machine learning model until the trained machine learning model is obtained.

[0051] Furthermore, the initial machine learning model is trained until the trained machine learning model is obtained, including: iteratively updating the parameters of the initial machine learning model according to the output of the initial machine learning model and the loss before the label, until the loss reaches the expected target, thereby obtaining the trained machine learning model.

[0052] After the training is completed, it is necessary to verify whether it meets the requirements. Therefore, in an embodiment of the present application, after obtaining the machine learning model after the training is completed, obtaining test data, and testing the machine learning model based on the test data.

[0053] As a specific application, the machine learning model takes the machine learning model of XGBoost based on the shared LSTM network as an example, and builds the machine learning model of catBoost based on the shared LSTM network. Then the data set collected from the real vehicle is labeled, and the machine learning model is trained through the processed data. The trained model can well handle the screening problem of different targets under different working conditions.

[0054] like Figure 2 As shown in the figure, the machine learning model of the comparison target catBoost based on the shared LSTM network is based on the actual road obstacle data set as input, and after the LSTM shared network comparison target catBoost learning model training, the training results are output to the target arbitration module. After the arbitration module arbitration strategy, the filtered target obstacle information is finally output.

[0055] like Figure 3 As shown, the workflow of the machine learning model of the comparison target catBoost based on the shared LSTM network is as follows:

[0056] The actual road obstacle data collected by the real vehicle is manually labeled and packaged into a labeled real vehicle obstacle data set according to the set data format as training data.

[0057] The first step is to perform data normalization to normalize the scattered data to the [0, 1] interval.

[0058] The second step is to fill in the data and manually fill the blank parts in the data structure caused by the missing number of obstacles with the default values.

[0059] The third step is to perform data time serialization, associate the original obstacle data samples with the time data, and obtain time series data samples for data training of the LSTM network.

[0060] The fourth step is to extract and associate data features. The obstacle information of different groups is associated according to the positional relationship. For obstacles with similar distances or similar lanes, the associated label is set to 1; for obstacles with far distances or non-adjacent lanes, the associated label is set to 0.

[0061] In the fifth step, the data is input into the machine learning model for training, and the stochastic gradient descent algorithm is used to learn the parameters of the comparison target catBoost model based on the shared LSTM network after establishment by continuously reducing the function value of the evaluation function, and the parameters are learned and updated in each iteration.

[0062] The sixth step is to output the test data for verification of the trained model and output the filtered target data.

[0063] like Figure 4 As shown in the figure, the shared network structure of the comparison target catBoost model based on the shared LSTM network is shown. The input layer of the model consists of two parts: the vehicle information and the surrounding obstacle vehicle information. The vehicle information mainly includes: the vehicle speed, steering wheel angle, yaw rate output by EBS, turn signal, accelerator pedal opening, brake pedal opening and other information; the obstacle vehicle information mainly includes: the lateral position coordinates, longitudinal position coordinates, lateral speed, longitudinal speed, lateral acceleration, longitudinal acceleration, turn signal information, brake light information, double flash warning light information and other information of surrounding obstacles. This information is grouped into a unified data structure for each individual as the input layer information of the shared model. These data sets are given as input to the hidden layer.

[0064] The hidden layer of the shared network structure of the embodiment of the present application includes three groups of fully connected layers: the first fully connected layer is a double-layer shared fully connected layer with activation functions of ArcTan and ISRU respectively; the fully connected layer outputs of the self-vehicle and the obstacle vehicle are coupled respectively. The result is output to the second fully connected layer, which is a double-layer shared fully connected layer with activation functions of SELU and Softsign respectively; the outputs of the fully connected layers are decoupled and output to the third fully connected layer respectively. The third fully connected layer is a single-layer shared fully connected layer with an activation function of SoftPlus, and the output result of the fully connected layer is given to the output layer as the final result.

[0065] S103: judging whether the target obstacle reaches a predetermined threat level according to the driving trajectory of the vehicle and the filtered information of the target obstacle.

[0066] In one embodiment of the present application, judging whether the target obstacle reaches a predetermined threat level according to the driving trajectory of the vehicle and the information of the filtered target obstacle includes: obtaining the horizontal coordinate and the vertical coordinate of the target obstacle according to the information of the target obstacle; obtaining a threshold range according to the driving trajectory of the vehicle; judging whether the horizontal coordinate and the vertical coordinate of the target obstacle are within the threshold range; if so, determining that the target obstacle reaches the predetermined threat level.

[0067] S104: If yes, output the target obstacle as the final screening target.

[0068] Specifically, the final output result indicates the most threatening target obstacle. Whether the target obstacle output by the machine learning model poses a threat to the vehicle's driving trajectory needs further verification to improve the accuracy of the result. Figure 5 As shown:

[0069] For the targets selected by machine learning, the position and speed information of the horizontal and vertical coordinates of the target, as well as the driving trajectory information of the vehicle, are extracted. The position and speed information of the target is information in the XY coordinate system, which needs to be converted into the SL coordinate system of the vehicle. After the coordinate conversion, if the SL coordinate and speed results are within the preset threshold range, the target is a threatening target and is output as the final screening target. If it is not within the preset threshold range, the target is not threatening and is not output.

[0070] According to the target screening method based on machine learning of the embodiment of the present application, road perception data is obtained, and the road perception data is input into the machine learning model to obtain the target obstacle screened out by the machine learning model from the road perception data, and according to the driving trajectory of the vehicle and the information of the screened target obstacle, it is determined that the target obstacle reaches a predetermined threat level and outputs the target obstacle as the final screening target. By screening targets with a trained model, it is possible to handle the screening problem of different targets under different working conditions well, and it has the advantage of high target screening accuracy. After the model screens out the target, it verifies whether the target is a threatening target obstacle through the driving trajectory of the vehicle. Therefore, it is possible to screen target obstacles more accurately and efficiently while ensuring safety, effectively improving driving comfort.

[0071] Figure 6 is a structural block diagram of a target screening system based on machine learning according to an embodiment of the present application. Figure 6 As shown, according to an embodiment of the present application, a target screening system based on machine learning includes: an acquisition module 610, a screening module 620 and an output module 630, wherein:

[0072] An acquisition module 610 is used to obtain road perception data;

[0073] A screening module 620, configured to input the road perception data into a machine learning model to obtain a target obstacle screened by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles;

[0074] The output module 630 is used to determine whether the target obstacle reaches a predetermined threat level according to the driving trajectory of the vehicle and the information of the screened target obstacle, and output the target obstacle as a final screening target when the target obstacle reaches the predetermined threat level.

[0075] According to the machine learning-based target screening system of the embodiment of the present application, road perception data is obtained, and the road perception data is input into the machine learning model to obtain the target obstacle screened out by the machine learning model from the road perception data, and according to the driving trajectory of the vehicle and the information of the screened target obstacle, it is determined that the target obstacle reaches a predetermined threat level and outputs the target obstacle as the final screening target. By screening targets with a trained model, it is possible to handle the screening problem of different targets under different working conditions well, and has the advantage of high target screening accuracy. After the model screens out the target, it verifies whether the target is a threatening target obstacle through the driving trajectory of the vehicle. Therefore, it is possible to screen target obstacles more accurately and efficiently while ensuring safety, effectively improving driving comfort.

[0076] For the specific definition of the target screening system based on machine learning, please refer to the definition of the target screening method based on machine learning above, which will not be repeated here. The various modules of the above-mentioned target screening system based on machine learning can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned various modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0077] In one embodiment, a vehicle is provided, comprising: a target screening system based on machine learning according to any of the above embodiments. After obtaining road perception data, the vehicle inputs the road perception data into a machine learning model to obtain target obstacles screened out by the machine learning model from the road perception data, and determines that the target obstacle reaches a predetermined threat level based on the vehicle's driving trajectory and the information of the screened target obstacles, and outputs the target obstacle as the final screening target. By screening targets with a trained model, it is possible to handle the screening problem of different targets under different working conditions, and has the advantage of high target screening accuracy. After the model screens out the target, it verifies whether the target is a threatening target obstacle through the vehicle's driving trajectory. Therefore, it is possible to screen target obstacles more accurately and efficiently while ensuring safety, and effectively improve driving comfort.

[0078] In addition, other structures and functions of the vehicle according to the embodiment of the present application are known to ordinary technicians in the field and will not be elaborated here.

[0079] In one embodiment, a computer device is provided. Figure 7 This is a block diagram of the computer device provided in the embodiment of the present application, refer to Figure 7 The computer device includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the aforementioned target screening method embodiment based on machine learning is implemented. For example, the following steps are performed: obtaining road perception data;

[0080] Inputting the road perception data into a machine learning model to obtain a target obstacle screened out by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles;

[0081] According to the driving trajectory of the vehicle and the filtered information of the target obstacle, determining whether the target obstacle reaches a predetermined threat level;

[0082] If yes, the target obstacle is output as the final screening target.

[0083] The embodiment of the present application further provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and the processor executes the computer program to implement the aforementioned target screening method embodiment based on machine learning. For example, the following steps are performed: obtaining road perception data;

[0084] Inputting the road perception data into a machine learning model to obtain a target obstacle screened out by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles;

[0085] According to the driving trajectory of the vehicle and the filtered information of the target obstacle, determining whether the target obstacle reaches a predetermined threat level;

[0086] If yes, the target obstacle is output as the final screening target.

[0087] The present application embodiment provides a computer program product, which includes instructions. When the instructions are executed, the method described in the embodiment of the present application is executed. For example, it can be executed Figure 1 The various steps of the machine learning-based target screening method shown, for example, perform: obtaining road perception data;

[0088] Inputting the road perception data into a machine learning model to obtain a target obstacle screened out by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles;

[0089] According to the driving trajectory of the vehicle and the filtered information of the target obstacle, determining whether the target obstacle reaches a predetermined threat level;

[0090] If yes, the target obstacle is output as the final screening target.

[0091] Those of ordinary skill in the art can understand that all or part of the processes in the methods for implementing the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0092] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0093] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A target screening method based on machine learning, characterized in that: include: Obtain road perception data; Inputting the road perception data into a machine learning model to obtain a target obstacle screened out by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles; According to the driving trajectory of the vehicle and the filtered information of the target obstacle, determining whether the target obstacle reaches a predetermined threat level; If yes, the target obstacle is output as the final screening target.

2. The target screening method based on machine learning according to claim 1, characterized in that: Before inputting the road perception data into the machine learning model to obtain the target obstacle screened out from the road perception data by the machine learning model, the method includes: Obtaining actual road obstacle data, and performing labeling processing on the actual road obstacle data to obtain the labeled actual road obstacle data set; An initial machine learning model is trained based on the labeled actual road obstacle data set to obtain the trained machine learning model.

3. The target screening method based on machine learning according to claim 2, characterized in that: The step of training an initial machine learning model according to the labeled actual road obstacle data set to obtain the trained machine learning model includes: Sequentially performing normalization processing, data filling processing, and data time serialization processing on the data in the actual road obstacle data set; Perform data feature extraction and association operations on the processed labeled actual road obstacle data set; After the data feature extraction and association operation is completed, the data is input into the initial machine learning model to train the initial machine learning model until the trained machine learning model is obtained.

4. The target screening method based on machine learning according to claim 3, characterized in that: The step of training the initial machine learning model until a trained machine learning model is obtained includes: According to the output of the initial machine learning model and the loss before the label, the parameters of the initial machine learning model are iteratively updated until the loss reaches the desired target, thereby obtaining the trained machine learning model.

5. The target screening method based on machine learning according to claim 4, characterized in that: After the machine learning model is trained, it also includes: Obtain test data, and test the machine learning model based on the test data.

6. The target screening method based on machine learning according to any one of claims 1 to 5, characterized in that: The step of judging whether the target obstacle reaches a predetermined threat level according to the driving trajectory of the vehicle and the filtered information of the target obstacle includes: Obtaining the horizontal coordinate and the vertical coordinate of the target obstacle according to the information of the target obstacle; Obtaining a threshold range according to the driving trajectory of the vehicle; Determining whether the horizontal coordinate and the vertical coordinate of the target obstacle are within the threshold range; If yes, it is determined that the target obstacle reaches the predetermined threat level.

7. The target screening method based on machine learning according to claim 1, characterized in that: The machine learning model is a machine learning model of the comparison target XGBoost based on a shared LSTM network.

8. A target screening system based on machine learning, characterized in that: include: An acquisition module, used to obtain road perception data; a screening module, configured to input the road perception data into a machine learning model to obtain a target obstacle screened by the machine learning model from the road perception data, wherein the machine learning model is pre-trained based on an actual road obstacle data set with labels, and the labels are used to characterize the target obstacles; The output module is used to determine whether the target obstacle reaches a predetermined threat level according to the driving trajectory of the vehicle and the information of the screened target obstacle, and output the target obstacle as a final screening target when the target obstacle reaches the predetermined threat level.

9. A vehicle, characterized in that: include: The machine learning-based target screening system according to claim 8.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the target screening method based on machine learning according to any one of claims 1-7 is implemented.

Images