Target area detection method and system, vehicle, storage medium and program product

Abstract

The invention provides a target area detection method and system, a vehicle, a storage medium and a program product. The method comprises the following steps: predicting a detection target according to image information and a target detection model; determining a target distance between the vehicle and the detection target and a target speed of the detection target according to the detection target; and braking the vehicle according to the detection target, the target distance, the target speed and the vehicle information. According to the method, the detection target is predicted through the image information and the target detection model which is obtained by additionally using pedestrian head area training, so that the predicted detection target is more efficient, accurate and reliable, missing detection and error detection under dangerous working conditions are avoided, the position of the detection target is accurately obtained according to the target distance and the like, and the detection accuracy is improved. And whether the detection target collides with the vehicle or not is judged, and then vehicle braking is controlled, so that the purpose that the braking system can efficiently, accurately and reliably detect the detection target in the dangerous working condition is achieved, traffic accidents are reduced or avoided, and the driving safety of the vehicle is improved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle technology, and in particular to a target area detection method based on a braking system, a braking system, a vehicle, a computer-readable storage medium, and a computer program product. Background Art

[0002] Target area detection is very important for vehicle driving safety. Among the related technologies, the vehicle-based AEB (Autonomous Emergency Braking) system can detect targets in the vehicle's target area in real time through sensors, millimeter-wave radars, lidars or cameras, etc., such as obstacles and pedestrians in front of the vehicle. When it is judged that there is a risk of collision, the AEB system will be automatically triggered to avoid or reduce the occurrence of traffic accidents.

[0003] However, the target area detection method described above cannot efficiently, accurately and reliably detect the detection target in high-risk conditions, which reduces the driving safety of the vehicle. High-risk conditions, such as the "ghosting" condition, are conditions where the pedestrian's body is partially or completely blocked, making it difficult for the system to accurately detect. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0005] To this end, one object of the present invention is to propose a target area detection method based on a braking system, which enables the braking system to efficiently, accurately and reliably detect detection targets in dangerous conditions, reduce or avoid the occurrence of traffic accidents, and improve vehicle driving safety. It has low implementation cost and strong scalability, and can be widely used in the field of vehicle safety technology.

[0006] To this end, a second object of the present invention is to provide a braking system.

[0007] To this end, a third object of the present invention is to provide a vehicle.

[0008] To this end, a fourth object of the present invention is to provide a computer-readable storage medium.

[0009] To this end, a fifth object of the present invention is to provide a computer program product.

[0010] In order to achieve the above-mentioned objectives, an embodiment of the first aspect of the present invention proposes a target area detection method based on a braking system, the method comprising: predicting a detection target based on image information and a target detection model; determining a target distance between a vehicle and the detection target and a target speed of the detection target based on the detection target; and braking the vehicle based on the detection target, the target distance, the target speed and the vehicle information.

[0011] According to the target area detection method based on the braking system of the embodiment of the present invention, the detection target is predicted through image information and a target detection model obtained by additionally using pedestrian head area training, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous working conditions are avoided. Dangerous working conditions, such as the pedestrian's body being partially or completely blocked, make it difficult for the system to accurately detect the working conditions, and then determine the target distance between the vehicle and the detection target and the target speed of the detection target based on the detection target, and combine the vehicle information to accurately obtain the position of the detection target, and judge whether the detection target will collide with the vehicle, and then control the braking of the vehicle, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous working conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. Moreover, the implementation cost is low and the scalability is strong, and it can be widely used in the field of vehicle safety technology.

[0012] In some embodiments, predicting the detection target based on image information and a target detection model includes: acquiring image information during vehicle driving; preprocessing the image information to obtain a preprocessed image; and predicting the detection target based on the preprocessed image and the target detection model.

[0013] In some embodiments, predicting the detection target based on the preprocessed image and the target detection model includes: determining parameter information of the target to be detected based on the preprocessed image and the target detection model; and predicting the detection target based on the parameter information.

[0014] In some embodiments, before predicting the detection target based on image information and the target detection model, it also includes: acquiring original image data of the vehicle's driving road, a target detection model and input requirement parameters of the target detection model; performing feature extraction and feature enhancement on the original image data according to the feature extraction layer, feature enhancement layer and input requirement parameters of the target detection model to obtain enhanced features; determining the predicted detection result according to the output requirements of the target detection model and the enhanced features to perform model training.

[0015] In some embodiments, before obtaining the input requirement parameters of the target detection model, it also includes: labeling the original image data set composed of the original image data.

[0016] In some embodiments, determining the target distance between the vehicle and the detection target based on the detection target includes: obtaining prior height data of the pedestrian; and determining the target distance based on sensor parameter information, the prior height data of the pedestrian, and the detection frame information of the detection target based on a geometric algorithm.

[0017] In some embodiments, determining a target distance between the vehicle and the detection target according to the detection target includes: determining the target distance between the vehicle and the detection target based on a deep learning algorithm and the detection target.

[0018] In some embodiments, determining the target speed of the detection target based on the detection target includes: acquiring image features of the detection target; determining adjacent image frame features based on the image features of the detection target; and determining the target speed of the detection target based on the adjacent image frame features and a speed measurement learning algorithm.

[0019] In some embodiments, braking the vehicle according to the detected target, the target distance, the target speed and the vehicle information includes: determining a target area according to a risk assessment algorithm, the detected target, the target distance, the target speed and the vehicle information; obtaining a risk coefficient of the target area; and triggering the braking module to brake the vehicle when the risk coefficient is greater than a preset coefficient threshold.

[0020] In order to achieve the above-mentioned purpose, an embodiment of the second aspect of the present invention proposes a braking system, which includes: a target detection module, which is used to predict a detection target based on image information and a target detection model; a distance measurement module, which is connected to the target detection module, and is used to determine the target distance between the vehicle and the detection target based on the detection target; a speed measurement module, which is connected to the target detection module, and is used to determine the target speed of the detection target based on the detection target; a control module, which is respectively connected to the distance measurement module and the speed measurement module, and is used to issue a braking command based on the detection target, the target distance, the target speed and the vehicle information; and a braking module, which is connected to the control module, and is used to brake the vehicle when the braking command is received.

[0021] According to the braking system of the embodiment of the present invention, the detection target is predicted through image information and the target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous working conditions are avoided. Dangerous working conditions, such as the pedestrian's body being partially or completely blocked, making it difficult for the system to accurately detect the working conditions, and then the target distance between the vehicle and the detection target and the target speed of the detection target are determined according to the detection target. In combination with the vehicle information, the position of the detection target is accurately obtained, and it is determined whether the detection target will collide with the vehicle, and then the vehicle braking is controlled, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous working conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. Moreover, the implementation cost is low and the scalability is strong, and it can be widely used in the field of vehicle safety technology.

[0022] In order to achieve the above object, an embodiment of a third aspect of the present invention provides a vehicle, comprising: a braking system as required by the above embodiment.

[0023] According to the vehicle of the embodiment of the present invention, the detection target is predicted through image information and the target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous conditions are avoided. Dangerous conditions, such as the pedestrian's body is partially or completely blocked, making it difficult for the system to accurately detect the condition, and then the target distance between the vehicle and the detection target and the target speed of the detection target are determined according to the detection target. In combination with the vehicle information, the position of the detection target is accurately obtained, and it is determined whether the detection target will collide with the vehicle, and then the vehicle braking is controlled, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. It has low implementation cost and strong scalability, and can be widely used in the field of vehicle safety technology.

[0024] In order to achieve the above-mentioned objectives, an embodiment of the fourth aspect of the present invention proposes a computer-readable storage medium, on which a target area detection program based on a braking system is stored. When the target area detection program based on a braking system is executed by a processor, a device equipped with the target area detection program based on a braking system implements the target area detection method based on a braking system as described in the above-mentioned embodiment.

[0025] In order to achieve the above-mentioned purpose, an embodiment of the fifth aspect of the present invention proposes a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the target area detection method based on the braking system as described in the above-mentioned embodiment.

[0026] According to the computer program product of the embodiment of the present invention, the detection target is predicted through image information and the target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous conditions are avoided. Dangerous conditions, such as the pedestrian's body is partially or completely blocked, making it difficult for the system to accurately detect the working conditions. Then, the target distance between the vehicle and the detection target and the target speed of the detection target are determined based on the detection target. In combination with the vehicle information, the position of the detection target is accurately obtained, and it is determined whether the detection target will collide with the vehicle, and then the vehicle braking is controlled, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. Moreover, the implementation cost is low and the scalability is strong, and it can be widely used in the field of vehicle safety technology.

[0027] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which: Figure 1 is a schematic diagram of a high-risk working condition according to an embodiment of the present invention; Figure 2 is a flow chart of a target area detection method based on a braking system according to an embodiment of the present invention; Figure 3 is a schematic diagram of the structure of an AEB system based on pedestrian head area detection according to an embodiment of the present invention; Figure 4 is a flowchart of constructing and training a target detection module according to an embodiment of the present invention; Figure 5 is a target detection module prediction flow chart according to one embodiment of the present invention; Figure 6 is a structural diagram of a speed measurement and distance measurement module of a system according to an embodiment of the present invention; Figure 7 is a structural flow chart of a control module according to an embodiment of the present invention; Figure 8 is a flow chart of a target area detection method based on a braking system according to another embodiment of the present invention; Fig. 9 is a block diagram of a braking system according to an embodiment of the present invention; Fig.10 is a block diagram of a vehicle according to one embodiment of the present invention.

[0029] Reference numerals: AEB system backbone 50; Data preparation stage 53; obstacle perception model structure 54; Speed ​​measurement module 60; Distance measurement module 61; Target detection module 200; control module 203; braking module 204; Braking system 100; Vehicle 99. DETAILED DESCRIPTION

[0030] The embodiments described with reference to the drawings are exemplary, and embodiments of the present invention are described in detail below.

[0031] With the rapid development of the vehicle industry, vehicle safety issues are receiving more and more attention. As a type of active vehicle safety technology, the AEB system has been widely used for target area detection. The AEB system uses sensors, millimeter-wave radars, lidars or cameras to detect obstacles and pedestrians in front of the vehicle in real time. When it is determined that there is a risk of collision, the system will automatically trigger the braking system to avoid or reduce the occurrence of traffic accidents.

[0032] However, the above mainstream AEB technologies are prone to missed detection or false detection of obscured pedestrians, and there is still much room for improvement. Figure 1 FIG. 1 is a schematic diagram of a high-risk working condition according to an embodiment of the present invention. Figure 1 In the "ghost poking out of the head" condition, the pedestrian's body is partially or completely blocked, making it difficult for the system to accurately identify the pedestrian, thereby increasing the risk of traffic accidents.

[0033] In related technologies, for example, preventive control is adopted for scenarios with a high probability of "ghosting" conditions, so that control actions can be taken in advance before the vehicle recognizes pedestrians to reduce the possibility of traffic accidents. The method used includes: obtaining historical vehicle behavior data when encountering pedestrians, and establishing a vehicle control behavior evaluation model to predict the probability of a side vehicle encountering a pedestrian in front, and calculate the probability of pedestrians in the blind spot of the sidewalk; based on the probability of pedestrians in the blind spot of the sidewalk, it is determined whether a preventive control strategy needs to be adopted.

[0034] However, the above method is limited to the situation where non-pedestrian information can be obtained in advance, such as the behavior of other vehicles when encountering pedestrians, and preventive control of high-risk "ghosting" conditions. Without additional information and relying only on the vehicle itself, accidents cannot be avoided. The applicability and scalability of the method are poor.

[0035] For another example, when a vehicle blind spot is determined, the driver is reminded in advance to reduce accidents caused by blind spots. The method includes obtaining environmental information in a preset area and first driving information of the vehicle, determining a target area and attribute information corresponding to the target area from the preset area based on an electronic map and environmental information, determining target distance information between the vehicle and the target area, and determining warning information based on the first driving information, attribute information corresponding to the target area, and target distance information.

[0036] However, the above method can help the vehicle increase its attention in the "ghosting" condition by analyzing environmental information to determine whether there is a risk area in the current vehicle scenario, thereby reducing the risk of traffic accidents. However, under complex road conditions, as the risk area increases, the threshold of early warning cannot be grasped, which will cause problems such as frequent early warning or no early warning, and the reliability of the method is poor.

[0037] For another example, a system is composed of a camera, a positioning module, a wireless network module, a cloud processing module, and a receiving warning module. When the system is working: the camera obtains video information of pedestrians on the roadside, and determines the real-time location information of pedestrians through the positioning module; using 5G technology, the video information is transmitted to the cloud processing module in real time through the wireless network module installed inside the camera; the cloud processing module selects key information sources through screening algorithm calculations, and transmits the key information to the car running on the road; the car's receiving warning module receives the video information transmitted from the cloud processing module, and judges the possibility of danger based on its own edge computing technology.

[0038] However, the above method requires the installation of additional cameras along the entire road and the use of 5G communications and high-precision positioning technology, which greatly increases the equipment and technical costs of the project.

[0039] Therefore, the target area detection method based on the braking system according to the embodiment of the present invention is adopted.

[0040] Combine the following Figure 1-Figure 8 A target area detection method based on a braking system according to an embodiment of the present invention is described.

[0041] like Figure 2 FIG. 1 is a flow chart of a method for detecting a target area based on a braking system according to an embodiment of the present invention. The method for detecting a target area based on a braking system according to an embodiment of the present invention at least comprises steps S1 to S3.

[0042] Step S1, predicting the detection target based on image information and the target detection model.

[0043] In an embodiment, the image information is image data of the road on which the vehicle is located, acquired by a sensor on the vehicle, such as a camera, and the amount of image information collected is determined based on the specific task scale and task objectives; the target detection model is a model for predicting detection targets in image information, and the detection targets may be obstacles and pedestrians that may collide with the vehicle; taking pedestrians as an example, the camera on the vehicle acquires image information during driving, and predicts the detection targets based on the image information and the target detection model. The pedestrian head area detection technology is additionally used in the prediction of the detection targets by the target detection model, so that obstructed pedestrians can be more sensitively perceived, thereby avoiding problems such as missed detection and false detection that are prone to occur under traditional methods, and improving the accuracy and reliability of detection.

[0044] Step S2, determining a target distance between the vehicle and the detection target and a target speed of the detection target according to the detection target.

[0045] In the embodiment, the target distance is the distance between the vehicle and the detection target; the target speed is the speed of the detection target; taking the target as a person, the target distance between the vehicle and the detection target and the target speed of the detection target are determined according to the detection target. In determining the target distance, the visual ranging technology is used to achieve real-time and accurate acquisition of the pedestrian distance. In determining the target speed, the pure visual speed measurement technology is used to achieve rapid and accurate acquisition of pedestrian movement information, and only the visual ranging module installed inside the vehicle is used, and no other additional equipment support is required to achieve the determination of the target distance, which is relatively low in cost and easy to deploy and expand.

[0046] Step S3, braking the vehicle according to the detected target, target distance, target speed and vehicle information.

[0047] In an embodiment, vehicle information includes the speed of the vehicle itself; a risk estimation algorithm is used to encode information such as target speed, target distance, and the speed of the vehicle itself and combine the features of the encoding feature network to input a multi-layer perceptron, and finally a risk area is generated, and the risk coefficient of each target area is calculated to determine whether the detection target in the risk area will collide with the vehicle. If it is determined that the detection target in the risk area will collide with the vehicle, the braking module is triggered to brake the vehicle, thereby achieving the braking system that can efficiently, accurately and reliably detect the detection target in dangerous conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. This can be achieved only by relying on sensors inside the vehicle, with low cost and strong scalability, and can be widely used in the field of vehicle safety technology.

[0048] According to the target area detection method based on the braking system of the embodiment of the present invention, the detection target is predicted through image information and a target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous conditions are avoided. Dangerous conditions, such as the pedestrian's body being partially or completely blocked, make it difficult for the system to accurately detect the working conditions, and then determine the target distance between the vehicle and the detection target and the target speed of the detection target based on the detection target, and combine the vehicle information to accurately obtain the location of the detection target, and judge whether the detection target will collide with the vehicle, and then control the braking of the vehicle, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. It has low implementation cost and strong scalability, and can be widely used in the field of vehicle safety technology.

[0049] In some embodiments, predicting the detection target based on image information and a target detection model includes: acquiring image information during vehicle driving; preprocessing the image information to obtain a preprocessed image; and predicting the detection target based on the preprocessed image and a target detection model.

[0050] In an embodiment, the image includes a picture; image information is obtained during the driving process of the vehicle, for example, a forward-looking camera is used to collect forward-looking pictures in real time during driving, and the resolution, field of view (FOV), picture size, etc. of the sensor are determined to ensure that the collected pictures are clear and effective; data preprocessing is performed to obtain a preprocessed image. Before the picture is input into the target detection model, it is usually necessary to preprocess it. This is a key step to improve the performance and robustness of the model, and mainly includes processing the picture into the size required for the input of the model and normalizing the picture; the detection target is predicted based on the preprocessed image and the target detection model to achieve accurate detection of the target.

[0051] In some embodiments, predicting the detection target based on the preprocessed image and the target detection model includes: determining parameter information of the target to be detected based on the preprocessed image and the target detection model; and predicting the detection target based on the parameter information.

[0052] In an embodiment, the parameter information of the target to be detected includes the category of the target, the size and position of the detection frame, and the corresponding detection score; the parameter information of the target to be detected is determined based on the preprocessed image and the target detection model, for example, in the model prediction stage, the pre-trained model is used to predict the processed image, and the category, size and position of the detection frame, and the corresponding detection score of the target are obtained, so as to prepare data for sorting the targets according to the scores; the detection target is predicted based on the parameter information, for example, in the model output stage, for each input image, the final detection score is sorted according to the detection score and returned for use by subsequent modules. It can be understood that the higher the score, the more accurate the target, and the image features outputted in the last layer in the feature enhancement layer are saved so as to be combined with the image features of the next frame to predict the target speed. In contrast, the prior art usually uses laser radar perception, but cannot perceive the obscured pedestrians under occlusion conditions.

[0053] like Figure 5 FIG. 2 is a flowchart of a target detection module prediction process according to an embodiment of the present invention. For example, the target detection module prediction process at least includes steps S30 to S33.

[0054] Step S30, determine the required image information and collect the front view image in real time.

[0055] Step S31, data preprocessing: image size and normalization.

[0056] Step S32, model prediction: obtain category, detection box, and detection score.

[0057] Step S33, model output: sort the detection results and save the image features.

[0058] like Figure 3 As shown, it is a schematic diagram of the structure of an AEB system based on pedestrian head area detection according to an embodiment of the present invention. Step S12, target detection module: the model detects pedestrians, vehicles, faces, etc. Step S12 uses the pre-trained detection model to predict the captured forward-looking data, including the above steps S30-S33. Step S10 is associated with step S12, and step S11 optimizes step S12, specifically: Step S10, model preparation: data annotation, model construction, and model training.

[0059] Step S11, adding the target detection task of the pedestrian head area.

[0060] Step S12, target detection module: the model detects pedestrians, vehicles, faces, etc.

[0061] In some embodiments, before predicting the detection target based on image information and the target detection model, it also includes: obtaining original image data of the vehicle's driving road, a target detection model and input requirement parameters of the target detection model; performing feature extraction and feature enhancement on the original image data according to the feature extraction layer, feature enhancement layer and input requirement parameters of the target detection model to obtain enhanced features; determining the predicted detection results according to the output requirements and enhanced features of the target detection model to perform model training.

[0062] In the embodiment, the input requirement parameters are the input parameters required by the target detection model, including the input size of the image, the number of channels, and the number of predicted categories; the original data of the real road is collected to obtain the original image data of the vehicle driving on the road, and the target detection model and the input requirement parameters of the target detection model are obtained to provide a data basis for obtaining the detection target in the image information; the existing calibration technology is used to calibrate the sensor (such as a camera) used for collection to obtain the internal and external parameters of the sensor. The parameter calculation of the sensor is crucial, and the accuracy of the internal and external parameters will greatly affect the detection effect of the algorithm; the target detection model is reasonably designed according to the use needs of the model, the inference speed, etc. to determine the input size of the image, the number of channels, and the predicted number of categories. The parameters such as the number of categories measured; according to the feature extraction layer and input requirement parameters of the target detection model, the original image data is subjected to feature extraction. The performance of the feature extraction layer directly determines the prediction result of the model. For this purpose, the currently recognized high-performance feature extraction layer can be selected, such as ResNet (residual network), MobileNet (mobile network), CSPDarken (sequential communication process processing), etc., which can also be replaced by the later advanced feature extraction layer; according to the feature extraction layer, feature enhancement layer and input requirement parameters of the target detection model, the original image data is subjected to feature extraction and feature enhancement to obtain enhanced features. In the feature enhancement layer, a structure such as FPN (Feature Pyramid Network, Feature Pyramid Network), PAN (Path Aggregation Network), etc., to achieve feature representations of different scales. FPN can fuse features of different levels through upsampling and downsampling operations to adapt to different target sizes. PAN realizes cross-level fusion of features through lateral connection and pyramid pooling operations, so that the network can pay attention to information at different levels at the same time; according to the output requirements of the target detection model and the enhanced features, the predicted detection results are determined to perform model training. For example, a detection head is used to decode the required predictions from high-dimensional features according to the output requirements of the model. The detection head mainly adopts the form of a multi-layer perceptron (MLP), which takes the enhanced features as input and outputs the detection results of faces, pedestrian heads and other general obstacles.

[0063] like Figure 4FIG. 1 is a flowchart of target detection module construction and training according to an embodiment of the present invention. For example, the target detection module construction and training process according to the embodiment of the present invention at least includes steps S20 to S28.

[0064] In the data preparation stage 53, the following is performed: Step S20, collecting data collected by sensors during driving.

[0065] Step S21, sensor calibration and calculation of internal and external parameters.

[0066] Step S22: data set annotation.

[0067] In the obstacle perception model structure 54, the following is performed: Step S23, determining the image input size, the number of channels and the number of prediction categories.

[0068] Step S24, feature extraction layer: such as ResNet, MobileNet, CSPDarken, etc.

[0069] Step S25, feature enhancement layer: FPN, PAN, etc.

[0070] Step S26, detecting the head.

[0071] Step S27, result output: pedestrian head, face, and other obstacles.

[0072] After the obstacle perception model structure 54, the following is performed: Step S28, model training.

[0073] like Figure 3 As shown, it is a schematic diagram of the structure of an AEB system based on pedestrian head area detection according to an embodiment of the present invention. Step S10, model preparation: data annotation, model construction, and model training. In step S10, the main task is to obtain a trained target detection model. The target detection model uses a high-performance image processing chip and a deep learning algorithm to identify pedestrians, vehicles, faces, pedestrian head areas, etc. Step S10 includes the above-mentioned steps S20-S28.

[0074] In some embodiments, before obtaining the input requirement parameters of the target detection model, it also includes: annotating the original image data set composed of the original image data.

[0075] In an embodiment, the original image data in the collected original image data set is annotated to realize the training process of the supervised model; the quality of data annotation directly affects the training effect of the model, and the main contents of the annotation include: determining the number of annotation categories of the data; determining the format of the annotated data to better describe the annotated data; determining the content that needs to be annotated, including whether to annotate the distance, speed, target center point coordinates, etc.; the detailed process of annotation, determining the specific steps and methods of annotation, etc.; wherein, the annotation method includes different techniques such as manual annotation, automatic annotation or semi-automatic annotation.

[0076] In some embodiments, determining the target distance between the vehicle and the detection target based on the detection target includes: obtaining prior height data of the pedestrian; and determining the target distance based on the sensor parameter information, the prior height data of the pedestrian, and the detection frame information of the detection target based on a geometric algorithm.

[0077] In an embodiment, pedestrian prior height data is obtained to serve as a data basis for calculating the target distance; the target distance is determined according to the sensor parameter information, the pedestrian prior height data, and the detection frame information of the detection target using a geometric algorithm. For example, in the input preparation stage of the ranging module, the predicted information of the 2D detection frame in the predicted detection target is received from the detection module, the internal and external parameters of the calibrated sensor are obtained, and the target distance is calculated using a traditional 2D ranging algorithm in combination with the pedestrian prior height data, so as to realize a method of increasing the target distance calculation.

[0078] In some embodiments, determining a target distance between the vehicle and the detection target according to the detection target includes: determining the target distance between the vehicle and the detection target based on a deep learning algorithm and the detection target.

[0079] In an embodiment, the target distance between the vehicle and the detection target is determined based on a deep learning algorithm and a detection target. For example, a deep learning ranging algorithm is used. The deep learning algorithm adopts the current advanced pure visual obstacle detection model, combined with the detection target, and directly outputs the target distance between the vehicle and the detection target. Among them, the ranging algorithm includes a geometric algorithm and a deep learning algorithm, etc. This part also adopts a multi-algorithm joint prediction method for redundancy, which has the advantages of stability and reliability. In practical applications, more redundant schemes for calculating the target distance can be selected according to different scenarios and needs. For example, laser radar assisted ranging can be added in scenarios with higher performance requirements to improve the response speed and accuracy of the system. In contrast, the existing technology mainly uses radar, millimeter wave or ultrasonic wave for ranging, among which laser radar is easily affected by environmental interference; millimeter wave radar performs poorly when detecting pedestrians due to insufficient resolution, environmental clutter, obstacle occlusion, complex posture, motion blur and other problems, and is prone to false alarms; and the detection range and accuracy of ultrasonic wave are both low.

[0080] In some embodiments, determining a target speed of a detection target based on a detection target includes: acquiring image features of the detection target; determining adjacent image frame features based on the image features of the detection target; and determining a target speed of the detection target based on adjacent image frame features and a speed measurement learning algorithm.

[0081] In an embodiment, the image features are picture features output and saved by the target detection model; the image features of the detected target are obtained, for example, in the input preparation stage of the speed measurement module, the prediction results from the target detection model are received, and the picture features saved in the model output are obtained; the adjacent image frame features are determined according to the image features of the detected target, for example, the image features of the frames before and after the current image are obtained according to the image features, and the adjacent image frame features are obtained to realize the input of the deep learning network, and the frame rate of the currently captured image is determined to realize the calculation of the target speed using the traditional method; the target speed of the detected target is determined according to the adjacent image frame features and the speed measurement learning algorithm, for example, the speed measurement learning algorithm part adopts a multi-algorithm joint prediction method for redundancy, which can make the system more stable and can largely avoid false detection and missed detection. The algorithm includes: deep learning algorithm and optical flow algorithm, etc. Among them, the deep learning algorithm adopts a pure visual 3D obstacle detection model based on continuous frame time series fusion, which can directly output the target speed of the detected target. On the other hand, the optical flow algorithm can be used to directly calculate the target speed through the current mature optical flow algorithm combined with the time interval of continuous frames. In practical applications, redundancy solutions can be added according to different scenarios and requirements, such as using a fusion algorithm of vision and millimeter-wave radar in complex scenarios to improve the reliability and accuracy of the system and obtain accurate target speed. Compared with the existing technology of speed measurement through millimeter-wave radar, the poor perception of pedestrians leads to a high missed detection rate and low reliability.

[0082] like Figure 6 FIG. 1 is a structural diagram of a speed measurement and distance measurement module of a system according to an embodiment of the present invention. For example, the structural flow of the speed measurement and distance measurement module of the system according to the embodiment of the present invention at least includes steps S40 to S46.

[0083] Step S40, detecting the prediction result of the module.

[0084] Step S41, input preparation: obtaining the features of previous and next frames, frame rate, etc.

[0085] Step S42, speed measurement algorithm: deep learning network, optical flow calculation, etc.

[0086] Step S43, output the target speed.

[0087] Step S44, input preparation: obtain sensor parameters, adult prior height, etc.

[0088] Step S45, ranging algorithm: deep learning network, geometric calculation, etc.

[0089] Step S46, output the target distance.

[0090] Among them, steps S41 to S43 are performed in the speed measurement module 60, and steps S44 to S46 are performed in the distance measurement module 61. Figure 3 As shown, step S13 adds the prior adult height to measure the distance of the detected head area. Step S14, distance measurement and speed measurement module. Step S13 optimizes step S14. Step S14 includes the above steps S40-S46. Step S14 partially takes the target information detected by the target detection module, such as obstacle information, as input, and further analyzes to obtain the accurate position and movement speed of the target, that is, the target distance and target speed.

[0091] In some embodiments, braking a vehicle based on a detected target, a target distance, a target speed, and vehicle information includes: determining a target area based on a risk assessment algorithm, a detected target, a target distance, a target speed, and vehicle information; obtaining a risk coefficient of the target area; and triggering a braking module to brake the vehicle when the risk coefficient is greater than a preset coefficient threshold.

[0092] In an embodiment, the vehicle information includes the speed of the vehicle itself; the target area is the area where the detection target is located, and the number of areas depends on the specific situation. For example, the detection area is divided into 100 target areas, and the risk coefficient of the detection target colliding with the vehicle in each area is independent; the preset coefficient threshold is a critical value for judging whether the detection target will collide with the vehicle; the target area is determined according to the risk assessment algorithm, the detection target, the target distance, the target speed and the vehicle information. For example, a risk estimation algorithm is used, and the risk estimation algorithm adopts a redundant form of joint prediction of multiple algorithms to improve the stability and reliability of the system. The algorithm includes: a deep learning network, a Time To Collision (TTC) algorithm; the deep learning network uses an additional feature encoder and uses a multi-layer perceptron (MLP) network for risk prediction; the feature encoder is used to encode features from the target detection network to obtain encoded features; then the target speed, target distance, the speed of the vehicle itself and other information are encoded and input into the multi-layer perceptron in combination with the features of the encoded feature network, and finally a risk area, that is, a target area, is generated. Alternatively, the risk area can also be obtained by using the traditional TTC algorithm and calculating using vehicle information, target speed, target distance and other information; the risk coefficient of the target area is obtained, for example, the risk coefficient of each target area is calculated to determine whether the detection target in the risk area will collide with the vehicle; when the risk coefficient is greater than a preset coefficient threshold, it is considered that the detection target in the risk area will collide with the vehicle, and the braking module is triggered to brake the vehicle. For example, the obtained risk area is judged according to the preset coefficient threshold (preset threshold). If the risk coefficient is greater than the preset coefficient threshold, it is judged that there is a risk and a signal is sent to the braking module. The braking module of the system is the part that performs emergency braking. When the control module triggers the braking module, the braking module will immediately perform an emergency braking operation to slow down or stop the vehicle. Among them, the design of the brake module should ensure that it can respond to the command of the control module in a very short time and stop the vehicle in time. In actual applications, the brake module can adopt various forms of brake mechanisms, such as disc brakes or drum brakes, etc., which can be selected and matched according to the model and specifications of the vehicle.

[0093] like Figure 7 FIG. 1 is a structural flow chart of a control module according to an embodiment of the present invention. For example, the structural flow chart of the control module according to the embodiment of the present invention at least includes steps S50 to S54.

[0094] Step S50, target detection network feature output, target distance, target speed, and vehicle information.

[0095] Step S51, risk estimation algorithm: deep network, TTC algorithm, etc.

[0096] Step S52, determine whether the risk factor exceeds a preset threshold. If yes, execute step S54; otherwise, execute step S53.

[0097] Step S53: No signal is transmitted to the brake module to trigger braking.

[0098] Step S54: transmitting a signal to the brake module to trigger braking.

[0099] Combination Figure 3 As shown, step S15, control module: determine whether to perform emergency braking. Step S16, braking module: perform emergency braking. The control module in step S15 is the control center of the entire system, which is used to further analyze the obtained target distance and target speed to obtain the risk area during driving, and then judge the risk according to the preset coefficient threshold. When the judgment result is that there is a risk of collision, the control module will automatically trigger the braking module to perform emergency braking. The logic algorithm of the control module can be adjusted and optimized according to actual needs to improve the response speed and accuracy of the system. In actual applications, the false alarm rate and missed alarm rate of the system can be balanced by adjusting the safety distance threshold and the control logic algorithm to adapt to different road environments and driving habits. Step S15 includes the above-mentioned steps S50-step S54.

[0100] Reference below Figure 8 The target area detection method based on the braking system according to the embodiment of the present invention is described in detail.

[0101] like Figure 8 FIG. 2 is a flow chart of a method for detecting a target area based on a braking system according to another embodiment of the present invention. The method for detecting a target area based on a braking system according to the embodiment of the present invention at least includes steps S80-S96.

[0102] Step S80, annotating the original image data set consisting of the original image data.

[0103] Step S81, obtaining original image data of the road on which the vehicle is traveling, a target detection model, and input requirement parameters of the target detection model.

[0104] Step S82, performing feature extraction and feature enhancement on the original image data according to the feature extraction layer, feature enhancement layer and input requirement parameters of the target detection model to obtain enhanced features.

[0105] Step S83, determining the predicted detection result according to the output requirements and enhanced features of the target detection model to perform model training.

[0106] Step S84, obtaining image information of the vehicle during driving.

[0107] Step S85, preprocessing the image information to obtain a preprocessed image.

[0108] Step S86, determining parameter information of the target to be detected based on the preprocessed image and the target detection model.

[0109] Step S87, predicting the detection target according to the parameter information.

[0110] Step S88, obtaining pedestrian prior height data.

[0111] Step S89, determining the target distance based on a geometric algorithm according to sensor parameter information, pedestrian prior height data, and detection frame information of the detection target.

[0112] Step S90, determining a target distance between the vehicle and the detection target based on the deep learning algorithm and the detection target.

[0113] Step S91, obtaining image features of the detection target.

[0114] Step S92: determining the features of adjacent image frames according to the image features of the detection target.

[0115] Step S93, determining the target speed of the detection target according to the features of adjacent image frames and the speed measurement learning algorithm.

[0116] Step S94, determining the target area according to the risk assessment algorithm, the detection target, the target distance, the target speed and the vehicle information.

[0117] Step S95, obtaining the risk coefficient of the target area.

[0118] Step S96, when the risk coefficient is greater than a preset coefficient threshold, triggering a braking module to brake the vehicle.

[0119] According to the target area detection method based on the braking system of the embodiment of the present invention, the detection target is predicted through image information and a target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous conditions are avoided. Dangerous conditions, such as the pedestrian's body being partially or completely blocked, make it difficult for the system to accurately detect the working conditions, and then determine the target distance between the vehicle and the detection target and the target speed of the detection target based on the detection target, and combine the vehicle information to accurately obtain the location of the detection target, and judge whether the detection target will collide with the vehicle, and then control the braking of the vehicle, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. It has low implementation cost and strong scalability, and can be widely used in the field of vehicle safety technology.

[0120] Reference below Fig. 9 A braking system according to an embodiment of the present invention is described.

[0121] like Fig. 9 The figure shows a block diagram of a braking system according to an embodiment of the present invention. The braking system 100 of the embodiment of the present invention comprises: a target detection module 200, which is used to predict the detection target according to the image information and the target detection model; a distance measurement module 61, which is connected to the target detection module, and is used to determine the target distance between the vehicle and the detection target according to the detection target; a speed measurement module 60, which is connected to the target detection module 200, and is used to determine the target speed of the detection target according to the detection target; a control module 203, which is respectively connected to the distance measurement module 61 and the speed measurement module 60, and is used to issue a braking command according to the detection target, the target distance, the target speed and the vehicle information; and a braking module 204, which is connected to the control module 203, and is used to brake the vehicle when receiving the braking command.

[0122] Specifically, the target detection module 200 is used to detect obstacles in front of the vehicle in real time. Compared with the traditional target detection module, which detects general obstacles such as pedestrians and vehicles, the present invention adds an additional pedestrian head area detection task target to enhance the model's detection ability for obscured pedestrians. At the same time, the system also uses existing mature face recognition technology to further perceive obscured pedestrians. This module uses advanced computer vision and deep learning algorithms to capture images in front of the vehicle in real time through the front camera, and processes and analyzes the images to quickly and accurately detect targets.

[0123] Distance measurement module 61: A key module for sensing the distance between the detection target and the vehicle. Compared with the traditional laser radar, which has no penetration and cannot detect obscured targets; although the millimeter wave radar has penetration, the detection effect on pedestrians is poor. This module adopts visual distance measurement technology, which has the advantages of non-contact, non-destructive, and real-time monitoring, and can adapt to different lighting conditions and environmental changes. This module is mainly based on deep learning algorithms, and the traditional method is a redundant method for joint distance measurement. In this regard, the present invention optimizes the pure visual solution, adds a priori adult height information to perform geometric calculations on the target distance, and accurately determines the position and distance of pedestrians.

[0124] Speed ​​measurement module 60: A key module for obtaining the speed of obscured pedestrians. Compared with traditional technologies, millimeter-wave radar has a poorer perception of pedestrians. This module also uses a pure visual deep learning algorithm to measure speed using the features of the previous and next frame images. In addition, it uses an optical flow algorithm as redundancy to estimate the speed of pedestrians at the same time, so as to quickly and accurately perceive the dynamic information of pedestrians.

[0125] Control module 203: The control center of the system receives information from the target detection module 200, the distance measurement module 61 and the speed measurement module 60, and makes judgments based on preset thresholds. When the judgment result is that there is a risk of collision, the control module 203 will automatically trigger the braking module to perform emergency braking. The logic algorithm of the control module can be adjusted and optimized according to actual needs to improve the response speed and accuracy of the system.

[0126] Braking module 204: The part that performs emergency braking. When the control module 203 triggers the braking module 204, the braking module 204 will immediately perform emergency braking operations to slow down or stop the vehicle. The design of the braking module 204 should ensure that it can respond to the command of the control module in a very short time and stop the vehicle in time.

[0127] According to the braking system 100 of the embodiment of the present invention, the detection target is predicted through image information and the target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous working conditions are avoided. Dangerous working conditions, such as the pedestrian's body being partially or completely blocked, making it difficult for the system to accurately detect the working conditions, and then the target distance between the vehicle and the detection target and the target speed of the detection target are determined according to the detection target. In combination with the vehicle information, the position of the detection target is accurately obtained, and it is determined whether the detection target will collide with the vehicle, and then the vehicle braking is controlled, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous working conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. Moreover, the implementation cost is low and the scalability is strong, and it can be widely used in the field of vehicle safety technology.

[0128] In some embodiments, the control module 203 predicts the detection target based on the image information and the target detection model, including: acquiring image information during vehicle driving; preprocessing the image information to obtain a preprocessed image; and predicting the detection target based on the preprocessed image and the target detection model.

[0129] In some embodiments, the control module 203 predicts the detection target based on the preprocessed image and the target detection model, including: determining parameter information of the target to be detected based on the preprocessed image and the target detection model; and predicting the detection target based on the parameter information.

[0130] In some embodiments, before the control module 203 predicts the detection target based on the image information and the target detection model, it also includes: acquiring the original image data of the vehicle's driving road, the target detection model and the input requirement parameters of the target detection model; performing feature extraction and feature enhancement on the original image data according to the feature extraction layer, feature enhancement layer and input requirement parameters of the target detection model to obtain enhanced features; determining the predicted detection results according to the output requirements and enhanced features of the target detection model to perform model training.

[0131] In some embodiments, before the control module 203 obtains the input requirement parameters of the target detection model, it also includes: annotating the original image data set composed of the original image data.

[0132] In some embodiments, the control module 203 determines the target distance between the vehicle and the detection target based on the detection target, including: obtaining pedestrian prior height data; determining the target distance based on sensor parameter information, pedestrian prior height data, and detection frame information of the detection target based on a geometric algorithm.

[0133] In some embodiments, the control module 203 determines the target distance between the vehicle and the detection target according to the detection target, including: determining the target distance between the vehicle and the detection target based on a deep learning algorithm and the detection target.

[0134] In some embodiments, the control module 203 determines the target speed of the detection target based on the detection target, including: acquiring image features of the detection target; determining adjacent image frame features based on the image features of the detection target; and determining the target speed of the detection target based on the adjacent image frame features and a speed measurement learning algorithm.

[0135] In some embodiments, the control module 203 brakes the vehicle according to the detected target, target distance, target speed and vehicle information, including: determining the target area according to the risk assessment algorithm, the detected target, the target distance, the target speed and the vehicle information; obtaining the risk coefficient of the target area; when the risk coefficient is greater than a preset coefficient threshold, triggering the braking module to brake the vehicle.

[0136] According to the braking system 100 of the embodiment of the present invention, the detection target is predicted through image information and the target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous working conditions are avoided. Dangerous working conditions, such as the pedestrian's body being partially or completely blocked, making it difficult for the system to accurately detect the working conditions, and then the target distance between the vehicle and the detection target and the target speed of the detection target are determined according to the detection target. In combination with the vehicle information, the position of the detection target is accurately obtained, and it is determined whether the detection target will collide with the vehicle, and then the vehicle braking is controlled, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous working conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. Moreover, the implementation cost is low and the scalability is strong, and it can be widely used in the field of vehicle safety technology.

[0137] Reference below Fig.10 A vehicle 99 according to an embodiment of the present invention will be described.

[0138] like Fig.10As shown, it is a vehicle block diagram of an embodiment of the present invention, and the vehicle 99 includes: a braking system 100 as required by the above embodiment.

[0139] According to the vehicle 99 of the embodiment of the present invention, the detection target is predicted through image information and the target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous conditions are avoided. Dangerous conditions, such as the pedestrian's body is partially or completely blocked, making it difficult for the system to accurately detect the condition, and then the target distance between the vehicle and the detection target and the target speed of the detection target are determined according to the detection target. Combined with the vehicle information, the position of the detection target is accurately obtained, and it is determined whether the detection target will collide with the vehicle, and then the vehicle braking is controlled, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. Moreover, the implementation cost is low and the scalability is strong, and it can be widely used in the field of vehicle safety technology.

[0140] The following describes a computer-readable storage medium according to an embodiment of the present invention.

[0141] The computer-readable storage medium of an embodiment of the present invention stores a target area detection program based on a braking system. When the target area detection program based on a braking system is executed by a processor, a device equipped with the target area detection program based on a braking system implements the target area detection method based on a braking system as in the above-mentioned embodiment.

[0142] The following describes a computer program product according to an embodiment of the present invention.

[0143] The computer program product of the embodiment of the present invention includes a computer program. When the computer program is executed by a processor, the target area detection method based on the braking system as described in the above embodiment is implemented.

[0144] According to the computer program product of the embodiment of the present invention, the detection target is predicted through image information and the target detection model obtained by additional training using the pedestrian head area, so that the predicted detection target is more efficient, accurate and reliable, and missed detection and false detection in dangerous conditions are avoided. Dangerous conditions, such as the pedestrian's body is partially or completely blocked, making it difficult for the system to accurately detect the working conditions. The target distance between the vehicle and the detection target and the target speed of the detection target are determined according to the detection target. In combination with the vehicle information, the position of the detection target is accurately obtained, and it is determined whether the detection target will collide with the vehicle, and then the vehicle braking is controlled, so that the braking system can efficiently, accurately and reliably detect the detection target in dangerous conditions, reduce or avoid the occurrence of traffic accidents, and improve the driving safety of the vehicle. Moreover, the implementation cost is low and the scalability is strong, and it can be widely used in the field of vehicle safety technology.

[0145] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example.

[0146] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A target area detection method based on a braking system, characterized in that: include: Predict the detection target based on image information and target detection model; Determine a target distance between the vehicle and the detection target and a target speed of the detection target according to the detection target; The vehicle is braked according to the detected target, the target distance, the target speed and vehicle information.

2. The target area detection method based on the braking system according to claim 1, characterized in that: Predict the detection target based on the image information and the target detection model, including: Acquire image information during vehicle driving; Preprocessing the image information to obtain a preprocessed image; Predicting a detection target based on the preprocessed image and the target detection model.

3. The target area detection method based on the braking system according to claim 2, characterized in that: Predicting a detection target according to the preprocessed image and the target detection model includes: Determining parameter information of a target to be detected according to the preprocessed image and the target detection model; The detection target is predicted according to the parameter information.

4. The target area detection method based on the braking system according to claim 1, characterized in that: Before predicting the detection target based on image information and target detection model, it also includes: Acquire original image data of the road on which the vehicle is traveling, the target detection model, and input requirement parameters of the target detection model; Performing feature extraction and feature enhancement on the original image data according to the feature extraction layer, feature enhancement layer and the input requirement parameters of the target detection model to obtain enhanced features; The predicted detection result is determined according to the output requirements of the target detection model and the enhanced features to perform model training.

5. The target area detection method based on the braking system according to claim 4, characterized in that: Before obtaining the input requirement parameters of the target detection model, it also includes: An original image data set consisting of the original image data is labeled.

6. The target area detection method based on the braking system according to claim 1, characterized in that: Determining a target distance between the vehicle and the detection target according to the detection target includes: Obtain pedestrian prior height data; The target distance is determined based on a geometric algorithm according to sensor parameter information, the pedestrian prior height data, and the detection frame information of the detection target.

7. The target area detection method based on the braking system according to claim 1, characterized in that: Determining a target distance between the vehicle and the detection target according to the detection target includes: A target distance between the vehicle and the detection target is determined based on a deep learning algorithm and the detection target.

8. The target area detection method based on the braking system according to claim 1, characterized in that: Determining a target speed of the detection target according to the detection target includes: Obtain image features of the detection target; Determining adjacent image frame features according to the image features of the detection target; The target speed of the detection target is determined according to the adjacent image frame features and a speed measurement learning algorithm.

9. The target area detection method based on the braking system according to claim 1, characterized in that: Braking the vehicle according to the detected target, the target distance, the target speed and vehicle information, comprising: Determine a target area according to a risk assessment algorithm, the detection target, the target distance, the target speed and the vehicle information; Obtaining a risk factor of the target area; When the risk coefficient is greater than a preset coefficient threshold, the braking module is triggered to brake the vehicle.

10. A braking system, characterized in that: include: The target detection module is used to predict the detection target based on the image information and the target detection model; A distance measuring module, connected to the target detection module, for determining a target distance between the vehicle and the detection target according to the detection target; A speed measurement module, connected to the target detection module, and configured to determine a target speed of the detection target according to the detection target; A control module, connected to the distance measuring module and the speed measuring module respectively, and configured to issue a braking instruction according to the detected target, the target distance, the target speed and vehicle information; A braking module is connected to the control module and is used to brake the vehicle when receiving the braking instruction.

11. A vehicle, characterized in that: include: The braking system of claim 10.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a target area detection program based on a braking system. When the target area detection program based on a braking system is executed by a processor, a device installed with the target area detection program based on a braking system implements the target area detection method based on a braking system as described in any one of claims 1-9.

13. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method for detecting a target area based on a braking system according to any one of claims 1 to 9 is implemented.

Images