Blind spot detection method and blind spot detection sensor

Abstract

The present invention discloses a blind spot detection method and a blind spot detection sensor, which belongs to the field of blind spot detection technology, and includes the following steps: S1: collecting real-time data of vehicle blind spot detection, processing the real-time data of vehicle blind spot detection, and determining vehicle blind spot detection characteristic data; S2: determining the optimal blind spot detection sensing model; S3: based on the optimal blind spot detection sensing model, predicting and analyzing the vehicle blind spot detection characteristic data; S4: formulating a vehicle blind spot detection sensing early warning control plan, and early warning control of the safety risks of vehicle blind spot detection sensing. The present invention solves the existing problems of not being able to perform real-time detection and sensing of vehicle blind spots, not being able to provide timely early warning control, resulting in poor blind spot detection sensing effects, and not being able to effectively reduce the occurrence of safety accidents. The present invention can perform real-time detection and sensing of vehicle blind spots, and can provide timely early warning control, can improve the blind spot detection sensing effect, and can effectively reduce the occurrence of safety accidents.

Description

Technical Field

[0001] The present invention relates to the technical field of blind spot detection, and in particular to a blind spot detection method and a blind spot detection sensor. Background Art

[0002] Blind spot detection is a system or technology used to detect potentially dangerous areas within the driver's field of view. These areas are often referred to as "blind spots" because they cannot be directly seen from the driver's perspective. As society's requirements for automobile safety become increasingly higher, the active and passive safety of vehicles have received more and more attention. Among them, the demand for blind spot detection in application scenarios such as autonomous driving and automatic parking is becoming increasingly obvious.

[0003] The Chinese patent with publication number CN216285715U discloses a blind spot detection sensor with a wider detection range. Through the design of the mounting assembly, when installing the main body, a suitable mounting position can be selected according to the mounting surface required. The guide rod connected to the mounting block can rotate around the central axis, so that the mounting block can be flexibly adjusted in position, which is convenient for cross-mounting with the guide rod or adjusting the guide rod to the central axis for single-line installation when limiting the main body, which is conducive to flexible selection according to the size of the mounting surface. However, the patent has the following defects:

[0004] Existing blind spot detection sensors cannot detect and sense vehicle blind spots in real time, and cannot provide timely early warning and control, resulting in poor blind spot detection and sensing effects, and cannot effectively reduce the occurrence of safety accidents. Summary of the invention

[0005] The purpose of the present invention is to provide a blind spot detection method and a blind spot detection sensor, which can detect and sense the blind spots of vehicles in real time, provide timely early warning control, improve the blind spot detection sensing effect, effectively reduce the occurrence of safety accidents, and solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A blind spot detection method comprises the following steps:

[0008] S1: Collecting vehicle front area detection data, vehicle side area detection data and vehicle rear area detection data, determining vehicle blind spot detection real-time data, processing the vehicle blind spot detection real-time data, and determining vehicle blind spot detection feature data;

[0009] S2: According to the blind spot detection sensing requirements, a blind spot detection sensing model is constructed, and the performance of the blind spot detection sensing model is tested and optimized to determine the optimal blind spot detection sensing model;

[0010] S3: Based on the optimal blind spot detection sensing model, predictive analysis is performed on the vehicle blind spot detection feature data to determine the vehicle blind spot detection sensing prediction analysis result;

[0011] S4: Based on the prediction and analysis results of vehicle blind spot detection sensing, formulate a vehicle blind spot detection sensing early warning and control plan to carry out early warning and control of the safety risks of vehicle blind spot detection sensing.

[0012] According to another aspect of the present invention, a blind spot detection sensor is provided, which is used to implement a blind spot detection method as described above, comprising:

[0013] Data acquisition module, used to collect real-time data of vehicle blind spot detection;

[0014] A data processing module is used to process real-time data of vehicle blind spot detection;

[0015] A model building module for building a blind spot detection sensing model;

[0016] Test optimization module, used to test and optimize the blind spot detection sensing model;

[0017] Prediction analysis module, used for predicting and analyzing vehicle blind spot detection feature data;

[0018] The early warning and control module is used to carry out early warning and control of safety risks sensed by vehicle blind spot detection.

[0019] Preferably, the data acquisition module includes:

[0020] The front acquisition unit is used to detect obstacles in the area in front of the vehicle in real time and obtain detection data of the area in front of the vehicle;

[0021] The side acquisition unit is used to detect obstacles in the side area of ​​the vehicle in real time and obtain detection data of the side area of ​​the vehicle;

[0022] The rear acquisition unit is used to detect obstacles in the rear area of ​​the vehicle in real time and obtain detection data of the rear area of ​​the vehicle;

[0023] Among them, based on the vehicle front area detection data, the vehicle side area detection data and the vehicle rear area detection data, the vehicle blind spot detection real-time data is determined.

[0024] Preferably, the data acquisition module further includes:

[0025] An initial detection control module, used to control the front acquisition unit, the side acquisition unit, the side acquisition unit and the rear acquisition unit to perform obstacle detection according to the initial detection angle;

[0026] A first angle expansion module, used for performing a first detection angle expansion of the obstacle detection angle corresponding to the other acquisition units that have not acquired the obstacle when any one of the front acquisition unit, the side acquisition unit, the side acquisition unit and the rear acquisition unit acquires an obstacle, wherein the expansion range of the obstacle detection angle is 15%-20% of the initial detection angle;

[0027] A first determination module, configured to determine that the obstacle is only on one side of the acquisition unit that has acquired the obstacle, and determine that the obstacle danger level is low if the obstacle detection angle corresponding to the other acquisition units that have not acquired the obstacle is still not detected after the obstacle detection angle corresponding to the other acquisition units that have not acquired the obstacle is expanded;

[0028] The secondary expansion module is used to determine the secondary angle expansion range according to the number of newly added acquisition units that detect obstacles after the obstacle detection angle corresponding to the other acquisition units that have not acquired obstacles is expanded and an obstacle is detected; wherein the upper limit value and the lower limit value corresponding to the secondary angle expansion range are obtained by the following formula:

[0029]

[0030]

[0031] Among them, A up and A down Indicates the upper and lower limits of the secondary angle expansion range; A up0 and A down0 represents the upper limit and lower limit corresponding to the first angle expansion range; n represents the number of other acquisition units that have not acquired obstacles when any one of the front acquisition unit, the side acquisition unit, the side acquisition unit and the rear acquisition unit acquires an obstacle; A i A represents the angle expansion value of the first detection angle expansion corresponding to the i-th other acquisition unit that has not collected any obstacles; ci represents the initial detection angle corresponding to the i-th acquisition unit that has not acquired any obstacle;

[0032] The second determination module is used to determine a secondary expanded angle value that meets the secondary expanded angle range according to the secondary expanded angle range, and use the secondary expanded angle value to perform a second detection angle expansion on the detection angle of the acquisition unit that still has not detected an obstacle after the first angle expansion range, and determine the current blind spot obstacle hazard level according to the obstacle detection situation after the second detection angle expansion.

[0033] Preferably, the second determination module comprises:

[0034] An angle expansion range extraction module, used to extract the angle expansion range corresponding to the second angle expansion;

[0035] A new number extraction module is added, which is used to extract the new number of acquisition units that detect obstacles after the second detection angle expansion;

[0036] A first danger level determination module, configured to determine that the current blind spot obstacle danger level is medium when the number of newly added acquisition units that detect obstacles after the second detection angle expansion is zero;

[0037] a target acquisition unit determination module, configured to, when the number of newly added acquisition units that detect obstacles after the second detection angle expansion is non-zero, use the acquisition units that detect obstacles after the second detection angle expansion as target acquisition units;

[0038] The danger determination coefficient acquisition module is used to extract the specific angle value of the first detection angle expansion and the specific angle value of the second detection angle expansion of the target acquisition unit to obtain the danger determination coefficient; wherein the danger determination coefficient is obtained by the following formula:

[0039]

[0040] Where R represents the risk determination coefficient; R0 represents the preset initial value of the risk coefficient; m represents the number of target acquisition units; A 01j A represents the specific angle value of the first detection angle expansion of the j-th target acquisition unit; 02j A represents the specific angle value of the second detection angle expansion of the j-th target acquisition unit; cj represents the initial detection angle of the jth target acquisition unit;

[0041] A second danger level determination module is used to determine that the danger level of the current blind spot obstacle is medium when the danger determination coefficient is lower than a preset danger coefficient threshold;

[0042] The third danger level determination module is used to determine that the danger level of the current blind spot obstacle is high when the danger determination coefficient is not lower than a preset danger coefficient threshold.

[0043] Preferably, the data processing module includes:

[0044] A data retrieval unit, used to retrieve real-time data of vehicle blind spot detection;

[0045] Obtain real-time data of vehicle blind spot detection;

[0046] Based on the sequential retrieval method, the real-time data of vehicle blind spot detection is retrieved;

[0047] Check the consistency of vehicle blind spot detection real-time data;

[0048] Check whether the real-time data of vehicle blind spot detection meets the requirements according to the reasonable value range and mutual relationship of each parameter in the real-time data of vehicle blind spot detection;

[0049] Remove inconsistent data that is beyond the normal range, logically unreasonable or contradictory in the real-time data of vehicle blind spot detection;

[0050] Process invalid and missing values ​​of real-time data of vehicle blind spot detection;

[0051] Remove invalid data and missing data that are of no value to blind spot detection sensing in the real-time data of vehicle blind spot detection;

[0052] Determine the real-time vehicle blind spot detection data that is valuable for blind spot detection sensing;

[0053] A feature extraction unit, used for extracting features from the retrieved vehicle blind spot detection real-time data;

[0054] Acquire real-time vehicle blind spot detection data that is valuable for blind spot detection sensing after retrieval;

[0055] Through image analysis and transformation, the features required in the real-time data of vehicle blind spot detection, which are valuable for blind spot detection, are extracted;

[0056] Determine the vehicle blind spot detection characteristic data.

[0057] Preferably, the model building module includes:

[0058] A data collection unit, used to collect vehicle blind spot detection history data;

[0059] According to the blind spot detection sensing requirements, a large amount of vehicle blind spot detection historical data that can reflect the blind spot detection sensing characteristics is collected;

[0060] A data selection unit, used for selecting vehicle blind spot detection historical data;

[0061] Obtaining collected historical data of vehicle blind spot detection;

[0062] Select vehicle blind spot detection historical data;

[0063] Determine a blind spot detection sensing training set and a blind spot detection sensing test set;

[0064] A model building unit, used for building a blind spot detection sensing model;

[0065] According to the blind spot detection sensing requirements, select the model architecture suitable for blind spot detection sensing;

[0066] Based on the blind spot detection sensing training set, a model architecture suitable for blind spot detection sensing is trained; and a blind spot detection sensing model is determined.

[0067] Preferably, the test optimization module includes:

[0068] A performance test unit, used to perform performance testing on the blind spot detection sensing model;

[0069] Obtaining a blind spot detection sensing model;

[0070] Based on the blind spot detection sensing test set, the performance of the blind spot detection sensing model is tested;

[0071] Determine performance test results based on blind spot detection sensing models;

[0072] An optimization and adjustment unit, used for optimizing and adjusting the blind spot detection sensing model;

[0073] Obtain performance test results based on blind spot detection sensing models;

[0074] Conduct in-depth research and analysis on the performance test results based on the blind spot detection sensing model;

[0075] Determine the optimal adjustment plan based on the blind spot detection sensing model;

[0076] Optimize and adjust the blind spot detection sensing model according to the optimization adjustment plan;

[0077] Determine the optimal blind spot detection sensing model.

[0078] Preferably, the prediction analysis module includes:

[0079] A data extraction unit, used to extract vehicle blind spot detection feature data;

[0080] Extract vehicle blind spot detection feature data based on blind spot detection sensing requirements;

[0081] A prediction and analysis unit, used for predicting and analyzing vehicle blind spot detection feature data;

[0082] Obtain the optimal blind spot detection sensing model;

[0083] Based on the optimal blind spot detection sensing model, predict and analyze the vehicle blind spot detection feature data; determine the vehicle blind spot detection sensing prediction analysis results;

[0084] Among them, the vehicle blind spot detection sensing prediction analysis result is that there is a safety risk in the vehicle blind spot or there is no safety risk in the vehicle blind spot.

[0085] Preferably, the early warning control module includes:

[0086] Strategy formulation unit, used to formulate vehicle blind spot detection, sensing, warning and control plans;

[0087] Obtain vehicle blind spot detection sensing prediction analysis results;

[0088] Conduct in-depth exploration and relevant analysis of the vehicle blind spot detection sensing prediction analysis results;

[0089] Determine the vehicle blind spot detection, sensing, warning and control plan;

[0090] Early warning and control unit, used for early warning and control of safety risks sensed by vehicle blind spot detection;

[0091] Obtain vehicle blind spot detection, sensing, warning and control solutions;

[0092] Based on the vehicle blind spot detection sensing early warning control solution, early warning control of the safety risks of vehicle blind spot detection sensing is carried out;

[0093] Among them, timely early warning alarms are issued for the safety risks of vehicle blind spot detection sensing, reminding drivers in time to pay attention to the safety risks in the vehicle's blind spots.

[0094] Compared with the prior art, the present invention has the following beneficial effects:

[0095] The present invention collects vehicle front area detection data, vehicle side area detection data and vehicle rear area detection data to determine vehicle blind spot detection real-time data, processes the vehicle blind spot detection real-time data, determines vehicle blind spot detection feature data, constructs a blind spot detection sensing model according to blind spot detection sensing requirements, performs performance testing and optimization adjustment on the blind spot detection sensing model, determines an optimal blind spot detection sensing model, predicts and analyzes the vehicle blind spot detection feature data based on the optimal blind spot detection sensing model, determines a vehicle blind spot detection sensing prediction and analysis result, formulates a vehicle blind spot detection sensing early warning control plan based on the vehicle blind spot detection sensing prediction and analysis result, performs early warning control on the safety risk of vehicle blind spot detection sensing, can perform real-time detection and sensing of vehicle blind spots, can provide timely early warning control, can improve the blind spot detection sensing effect, and can effectively reduce the occurrence of safety accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 is a flow chart of the blind spot detection method of the present invention;

[0097] Figure 2 A tree diagram of the blind spot detection sensor of the present invention. DETAILED DESCRIPTION

[0098] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0099] In order to solve the problem that the existing blind spot detection sensors cannot detect the vehicle blind spots in real time and cannot provide timely warning control, resulting in poor blind spot detection and sensing effects and cannot effectively reduce the occurrence of safety accidents, please refer to Figure 1-Figure 2 , this embodiment provides the following technical solutions:

[0100] A blind spot detection sensor comprises: a data acquisition module, a data processing module, a model building module, a test optimization module, a prediction analysis module and an early warning control module.

[0101] It should be noted that through the interactive communication between the data acquisition module, data processing module, model building module, test optimization module, prediction analysis module and early warning control module, the vehicle blind spots can be detected and sensed in real time, and timely early warning control can be provided, which can improve the blind spot detection and sensing effect and effectively reduce the occurrence of safety accidents.

[0102] Among them, the data acquisition module is used to collect real-time data of vehicle blind spot detection;

[0103] In this embodiment, as a preferred technical solution of the present invention, the data acquisition module includes:

[0104] The front acquisition unit is used to detect obstacles in the area in front of the vehicle in real time and obtain detection data of the area in front of the vehicle;

[0105] It should be noted that by using lidar to detect obstacles in the area in front of the vehicle in real time and illuminating the ground with a laser beam and measuring the reflected light, high-resolution depth data can be obtained to identify blind spots in the area in front of the vehicle.

[0106] The side acquisition unit is used to detect obstacles in the side area of ​​the vehicle in real time and obtain detection data of the side area of ​​the vehicle;

[0107] It should be noted that a 360-degree panoramic camera is used to perform real-time detection of obstacles in the side area of ​​the vehicle. The 360-degree panoramic camera is installed in the side area of ​​the vehicle so that it can detect other vehicles and obstacles that may appear in the side area of ​​the vehicle during driving.

[0108] The rear acquisition unit is used to detect obstacles in the rear area of ​​the vehicle in real time and obtain detection data of the rear area of ​​the vehicle;

[0109] It should be noted that ultrasonic sensors are used to detect obstacles in the rear area of ​​the vehicle in real time. The technology of transmitting ultrasonic waves and receiving echoes can effectively detect the blind spots in the rear area of ​​the vehicle. Ultrasonic sensors are usually installed at the rear of the vehicle so that they can detect other vehicles and obstacles that may appear in the rear area of ​​the vehicle during driving.

[0110] Among them, based on the vehicle front area detection data, the vehicle side area detection data and the vehicle rear area detection data, the vehicle blind spot detection real-time data is determined.

[0111] Specifically, the data acquisition module also includes:

[0112] An initial detection control module, used to control the front acquisition unit, the side acquisition unit, the side acquisition unit and the rear acquisition unit to perform obstacle detection according to the initial detection angle;

[0113] A first angle expansion module, used for performing a first detection angle expansion of the obstacle detection angle corresponding to the other acquisition units that have not acquired the obstacle when any one of the front acquisition unit, the side acquisition unit, the side acquisition unit and the rear acquisition unit acquires an obstacle, wherein the expansion range of the obstacle detection angle is 15%-20% of the initial detection angle;

[0114] A first determination module, configured to determine that the obstacle is only on one side of the acquisition unit that has acquired the obstacle, and determine that the obstacle danger level is low if the obstacle detection angle corresponding to the other acquisition units that have not acquired the obstacle is still not detected after the obstacle detection angle corresponding to the other acquisition units that have not acquired the obstacle is expanded;

[0115] The secondary expansion module is used to determine the secondary angle expansion range according to the number of newly added acquisition units that detect obstacles after the obstacle detection angle corresponding to the other acquisition units that have not acquired obstacles is expanded and an obstacle is detected; wherein the upper limit value and the lower limit value corresponding to the secondary angle expansion range are obtained by the following formula:

[0116]

[0117]

[0118] Among them, A up and A down Indicates the upper and lower limits of the secondary angle expansion range; A up0 and A down0represents the upper limit and lower limit corresponding to the first angle expansion range; n represents the number of other acquisition units that have not acquired obstacles when any one of the front acquisition unit, the side acquisition unit, the side acquisition unit and the rear acquisition unit acquires an obstacle; A i A represents the angle expansion value of the first detection angle expansion corresponding to the i-th other acquisition unit that has not collected any obstacles; ci represents the initial detection angle corresponding to the i-th acquisition unit that has not acquired any obstacle;

[0119] The second determination module is used to determine a secondary expanded angle value that meets the secondary expanded angle range according to the secondary expanded angle range, and use the secondary expanded angle value to perform a second detection angle expansion on the detection angle of the acquisition unit that still has not detected an obstacle after the first angle expansion range, and determine the current blind spot obstacle hazard level according to the obstacle detection situation after the second detection angle expansion.

[0120] The technical effect of the above technical solution is: through the initial detection control module, the system can perform all-round obstacle detection around the vehicle (front, side, rear) according to the preset initial detection angle to ensure basic detection coverage. When any acquisition unit detects an obstacle, the first angle expansion module can intelligently expand the detection angle of other acquisition units that have not detected the obstacle, thereby increasing the breadth and depth of detection and improving the comprehensiveness of detection.

[0121] If no obstacle is detected after the first angle expansion, the first determination module can determine that the obstacle is only on the detected side and assess it as a low-level danger, which helps the system to respond quickly while avoiding unnecessary overreaction. If other acquisition units also detect obstacles after the first expansion, the secondary expansion module can dynamically calculate the secondary angle expansion range according to the number of newly added acquisition units, realizing more refined detection strategy adjustments.

[0122] Through the secondary angle expansion and the second judgment module, the system can further refine the risk level assessment of blind spot obstacles. Based on the obstacle detection after the second detection angle expansion, the system can more accurately determine the location, distance and possible threat of the obstacle, thereby providing a more accurate basis for subsequent decisions (such as obstacle avoidance, deceleration, etc.).

[0123] Overall, this technical solution significantly improves the vehicle's obstacle detection and response capabilities in complex environments through intelligent detection and evaluation mechanisms, helps reduce traffic accidents caused by blind spot obstacles, and improves driving safety. At the same time, through refined detection and evaluation, the system can provide users with more timely and accurate driving assistance information and improve user experience. The angle expansion range and hazard level assessment mechanism in this technical solution are based on formulas and algorithms, and have strong adaptability and scalability. With the continuous advancement of sensor technology and algorithms, the system can easily adjust parameters or optimize algorithms to further improve the accuracy and efficiency of detection and evaluation.

[0124] Specifically, the second determination module includes:

[0125] An angle expansion range extraction module, used to extract the angle expansion range corresponding to the second angle expansion;

[0126] A new number extraction module is added, which is used to extract the new number of acquisition units that detect obstacles after the second detection angle expansion;

[0127] A first danger level determination module, configured to determine that the current blind spot obstacle danger level is medium when the number of newly added acquisition units that detect obstacles after the second detection angle expansion is zero;

[0128] a target acquisition unit determination module, configured to, when the number of newly added acquisition units that detect obstacles after the second detection angle expansion is non-zero, use the acquisition units that detect obstacles after the second detection angle expansion as target acquisition units;

[0129] The danger determination coefficient acquisition module is used to extract the specific angle value of the first detection angle expansion and the specific angle value of the second detection angle expansion of the target acquisition unit to obtain the danger determination coefficient; wherein the danger determination coefficient is obtained by the following formula:

[0130]

[0131] Where R represents the risk determination coefficient; R0 represents the preset initial value of the risk coefficient; m represents the number of target acquisition units; A 01j A represents the specific angle value of the first detection angle expansion of the j-th target acquisition unit; 02j A represents the specific angle value of the second detection angle expansion of the j-th target acquisition unit; cj represents the initial detection angle of the jth target acquisition unit;

[0132] A second danger level determination module is used to determine that the danger level of the current blind spot obstacle is medium when the danger determination coefficient is lower than a preset danger coefficient threshold;

[0133] The third danger level determination module is used to determine that the danger level of the current blind spot obstacle is high when the danger determination coefficient is not lower than a preset danger coefficient threshold.

[0134] The technical effect of the above technical solution is: by introducing the second judgment module, the technical solution can make a more detailed and accurate assessment of the danger level of blind spot obstacles based on the results of the second detection angle expansion. This helps the vehicle system to more accurately judge the actual threat level of the obstacle, so as to take corresponding measures such as obstacle avoidance or deceleration. The angle expansion range extraction module and the newly added number extraction module can extract key information in the detection process in real time, such as the angle expansion range and the number of newly detected obstacles. This information provides an important basis for subsequent danger level judgment, allowing the system to dynamically adapt to different situations and flexibly adjust the judgment strategy.

[0135] The danger determination coefficient acquisition module calculates the danger determination coefficient by comprehensively considering the specific angle values ​​of the first and second detection angle expansions of the target acquisition unit, as well as the initial detection angle. This comprehensive evaluation method based on multiple factors can more comprehensively reflect the actual situation and potential threats of obstacles and improve the accuracy and reliability of the evaluation. The technical solution adopts a hierarchical judgment mechanism, that is, the danger level of blind spot obstacles is divided into intermediate and advanced levels according to factors such as the number of new additions and the danger determination coefficient. This hierarchical judgment mechanism helps the system to more clearly identify different levels of dangerous situations and take corresponding treatment measures. Through a more detailed obstacle danger level assessment, the vehicle system can more accurately judge the threat level of the surrounding environment and take measures such as obstacle avoidance and deceleration in a timely manner to avoid potential safety hazards. At the same time, this intelligent evaluation mechanism can also enhance the user experience and make drivers more trustful and dependent on the vehicle system. The evaluation model and algorithm in the technical solution have strong adaptability and scalability. With the continuous advancement of sensor technology, data processing technology and algorithms, the system can easily further improve the accuracy and efficiency of the evaluation by adjusting parameters and optimizing algorithms. In addition, the technical solution can also be expanded and customized according to actual needs to meet the needs of different vehicles and scenarios.

[0136] Among them, the data processing module is used to process the real-time data of vehicle blind spot detection;

[0137] In this embodiment, as a preferred technical solution of the present invention, the data processing module includes:

[0138] A data retrieval unit, used to retrieve real-time data of vehicle blind spot detection;

[0139] Obtain real-time data of vehicle blind spot detection;

[0140] Based on the sequential retrieval method, the real-time data of vehicle blind spot detection is retrieved;

[0141] Check the consistency of vehicle blind spot detection real-time data;

[0142] Check whether the real-time data of vehicle blind spot detection meets the requirements according to the reasonable value range and mutual relationship of each parameter in the real-time data of vehicle blind spot detection;

[0143] Remove inconsistent data that is beyond the normal range, logically unreasonable or contradictory in the real-time data of vehicle blind spot detection;

[0144] Process invalid and missing values ​​of real-time data of vehicle blind spot detection;

[0145] Remove invalid data and missing data that are of no value to blind spot detection sensing in the real-time data of vehicle blind spot detection;

[0146] Determine the real-time vehicle blind spot detection data that is valuable for blind spot detection sensing;

[0147] A feature extraction unit, used for extracting features from the retrieved vehicle blind spot detection real-time data;

[0148] Acquire real-time vehicle blind spot detection data that is valuable for blind spot detection sensing after retrieval;

[0149] Through image analysis and transformation, the features required in the real-time data of vehicle blind spot detection, which are valuable for blind spot detection, are extracted;

[0150] Determine the vehicle blind spot detection characteristic data.

[0151] Among them, the model building module is used to build a blind spot detection sensing model;

[0152] In this embodiment, as a preferred technical solution of the present invention, the model building module includes:

[0153] A data collection unit, used to collect vehicle blind spot detection history data;

[0154] According to the blind spot detection sensing requirements, a large amount of vehicle blind spot detection historical data that can reflect the blind spot detection sensing characteristics is collected;

[0155] A data selection unit, used for selecting vehicle blind spot detection historical data;

[0156] Obtaining collected historical data of vehicle blind spot detection;

[0157] Select vehicle blind spot detection historical data;

[0158] Determine a blind spot detection sensing training set and a blind spot detection sensing test set;

[0159] A model building unit, used for building a blind spot detection sensing model;

[0160] According to the blind spot detection sensing requirements, select the model architecture suitable for blind spot detection sensing;

[0161] Based on the blind spot detection sensing training set, a model architecture suitable for blind spot detection sensing is trained; and a blind spot detection sensing model is determined.

[0162] Among them, the test optimization module is used to test and optimize the blind spot detection sensing model;

[0163] In this embodiment, as a preferred technical solution of the present invention, the test optimization module includes: a performance testing unit, which is used to perform performance testing on the blind spot detection sensing model;

[0164] Obtaining a blind spot detection sensing model;

[0165] Based on the blind spot detection sensing test set, the performance of the blind spot detection sensing model is tested;

[0166] Determine performance test results based on blind spot detection sensing models;

[0167] An optimization and adjustment unit, used for optimizing and adjusting the blind spot detection sensing model;

[0168] Obtain performance test results based on blind spot detection sensing models;

[0169] Conduct in-depth research and analysis on the performance test results based on the blind spot detection sensing model;

[0170] Determine the optimal adjustment plan based on the blind spot detection sensing model;

[0171] Optimize and adjust the blind spot detection sensing model according to the optimization adjustment plan;

[0172] Determine the optimal blind spot detection sensing model.

[0173] Among them, the prediction and analysis module is used to perform prediction and analysis on the vehicle blind spot detection feature data;

[0174] In this embodiment, as a preferred technical solution of the present invention, the prediction and analysis module includes: a data extraction unit for extracting vehicle blind spot detection feature data;

[0175] Extract vehicle blind spot detection feature data based on blind spot detection sensing requirements;

[0176] A prediction and analysis unit, used for predicting and analyzing vehicle blind spot detection feature data;

[0177] Obtain the optimal blind spot detection sensing model;

[0178] Based on the optimal blind spot detection sensing model, predict and analyze the vehicle blind spot detection feature data;

[0179] Determine the vehicle blind spot detection sensing prediction analysis results;

[0180] Among them, the vehicle blind spot detection sensing prediction analysis result is that there is a safety risk in the vehicle blind spot or there is no safety risk in the vehicle blind spot.

[0181] Among them, the early warning and control module is used to carry out early warning and control of safety risks sensed by vehicle blind spot detection.

[0182] In this embodiment, as a preferred technical solution of the present invention, the early warning control module includes:

[0183] Strategy formulation unit, used to formulate vehicle blind spot detection, sensing, warning and control plans;

[0184] Obtain vehicle blind spot detection sensing prediction analysis results;

[0185] Conduct in-depth exploration and relevant analysis of the vehicle blind spot detection sensing prediction analysis results;

[0186] Determine the vehicle blind spot detection, sensing, warning and control plan;

[0187] Early warning and control unit, used for early warning and control of safety risks sensed by vehicle blind spot detection;

[0188] Obtain vehicle blind spot detection, sensing, warning and control solutions;

[0189] Based on the vehicle blind spot detection sensing early warning control solution, early warning control of the safety risks of vehicle blind spot detection sensing is carried out;

[0190] Among them, timely early warning alarms are issued for the safety risks of vehicle blind spot detection sensing, reminding drivers in time to pay attention to the safety risks in the vehicle's blind spots.

[0191] Therefore, by collecting vehicle front area detection data, vehicle side area detection data and vehicle rear area detection data, the vehicle blind spot detection real-time data is determined, the vehicle blind spot detection real-time data is processed, the vehicle blind spot detection feature data is determined, and a blind spot detection sensing model is constructed according to the blind spot detection sensing requirements. The blind spot detection sensing model is performance tested and optimized to determine the optimal blind spot detection sensing model. Based on the optimal blind spot detection sensing model, the vehicle blind spot detection feature data is predictively analyzed to determine the vehicle blind spot detection sensing prediction analysis results. Based on the vehicle blind spot detection sensing prediction analysis results, a vehicle blind spot detection sensing early warning control plan is formulated to carry out early warning control of the safety risks of vehicle blind spot detection sensing. The vehicle blind spots can be detected and sensed in real time, and timely early warning control can be provided, which can improve the blind spot detection sensing effect and effectively reduce the occurrence of safety accidents.

[0192] In order to better illustrate the blind spot detection process, this embodiment now provides a blind spot detection method, which is implemented based on a blind spot detection sensor as described above and includes the following steps:

[0193] S1: Collecting vehicle front area detection data, vehicle side area detection data and vehicle rear area detection data, determining vehicle blind spot detection real-time data, processing the vehicle blind spot detection real-time data, and determining vehicle blind spot detection feature data;

[0194] S2: According to the blind spot detection sensing requirements, a blind spot detection sensing model is constructed, and the performance of the blind spot detection sensing model is tested and optimized to determine the optimal blind spot detection sensing model;

[0195] S3: Based on the optimal blind spot detection sensing model, predictive analysis is performed on the vehicle blind spot detection feature data to determine the vehicle blind spot detection sensing prediction analysis result;

[0196] S4: Based on the prediction and analysis results of vehicle blind spot detection sensing, formulate a vehicle blind spot detection sensing early warning and control plan to carry out early warning and control of the safety risks of vehicle blind spot detection sensing.

[0197] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0198] Although 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 appended claims and their equivalents.

Claims

1. A blind spot detection method, based on a blind spot detection sensor, characterized in that: The steps include: S1: Collecting vehicle front area detection data, vehicle side area detection data and vehicle rear area detection data, determining vehicle blind spot detection real-time data, processing the vehicle blind spot detection real-time data, and determining vehicle blind spot detection feature data; S2: According to the blind spot detection sensing requirements, a blind spot detection sensing model is constructed, and the performance of the blind spot detection sensing model is tested and optimized to determine the optimal blind spot detection sensing model; S3: Based on the optimal blind spot detection sensing model, predictive analysis is performed on the vehicle blind spot detection feature data to determine the vehicle blind spot detection sensing prediction analysis result; S4: Based on the prediction and analysis results of vehicle blind spot detection sensing, formulate a vehicle blind spot detection sensing warning and control plan to conduct early warning and control of the safety risks of vehicle blind spot detection sensing; A blind spot detection sensor, comprising: Data acquisition module, used to collect real-time data of vehicle blind spot detection; A data processing module is used to process real-time data of vehicle blind spot detection; A model building module for building a blind spot detection sensing model; Test optimization module, used to test and optimize the blind spot detection sensing model; Prediction analysis module, used for predicting and analyzing vehicle blind spot detection feature data; Early warning and control module, used to carry out early warning and control of safety risks sensed by vehicle blind spot detection; The data acquisition module comprises: The front acquisition unit is used to detect obstacles in the area in front of the vehicle in real time and obtain detection data of the area in front of the vehicle; The side acquisition unit is used to detect obstacles in the side area of ​​the vehicle in real time and obtain detection data of the side area of ​​the vehicle; The rear acquisition unit is used to detect obstacles in the rear area of ​​the vehicle in real time and obtain detection data of the rear area of ​​the vehicle; Among them, based on the detection data of the front area of ​​the vehicle, the detection data of the side area of ​​the vehicle and the detection data of the rear area of ​​the vehicle, the real-time data of the vehicle blind spot detection is determined; The data acquisition module also includes: An initial detection control module is used to control the front acquisition unit, the side acquisition unit and the rear acquisition unit to perform obstacle detection according to the initial detection angle; A first angle expansion module is used for, when any one of the front collection unit, the side collection unit and the rear collection unit collects an obstacle, to expand the obstacle detection angle corresponding to the other collection units that have not collected the obstacle for the first time, wherein the expansion range of the obstacle detection angle is 15%-20% of the initial detection angle; A first determination module, configured to determine that the obstacle is only on one side of the acquisition unit that has acquired the obstacle, and determine that the obstacle danger level is low if the obstacle detection angle corresponding to the other acquisition units that have not acquired the obstacle is still not detected after the obstacle detection angle corresponding to the other acquisition units that have not acquired the obstacle is expanded; The secondary expansion module is used to determine the secondary angle expansion range according to the number of newly added acquisition units that detect obstacles after the obstacle detection angle corresponding to the other acquisition units that have not acquired obstacles is expanded and an obstacle is detected; wherein the upper limit value and the lower limit value corresponding to the secondary angle expansion range are obtained by the following formula: Among them, A up and A down Indicates the upper and lower limits of the secondary angle expansion range; A up0 and A down0 represents the upper limit and lower limit corresponding to the first angle expansion range; n represents the number of other acquisition units that have not acquired obstacles when any one of the front acquisition unit, the side acquisition unit and the rear acquisition unit acquires an obstacle; A i A represents the angle expansion value of the first detection angle expansion corresponding to the i-th other acquisition unit that has not collected any obstacles; ci represents the initial detection angle corresponding to the i-th acquisition unit that has not acquired any obstacle; A second determination module is used to determine a secondary expanded angle value that meets the secondary expanded angle range according to the secondary expanded angle range, and use the secondary expanded angle value to perform a second detection angle expansion on the detection angle of the acquisition unit that still has not detected an obstacle after the primary expanded angle range, and determine the current blind spot obstacle hazard level according to the obstacle detection situation after the second detection angle expansion; The second determination module comprises: An angle expansion range extraction module, used to extract the angle expansion range corresponding to the second angle expansion; A new number extraction module is added, which is used to extract the new number of acquisition units that detect obstacles after the second detection angle expansion; A first danger level determination module, configured to determine that the current blind spot obstacle danger level is medium when the number of newly added acquisition units that detect obstacles after the second detection angle expansion is zero; a target acquisition unit determination module, configured to, when the number of newly added acquisition units that detect obstacles after the second detection angle expansion is non-zero, use the acquisition units that detect obstacles after the second detection angle expansion as target acquisition units; The danger determination coefficient acquisition module is used to extract the specific angle value of the first detection angle expansion and the specific angle value of the second detection angle expansion of the target acquisition unit to obtain the danger determination coefficient; wherein the danger determination coefficient is obtained by the following formula: Where R represents the risk determination coefficient; R0 represents the preset initial value of the risk coefficient; m represents the number of target acquisition units; A 01j A represents the specific angle value of the first detection angle expansion of the j-th target acquisition unit; 02j A represents the specific angle value of the second detection angle expansion of the j-th target acquisition unit; cj represents the initial detection angle of the jth target acquisition unit; A second danger level determination module is used to determine that the danger level of the current blind spot obstacle is medium when the danger determination coefficient is lower than a preset danger coefficient threshold; The third danger level determination module is used to determine that the danger level of the current blind spot obstacle is high when the danger determination coefficient is not lower than a preset danger coefficient threshold.

2. A blind spot detection method according to claim 1, characterized in that: The data processing module comprises: A data retrieval unit, used for retrieving real-time data of vehicle blind spot detection; Obtain real-time data of vehicle blind spot detection; Based on the sequential retrieval method, the real-time data of vehicle blind spot detection is retrieved; Check the consistency of vehicle blind spot detection real-time data; Check whether the real-time data of vehicle blind spot detection meets the requirements according to the reasonable value range and mutual relationship of each parameter in the real-time data of vehicle blind spot detection; Remove inconsistent data that is beyond the normal range, logically unreasonable or contradictory in the real-time data of vehicle blind spot detection; Process invalid and missing values ​​of real-time data of vehicle blind spot detection; Remove invalid data and missing data that are of no value to blind spot detection sensing in the real-time data of vehicle blind spot detection; Determine the real-time vehicle blind spot detection data that is valuable for blind spot detection sensing; A feature extraction unit, used for extracting features from the retrieved vehicle blind spot detection real-time data; Acquire real-time vehicle blind spot detection data that is valuable for blind spot detection sensing after retrieval; Through image analysis and transformation, the features required in the real-time data of vehicle blind spot detection, which are valuable for blind spot detection, are extracted; Determine vehicle blind spot detection characteristic data.

3. A blind spot detection method according to claim 2, characterized in that: The model building module includes: A data collection unit, used to collect vehicle blind spot detection history data; According to the blind spot detection sensing requirements, a large amount of vehicle blind spot detection historical data that can reflect the blind spot detection sensing characteristics is collected; A data selection unit, used for selecting vehicle blind spot detection historical data; Obtaining collected historical data of vehicle blind spot detection; Select vehicle blind spot detection historical data; Determine a blind spot detection sensing training set and a blind spot detection sensing test set; A model building unit, used for building a blind spot detection sensing model; According to the blind spot detection sensing requirements, select the model architecture suitable for blind spot detection sensing; Based on the blind spot detection and sensing training set, the model architecture suitable for blind spot detection and sensing is trained; Determine the blind spot detection sensing model.

4. A blind spot detection method according to claim 3, characterized in that: The test optimization module includes: A performance test unit, used to perform performance testing on the blind spot detection sensing model; Obtaining a blind spot detection sensing model; Based on the blind spot detection sensing test set, the performance of the blind spot detection sensing model is tested; Determine performance test results based on blind spot detection sensing models; An optimization and adjustment unit, used for optimizing and adjusting the blind spot detection sensing model; Obtain performance test results based on blind spot detection sensing models; Conduct in-depth research and analysis on the performance test results based on the blind spot detection sensing model; Determine the optimal adjustment plan based on the blind spot detection sensing model; Optimize and adjust the blind spot detection sensing model according to the optimization adjustment plan; Determine the optimal blind spot detection sensing model.

5. A blind spot detection method according to claim 4, characterized in that: The prediction analysis module includes: A data extraction unit, used to extract vehicle blind spot detection feature data; Extract vehicle blind spot detection feature data based on blind spot detection sensing requirements; A prediction and analysis unit, used for predicting and analyzing vehicle blind spot detection feature data; Obtain the optimal blind spot detection sensing model; Based on the optimal blind spot detection sensing model, predict and analyze the vehicle blind spot detection feature data; Determine the vehicle blind spot detection sensing prediction analysis results; Among them, the vehicle blind spot detection sensing prediction analysis result is that there is a safety risk in the vehicle blind spot or there is no safety risk in the vehicle blind spot.

6. A blind spot detection method according to claim 5, characterized in that: The early warning control module includes: Strategy formulation unit, used to formulate vehicle blind spot detection, sensing, warning and control plans; Obtain vehicle blind spot detection sensing prediction analysis results; Conduct in-depth exploration and relevant analysis of the vehicle blind spot detection sensing prediction analysis results; Determine the vehicle blind spot detection, sensing, warning and control plan; Early warning and control unit, used for early warning and control of safety risks sensed by vehicle blind spot detection; Obtain vehicle blind spot detection, sensing, warning and control solutions; Based on the vehicle blind spot detection sensing early warning control solution, early warning control of the safety risks of vehicle blind spot detection sensing is carried out; Among them, timely early warning alarms are issued for the safety risks of vehicle blind spot detection sensing, reminding drivers in time to pay attention to the safety risks in the vehicle's blind spots.

Images