A method and system for identifying vehicle type using a front laser radar in a high-speed toll collection system

Abstract

The present invention provides a method and system for identifying vehicle types using a front laser radar of a high-speed toll collection system, and relates to the technical field of laser radar systems. The present invention integrates two laser radars at different angles, and combines a dynamic adjustment strategy for an environmental impact coefficient and a cleanliness impact coefficient, and finally adjusts the scanning parameters of the laser radar by adjusting an index, and finally determines the adjusted angle of an angle θ by generating a first comparison result and a second comparison result, and at the same time, grades the adjustment amount of the angle θ, and performs fine-tuning and calibration of corresponding values ​​on a first threshold and a second threshold for next vehicle type identification, so that the two laser radars at different angles can dynamically adjust the angle θ according to actual environmental data and pollutant characteristic data, thereby improving the accuracy of the output result of the vehicle type identification method.

Description

Technical Field

[0001] The present invention relates to the technical field of laser radar systems, and in particular to a method and system for identifying vehicle types using a front laser radar in a high-speed toll collection system. Background Art

[0002] The automation and intelligence of highway toll collection systems have become one of the key technologies to improve traffic efficiency and reduce operating costs. Among many technologies, vehicle type recognition technology has attracted widespread attention due to its importance to toll collection accuracy and system efficiency. Traditional vehicle type recognition technology mainly relies on image processing and radar technology, and uses pre-built neural network models to identify vehicle types. However, the stability and accuracy of these technologies in complex environments are still facing challenges, especially in high-speed driving environments, where factors such as vehicle posture changes, weather conditions, and road pollutants have a significant impact on recognition accuracy.

[0003] In the prior art, the publication number is CN112365718B, and the name is a laser vehicle model recognition method and device, which uses the ranging data of two vertically installed laser ranging radars to obtain the point cloud data of the vehicle in three-dimensional space through splicing, combination and conversion. According to the change amount of the point cloud data in the time domain, the projection on the three-dimensional space surface, and the coordinate processing in the three-dimensional space, the vehicle model is obtained through the vehicle model discrimination algorithm, which realizes the intuitive determination of the vehicle model from the laser radar data, and significantly improves the accuracy of vehicle model recognition;

[0004] Nevertheless, the technology still faces challenges in practical applications, especially in terms of the impact of environmental factors (temperature, humidity, visibility, atmospheric optical thickness) and the cleanliness of the lidar itself on data quality. Existing technologies fail to provide effective dynamic adjustment and optimization mechanisms, which to some extent affects the stability and robustness of the vehicle identification system.

[0005] The above information disclosed in the above Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for identifying vehicle types using a front laser radar in a high-speed toll collection system, so as to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for identifying vehicle type using a front laser radar in a high-speed toll collection system, the specific steps comprising:

[0009] Step S1: Determine that the angle between the two laser radars in the laser radar vehicle type recognition module is θ, and mark the vertically downward laser radar as i, and mark the obliquely downward laser radar as j;

[0010] Step S2: collect environmental data around the two laser radars in real time, and monitor pollutant characteristic data on the detection end surfaces of laser radar i and laser radar j respectively, transmit the environmental data and pollutant characteristic data to the central processing unit through a wireless network, and perform data preprocessing;

[0011] Step S3: Obtain the pre-processed environmental data for analysis and generate an environmental impact coefficient, which is used to evaluate the impact of the environmental data on the accuracy of the laser radar;

[0012] Step S4: obtaining the pre-processed pollutant characteristic data on the monitoring laser radar i, and analyzing and processing it to generate a first cleanliness influence coefficient; obtaining the pre-processed pollutant characteristic data on the laser radar j, and analyzing and processing it to generate a second cleanliness influence coefficient;

[0013] Step S5: Obtaining the environmental impact coefficient and the first cleanliness impact coefficient, analyzing and processing them, and constructing a first adjustment index, where the first adjustment index is used to generate a first adjustment strategy for dynamically adjusting the scanning parameters of the laser radar i;

[0014] Step S6: Obtain the environmental impact coefficient and the second cleanliness impact coefficient, analyze and process them, and construct a second adjustment index, where the second adjustment index is used to generate a second adjustment strategy for dynamically adjusting the scanning parameters of the laser radar j;

[0015] Step S7: setting adjustment thresholds of the first adjustment strategy and the second adjustment strategy, which are the first threshold and the second threshold respectively, and comparing and analyzing the first adjustment strategy with the first threshold to generate a first comparison result; and comparing and analyzing the second adjustment strategy with the second threshold to generate a second comparison result;

[0016] Step S8: after obtaining the first comparison result and the second comparison result, a comprehensive analysis is performed, and finally the angle after the included angle θ is adjusted is determined;

[0017] Step S9: the adjustment amount of the angle θ is graded, the larger the adjustment amount, the higher the grade, and the first threshold and the second threshold are fine-tuned and calibrated to the corresponding values ​​for the next vehicle model recognition according to different adjustment amount grades;

[0018] Step S10: According to the adjusted angle and scanning parameters, the vehicle is scanned by laser radar i and laser radar j to obtain vehicle information, and the vehicle model recognition is completed through the pre-trained neural network algorithm, the original data is matched and recognized, and output content is generated. The output content includes vehicle model, vehicle height, vehicle length, and vehicle speed data.

[0019] Furthermore, after obtaining the first comparison result and the second comparison result, a comprehensive analysis is performed, and finally the angle after the angle θ is adjusted is determined, which specifically includes the following contents:

[0020] Define the angle θ after comprehensive adjustment: , The calculation formula is:

[0021] ;

[0022] in, is the adjusted angle calculated based on the first comparison result, is the adjusted angle calculated based on the second comparison result, is the adjustment factor used to Adjust the positive and negative of

[0023] set up The value of is chosen as ;

[0024] when and When the angle is reduced: The value range is ;

[0025] when and When the angle is increased, The value range is ;

[0026] when and When any one of them increases the angle, the other decreases the angle;

[0027] When the weight of lidar i is greater than that of lidar j, ;

[0028] When the weight of laser radar i is less than that of laser radar j, ;

[0029] When the weight of lidar i is equal to that of lidar j, .

[0030] Furthermore, the adjustment amount of the angle θ is divided into the following levels, and the current adjustment amount of the angle θ is set to ;

[0031] Low adjustment means that the angle adjustment is less than 5%;

[0032] Medium adjustment means that the angle adjustment is between 5%-20%;

[0033] A high adjustment means that the angle adjustment is greater than 20%;

[0034] 1) When the adjustment level is low, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle model identification. The fine-tuning formula is:

[0035] ;

[0036] in, is the adjusted first threshold; is the adjusted second threshold; is the current adjustment amount;

[0037] 2) When the adjustment level is medium adjustment, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle model identification. The fine-tuning formula is:

[0038] ;

[0039] 3) When the adjustment level is a high adjustment, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle type recognition. The fine-tuning formula is:

[0040] .

[0041] A front laser radar vehicle type recognition system for a high-speed toll collection system, the system is used to execute the front laser radar vehicle type recognition method for the high-speed toll collection system, comprising:

[0042] Angle marking module: used to determine the angle between the two laser radars in the laser radar vehicle identification module as θ, and mark the vertically downward laser radar as i, and the tilted downward laser radar as j;

[0043] Data processing module: used to collect the environmental data around the two laser radars in real time, and monitor the pollutant characteristic data on the detection end surface of laser radar i and laser radar j respectively, transmit the environmental data and pollutant characteristic data to the central processing unit through the wireless network, and perform data preprocessing;

[0044] Environmental impact coefficient generation module: used to obtain pre-processed environmental data for analysis and generate environmental impact coefficients, which are used to evaluate the impact of environmental data on the accuracy of the lidar;

[0045] Cleanliness influence coefficient generation module: used to obtain the pre-processed pollutant characteristic data on the monitoring laser radar i, and analyze and process it to generate a first cleanliness influence coefficient; obtain the pre-processed pollutant characteristic data on the laser radar j, and analyze and process it to generate a second cleanliness influence coefficient;

[0046] A first adjustment index generation module: used to obtain the environmental impact coefficient and the first cleanliness impact coefficient, and perform analysis and processing to construct a first adjustment index, which is used to generate a first adjustment strategy for dynamically adjusting the scanning parameters of the laser radar i;

[0047] A second adjustment index generation module: used to obtain the environmental impact coefficient and the second cleanliness impact coefficient, and perform analysis and processing to construct a second adjustment index, which is used to generate a second adjustment strategy for dynamically adjusting the scanning parameters of the laser radar j;

[0048] Comparison module: used for setting adjustment thresholds of the first adjustment strategy and the second adjustment strategy, which are the first threshold and the second threshold respectively, and comparing and analyzing the first adjustment strategy with the first threshold to generate a first comparison result; and comparing and analyzing the second adjustment strategy with the second threshold to generate a second comparison result;

[0049] Angle adjustment module: used to obtain the first comparison result and the second comparison result, perform comprehensive analysis, and finally determine the angle after the angle θ is adjusted;

[0050] Fine-tuning calibration module: used to classify the adjustment amount of the angle θ into grades. The larger the adjustment amount, the higher the grade. According to different adjustment amount grades, the first threshold and the second threshold for the next vehicle model recognition are fine-tuned and calibrated to the corresponding values.

[0051] Vehicle type recognition module: It is used to scan the vehicle through LiDAR i and LiDAR j according to the adjusted angle and scanning parameters to obtain vehicle information, and complete the vehicle type recognition work through the pre-trained neural network algorithm, match and identify the original data, and generate output content. The output content includes vehicle type, vehicle height, vehicle length, and vehicle speed data.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] By integrating two laser radars at different angles and combining the dynamic adjustment strategy of the environmental impact coefficient and the cleanliness impact coefficient, the scanning parameters of the laser radar are finally adjusted by adjusting the index, and the adjusted angle θ is finally determined by generating the first comparison result and the second comparison result. At the same time, the adjustment amount of the angle θ is graded, and the first threshold and the second threshold are fine-tuned and calibrated to the corresponding values ​​for the next vehicle model recognition. The two laser radars at different angles can dynamically adjust the angle θ according to the actual environmental data and pollutant characteristic data, thereby improving the accuracy of the output results of the vehicle model recognition method. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0055] Figure 2 It is a system module block diagram of the present invention. DETAILED DESCRIPTION

[0056] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0057] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0058] Embodiment 1:

[0059] See also Figure 1 , the present invention provides a technical solution:

[0060] A method for identifying vehicle types using a front laser radar of a high-speed toll collection system, the method being applied to a laser radar vehicle type identification module, comprising the following steps:

[0061] Step S1: Determine that the angle between the two laser radars in the laser radar vehicle type recognition module is θ, and mark the vertically downward laser radar as i, and mark the obliquely downward laser radar as j;

[0062] Step S2: collect environmental data around the two laser radars in real time, and monitor pollutant characteristic data on the detection end surfaces of laser radar i and laser radar j respectively, transmit the environmental data and pollutant characteristic data to the central processing unit through a wireless network, and perform data preprocessing;

[0063] Step S3: Obtain the pre-processed environmental data for analysis and generate an environmental impact coefficient, which is used to evaluate the impact of the environmental data on the accuracy of the laser radar;

[0064] Step S4: obtaining the pre-processed pollutant characteristic data on the monitoring laser radar i, and analyzing and processing it to generate a first cleanliness influence coefficient; obtaining the pre-processed pollutant characteristic data on the laser radar j, and analyzing and processing it to generate a second cleanliness influence coefficient;

[0065] Step S5: Obtaining the environmental impact coefficient and the first cleanliness impact coefficient, analyzing and processing them, and constructing a first adjustment index, where the first adjustment index is used to generate a first adjustment strategy for dynamically adjusting the scanning parameters of the laser radar i;

[0066] Step S6: Obtain the environmental impact coefficient and the second cleanliness impact coefficient, analyze and process them, and construct a second adjustment index, where the second adjustment index is used to generate a second adjustment strategy for dynamically adjusting the scanning parameters of the laser radar j;

[0067] Step S7: setting adjustment thresholds of the first adjustment strategy and the second adjustment strategy, which are the first threshold and the second threshold respectively, and comparing and analyzing the first adjustment strategy with the first threshold to generate a first comparison result; and comparing and analyzing the second adjustment strategy with the second threshold to generate a second comparison result;

[0068] Step S8: after obtaining the first comparison result and the second comparison result, a comprehensive analysis is performed, and finally the angle after the included angle θ is adjusted is determined;

[0069] Step S9: the adjustment amount of the angle θ is graded, the larger the adjustment amount, the higher the grade, and the first threshold and the second threshold are fine-tuned and calibrated to the corresponding values ​​for the next vehicle model recognition according to different adjustment amount grades;

[0070] Step S10: According to the adjusted angle and scanning parameters, the vehicle is scanned by laser radar i and laser radar j to obtain vehicle information, and the vehicle model recognition is completed through the pre-trained neural network algorithm, the original data is matched and recognized, and output content is generated. The output content includes vehicle model, vehicle height, vehicle length, and vehicle speed data.

[0071] To further illustrate, the angle between the two laser radars in the laser radar vehicle identification module is determined to be θ, and the vertically downward laser radar is marked as i, and the tilted downward laser radar is marked as j, which specifically includes the following:

[0072] 1) Install the LiDAR equipment:

[0073] Install a LiDAR vehicle type recognition module in the front area of ​​the designed high-speed toll collection system. The module contains two LiDAR devices.

[0074] One of the laser radars is installed vertically downward, marked as laser radar i; the other laser radar is installed obliquely downward, forming an angle θ with laser radar i, marked as laser radar j;

[0075] 2) Measure the angle θ:

[0076] Use a high-precision angle measuring instrument, which in this embodiment is a digital goniometer or a laser rangefinder, to measure the angle θ between the two laser radars; the angle θ meets the design requirements to provide the best vehicle model recognition effect;

[0077] 3) Adjust the installation angle of the LiDAR:

[0078] According to the measured angle θ, fine-tuning is performed to ensure that the angle between the two laser radars accurately meets the preset design value; the preset design value is determined according to the setting standard of the manufacturer of the laser radar vehicle identification module, and the installation angle of the laser radar is fixed using an adjustment bracket and a locking device to avoid angle deviation due to vibration or other factors;

[0079] 4) Label lidar i and lidar j:

[0080] The laser radar installed vertically downward is marked as laser radar i, and the installation position parameters are recorded. In this embodiment, the installation position parameters are height and distance;

[0081] The laser radar installed tilted downward is marked as laser radar j, and the angle θ between it and laser radar i and the installation position parameters are recorded;

[0082] Environmental data include temperature, humidity, visibility, and atmospheric optical thickness, which are marked as T, H, V, and AOt respectively. The specific acquisition steps are as follows:

[0083] Environmental sensors are installed around the LiDAR vehicle identification module to monitor environmental data in real time. These sensors include temperature sensors, humidity sensors, visibility sensors, and atmospheric optical thickness sensors.

[0084] The pollutant characteristic data include particle deposition, transmittance, and adhesion coefficient. The particle deposition, transmittance, and adhesion coefficient of the monitoring lidar i are marked as , and ;

[0085] The particle deposition amount, transmittance, and adhesion coefficient of the monitoring laser radar j are marked as , and ;

[0086] For particle deposition, it refers to the number of particles deposited per unit area, expressed as the number of particles / cm², measured by a particle counter; for particle deposition, it refers to the number of particles deposited per unit area, expressed as the number of particles / cm², measured by a particle counter; and , the mathematical expression is as follows:

[0087]

[0088]

[0089] in, is the number of particles deposited on lidar i per unit time; is the number of particles deposited on the laser radar j per unit time; A is the surface area of ​​the laser radar detection end. In this embodiment, the surface areas of the detection ends of the laser radar i and the laser radar j are both limited to A;

[0090] Adjustment suggestions: According to the severity of the pollutants, collect and record the amount of particulate matter deposition in real time. When there are a lot of pollutants, reduce the angle to enhance data concentration;

[0091] For transmittance, the transmittance of laser through the detection end surface is expressed as a percentage and monitored by a transmission photometer; for transmittance and , the mathematical expression is as follows:

[0092]

[0093]

[0094] in,

[0095] is the intensity of the transmitted light through the detection end of the laser radar i;

[0096] is the intensity of the incident light incident on the detection end of the laser radar i;

[0097] is the intensity of the transmitted light through the detection end of the laser radar j;

[0098] is the intensity of the incident light incident on the detection end of the laser radar j;

[0099] Adjustment suggestions: Real-time monitoring of transmittance changes. When there are many pollutants, reducing the angle can improve data quality;

[0100] Adhesion coefficient reflects the adhesion ability of pollutants. It is calculated by dividing the adhesion force measured experimentally by the surface area of ​​the detection end. The unit is N / m². and , the mathematical expression is as follows:

[0101]

[0102]

[0103] in, is the adhesion force of the pollutant on the surface of the lidar i detection end; is the adhesion force of the pollutant on the surface of the laser radar detection end; A is the surface area of ​​the laser radar detection end;

[0104] Adjustment suggestions: When there are more pollutants, the adhesion coefficient increases and the angle is reduced to enhance data accuracy;

[0105] The environmental data and pollutant characteristic data are transmitted to the central processing unit via the wireless network and the data is pre-processed, including:

[0106] Step 3.1: Pack the collected environmental data and pollutant characteristic data into a data package:

[0107] Step 3.2: Install a wireless transmission module near the laser radar vehicle identification module. In this embodiment, it is a Wi-Fi or 4G / 5G module, which is used to transmit the packaged data to the central processing unit through the wireless network;

[0108] Step 3.3, transmitting the data packet to the central processing unit via the wireless network;

[0109] The central processing unit receives and pre-processes environmental data and pollutant characteristic data;

[0110] Step 4.1, the central processing unit receives environmental data and pollutant characteristic data packets;

[0111] Step 4.2, parse the received data according to the data packet format and store it in the corresponding variable;

[0112] Step 4.3: Preprocess the parsed data, including but not limited to: Data cleaning: removing noise and outliers;

[0113] Data correction: Correct the data according to known calibration parameters;

[0114] Data format conversion: converting data into the required format for subsequent processing and analysis;

[0115] Ensure that the parameters of environmental data and pollutant characteristic data are in the same range (0,1) for comprehensive evaluation, and use the minimum-maximum normalization method for each parameter to perform normalization; the normalization formula is as follows:

[0116]

[0117] Where X is the original data value; The minimum value in the data set; is the maximum value in the data set; is the normalized data value.

[0118] To further explain, the environmental impact coefficient is defined as EIC, and the calculation formula is as follows:

[0119]

[0120] in, is the adjustment coefficient, controlling the amplitude of the overall influence coefficient. is the adjustment parameter of the temperature influence coefficient, is the adjustment parameter of humidity influence coefficient, The value range is (0,1). The specific value of is determined by the expert group through experimental data, and the determination method is the existing technology and will not be described in detail;

[0121] Reflects the effect of temperature on the laser radar recognition accuracy. The higher or lower the temperature, the greater the impact on the recognition accuracy.

[0122] It represents the influence of humidity on the recognition accuracy of LiDAR. The higher the humidity, the greater the influence on the recognition accuracy. And H is always greater than ;

[0123] Indicates the impact of visibility on the laser radar recognition accuracy. The lower the visibility, the greater the impact on the recognition accuracy.

[0124] Represents the effect of atmospheric optical thickness on the recognition accuracy of lidar. The greater the atmospheric optical thickness, the greater its effect on the recognition accuracy, where AOt is always less than 1;

[0125] when When , it means the environmental impact is greater. The value needs to be ensured ;

[0126] And when hour, The value is also close to 1;

[0127] Set the value range of EIC to (0,1);

[0128] When EIC is closer to 0, it means that the environment has a greater impact on the recognition accuracy, and the laser radar recognition accuracy is lower;

[0129] When EIC is closer to 1, it means that the environment has less impact on the recognition accuracy and the lidar recognition accuracy is higher.

[0130] Further explanation: the first cleanliness influence coefficient is defined as , the calculation formula is as follows:

[0131]

[0132] The second cleanliness influence coefficient is , the calculation formula is as follows:

[0133]

[0134] Among them: a1 and b1 are adjustment coefficients, which control the amplitude of the overall influence coefficient, a2 and b2 are adjustment parameters of the particle deposition influence coefficient, a3 and b3 are adjustment parameters of the transmittance influence coefficient, the value range of a1, a2, a3, b1, b2, b3 is (0, 1), the specific values ​​of a1, a2, a3, b1, b2, b3 are determined by the expert group through experimental data, and the determination method is the existing technology, which will not be repeated;

[0135] Both represent the effect of particle deposition on the cleanliness of the lidar. The greater the deposition, the lower the cleanliness.

[0136] Both represent the effect of transmittance on the cleanliness of the LiDAR. The higher the transmittance, the higher the cleanliness. ;

[0137] Both represent the effect of adhesion coefficient on the cleanliness of the lidar. The stronger the adhesion, the lower the cleanliness.

[0138] set up and The value range of is (0,1), which indicates the degree of influence of cleanliness;

[0139] when or The closer it is to 0, the greater the influence of pollutant characteristic data on the cleanliness of the lidar, and the lower the cleanliness of the lidar;

[0140] when or The closer it is to 1, the smaller the impact of pollutant characteristic data on the cleanliness of the lidar is, and the higher the cleanliness of the lidar is.

[0141] Further explanation: the first adjustment index corresponding to the laser radar i is defined as , the calculation formula is as follows:

[0142]

[0143] in, and They are EIC and The weight of ; EIC stands for Environmental Impact Coefficient; represents the first cleanliness influence coefficient; represents a first adjustment index, used to generate a first adjustment strategy for dynamically adjusting the scanning parameters of the laser radar i;

[0144] When EIC decreases, When increasing, and The weight distribution of EIC is adjusted to reflect the corresponding changes in EIC. The impact ratio The corresponding changes are more important, setting > to ensure that changes in EIC The impact is more prominent;

[0145] When EIC increases, When it decreases, and Adjust the weight distribution and set > to ensure The changes in The impact is more prominent;

[0146] set up The value range of is (0,1);

[0147] when The closer it is to 0, the greater the environmental impact and cleanliness impact, and the greater the adjustment of the scanning parameters of LiDAR i;

[0148] when The closer it is to 1, the smaller the environmental impact and cleanliness impact are, and the smaller the need to adjust the scanning parameters of LiDAR i is;

[0149] The first adjustment strategy includes:

[0150] The scanning parameters include scanning frequency, scanning angle and transmission power. The comparison threshold of the first adjustment index is set to Q1, the value range of Q1 is 0.35≤Q1≤0.68, and the initial reference value of Q1 is set to 0.46; Q1 is determined by the expert group through experimental data, and the determination method is the existing technology and will not be described in detail;

[0151] 1) When the first adjustment index is greater than the threshold Q1, the scanning frequency is increased by 20% to improve the recognition accuracy; the scanning angle is increased by 15% to cover a larger area; the transmission power is increased by 25% to improve the signal strength;

[0152] 2) If EIC < 0.3, based on step 1), the scanning frequency is increased by 5% and the transmission power is increased by 10%;

[0153] like , add 5% to the scanning angle in step 1);

[0154] When the first adjustment index is equal to the threshold Q1, the scanning frequency, scanning angle and transmission power remain unchanged;

[0155] 3) When the first adjustment index is less than the threshold Q1, the scanning frequency is reduced by 15%; the scanning angle is reduced by 10%; and the transmission power is reduced by 20%;

[0156] 4) If EIC>0.7, based on step 3), the scanning frequency is reduced by 5% and the transmission power is reduced by 5%;

[0157] like , reduce the scanning angle in step 3) by 5%.

[0158] Further explanation: the second adjustment index corresponding to the laser radar j is defined as , the calculation formula is as follows:

[0159]

[0160] in, and They are EIC and The weight of ;

[0161] When EIC decreases, When increasing, and The weight distribution of EIC is adjusted to reflect the corresponding changes in EIC. The impact ratio The corresponding changes are more important, setting > to ensure that changes in EIC The impact is more prominent;

[0162] When EIC increases, When it decreases, and Adjust the weight distribution and set > to ensure The changes in The impact is more prominent;

[0163] set up The value range of is (0,1);

[0164] when The closer it is to 0, the greater the impact of the environment and cleanliness, and the greater the adjustment of the scanning parameters of the laser radar j;

[0165] when The closer it is to 1, the smaller the impact of the environment and cleanliness, and the smaller the need to adjust the scanning parameters of the laser radar j;

[0166] The second adjustment strategy includes the following:

[0167] The comparison threshold of the second adjustment index is set to Q2, the value range of Q2 is 0.38≤Q2≤0.75, and the initial reference value of Q2 is set to 0.52; Q2 is determined by the expert group through experimental data, and the determination method is the existing technology and will not be described in detail;

[0168] 1.1) When the second adjustment index is greater than the threshold Q2, the scanning frequency is reduced by 10%, the scanning angle is reduced by 5%, and the transmission power is reduced by 15%;

[0169] 1.2) When EIC>0.8, based on step 1.1), the scanning frequency is reduced by 5% and the transmission power is reduced by 6%;

[0170] when , the scanning angle is reduced by 4% based on step 1.1);

[0171] 2.1) When the second adjustment index is equal to the threshold Q2, the scanning frequency, scanning angle and transmission power remain unchanged;

[0172] 3.1) When the second adjustment index is less than the threshold Q2, the scanning frequency increases by 20% and the scanning angle increases by 15%.

[0173] Transmit power increased by 25%;

[0174] 3.2) When , based on step 3.1), the scanning frequency increases by 5% and the transmission power increases by 6%;

[0175] when , the scanning angle is increased by 6% based on that in step 3.1).

[0176] Further, the first adjustment strategy is compared and analyzed with the first threshold to generate a first comparison result, including the following contents:

[0177] Set the first threshold of the comparison threshold Q1 to r1;

[0178] 1) When the comparison threshold Q1 is greater than the first threshold r1, the fine-tuning formula for Q1 is as follows:

[0179]

[0180] in, ; is the adjusted threshold;

[0181] The angle between LiDAR i and LiDAR j The adjustment formula is:

[0182]

[0183] in, , is the angle adjustment amount; is the adjusted angle;

[0184] 2) When the comparison threshold Q1 is equal to the fine-tuning threshold r1, Q1 remains unchanged;

[0185] The angle between LiDAR i and LiDAR j remain unchanged;

[0186] 3) When the comparison threshold Q1 is less than the first threshold r1, the fine-tuning formula for Q1 is as follows:

[0187]

[0188] in, ; is the adjusted threshold;

[0189] The angle between LiDAR i and LiDAR j The adjustment formula is:

[0190]

[0191] in, , is the adjusted angle.

[0192] Further, the second adjustment strategy is compared and analyzed with the second threshold to generate a second comparison result, including the following contents:

[0193] Set the second threshold of the comparison threshold Q2 to r2;

[0194] 1) When the comparison threshold Q2 is greater than the second threshold r2, the fine-tuning formula for Q2 is as follows:

[0195]

[0196] in, , is the adjusted threshold;

[0197] The angle adjustment formula between laser radar i and laser radar j is:

[0198]

[0199] in, ; is the adjusted angle;

[0200] 2) When the comparison threshold Q2 is equal to the second threshold r2, Q2 remains unchanged; the angle between the laser radar i and the laser radar j remains unchanged;

[0201] 3) When the comparison threshold Q2 is less than the second threshold r2, the fine-tuning formula for Q2 is as follows:

[0202]

[0203] in, , is the adjusted threshold;

[0204] The formula for adjusting the angle between LiDAR i and LiDAR j is:

[0205]

[0206] in, ; is the adjusted angle;

[0207] Further description, according to the first comparison result and the second comparison result, the angle θ between the laser radar i and the laser radar j is dynamically adjusted, specifically including the following contents:

[0208] Define the angle θ after comprehensive adjustment: , The calculation formula is:

[0209]

[0210] in, is the adjusted angle calculated based on the first comparison result, is the adjusted angle calculated based on the second comparison result, is the adjustment factor used to Adjust the positive and negative of

[0211] set up The value of is chosen as ;

[0212] when and When the angle is reduced: The value range is ;

[0213] Both adjustments are reduced, indicating that the angle needs to be further reduced to improve scanning accuracy;

[0214] when and When the angle is increased, The value range is ; Both adjustment amounts are increased, and the angle needs to be further increased to expand the scanning range;

[0215] when and When any one of them increases the angle, the other decreases the angle;

[0216] When the weight of lidar i is greater than that of lidar j, ;

[0217] When the weight of laser radar i is less than that of laser radar j, ;

[0218] When LiDAR i is equal to the weight of LiDAR j, .

[0219] Embodiment 2:

[0220] Further explanation based on the first embodiment, in order to verify the angle between the laser radar i and the laser radar j The effect of dynamic adjustment, the test uses two laser radars i and laser radar j, respectively configures the first adjustment strategy and the second adjustment strategy, and sets the corresponding thresholds; in the test, by comparing the comparison results of the adjustment strategy and the threshold, the angle θ is dynamically adjusted; the specific steps are as follows:

[0221] 1) Experimental preparation:

[0222] The two laser radars i and j are installed on the mobile platform respectively, ensuring that the initial angle between them is θ;

[0223] Setting the comparison threshold Q1 of the first adjustment strategy and the comparison threshold Q2 of the second adjustment strategy to different initial values, and setting corresponding first thresholds r1 and r2;

[0224] The initial angle θ is 20 degrees;

[0225] The following table records the detailed data under different test conditions:

[0226] Table 1

[0227] Adjustment strategy Initial threshold Q1 / Q2 The first threshold r1 / r2 <![CDATA[Adjusted threshold Q1 new / Q2 new > Initial angle θ <![CDATA[Adjusted included angle θ 1 / θ 2 > <![CDATA[The included angle θ after dynamic adjustment new > LiDAR Q1=50 r1=45 <![CDATA[Q1 new =50.25]]> θ=20∘ <![CDATA[θ 1 =30∘]]> <![CDATA[θ new =20∘]]> LiDAR Q2=60 r2=55 <![CDATA[Q2 new =60.25]]> θ=20∘ <![CDATA[θ 2 =30∘]]> <![CDATA[θ new =20∘]]> LiDAR Q1=55 r1=50 <![CDATA[Q1 new =55.25]]> θ=19∘ <![CDATA[θ 1 =29∘]]> <![CDATA[θ new =19.5∘]]> LiDAR Q2=65 r2=60 <![CDATA[Q2 new =65.25]]> θ=19∘ <![CDATA[θ 2 =29∘]]> <![CDATA[θ new =19.5∘]]> LiDAR Q1=45 r1=45 <![CDATA[Q1 new =45]]> θ=18∘ <![CDATA[θ 1 =18∘]]> <![CDATA[θ new =18∘]]> LiDAR Q2=55 r2=55 <![CDATA[Q2 new =55]]> θ=18∘ <![CDATA[θ 2 =18∘]]> <![CDATA[θ new =18∘]]> LiDAR Q1=40 r1=45 <![CDATA[Q1 new =39.75]]> θ=17∘ <![CDATA[θ 1 =15∘]]> <![CDATA[θ new =16∘]]> LiDAR Q2=50 r2=55 <![CDATA[Q2 new =49.75]]> θ=17∘ <![CDATA[θ 2 =15∘]]> <![CDATA[θ new =16∘]]>

[0228] By comparing the data in the table, we can clearly see the beneficial effects of the dynamic adjustment strategy proposed in the present invention in practical applications. Specifically, its creativity and novelty are demonstrated in the following aspects:

[0229] 1) Dynamically adjust the angle:

[0230] The table shows the angle adjustment under different initial thresholds and fine-tuning thresholds. In the first row of data, the initial threshold Q1 of LiDAR i is 50, the first threshold r1 is 45, and the adjusted threshold is 50.25, resulting in an increase in the angle from 20 degrees to 30 degrees; this indicates that when the comparison threshold is greater than the fine-tuning threshold, the angle will increase accordingly, thereby expanding the scanning range;

[0231] 2) Keep the angle stable:

[0232] In the third row of data, the initial threshold Q1 and the first threshold r1 of LiDAR i are both 45. The adjusted threshold The angle remains unchanged and 18 degrees; this indicates that when the comparison threshold is equal to the fine-tuning threshold, the angle remains unchanged, ensuring the stability of the scanning accuracy;

[0233] 3) Comprehensive adjustment effect:

[0234] Through the comprehensive adjustment formula, the last column in the table shows the dynamically adjusted angle ; In the first row of data, the angle after comprehensive adjustment is still 20 degrees, which shows that under certain conditions, the dynamic adjustment strategy can maintain the stability of the angle;

[0235] The impact of threshold adjustment on the angle:

[0236] When the comparison threshold Q1 or Q2 is greater than the first threshold r1 or r2, the adjusted threshold or will increase, resulting in an increase in the angle; in the first row of data, Q1 increases from 50 to 50.25, and the angle increases from 20 degrees to 30 degrees;

[0237] When the comparison threshold Q1 or Q2 is less than the first threshold r1 or r2, the adjusted threshold or will decrease, resulting in a decrease in the included angle; in the seventh row of data, Q1 decreases from 40 to 39.75, and the included angle decreases from 17 degrees to 15 degrees;

[0238] Numerical interaction of angle adjustment:

[0239] Angle adjustment It is proportional to the difference between the comparison threshold and the fine-tuning threshold; in the first row of data, , resulting in an increase of 10 degrees in the angle;

[0240] The angle after comprehensive adjustment The comprehensive adjustment formula is used to calculate the and The difference ensures the stability and accuracy of the angle;

[0241] Through these quantitative data, we can clearly see the dynamic adjustment effect of the present invention under different conditions, as well as its advantages in maintaining angle stability and accuracy.

[0242] Embodiment three:

[0243] Further explanation is given on the basis of the first embodiment, the adjustment amount of the angle θ is divided into the following levels, and the current adjustment amount is set as ;

[0244] A low adjustment means an adjustment of less than 5%;

[0245] Medium adjustment means the adjustment is between 5%-20%;

[0246] A high adjustment means an adjustment greater than 20%;

[0247] 1) When the adjustment level is low, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle model identification. The fine-tuning formula is:

[0248]

[0249] in, is the adjusted first threshold; is the adjusted second threshold; is the current adjustment amount;

[0250] 2) When the adjustment level is medium adjustment, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle model identification. The fine-tuning formula is:

[0251]

[0252] 3) When the adjustment level is a high adjustment, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle type recognition. The fine-tuning formula is:

[0253]

[0254] when When increasing:

[0255] The corresponding fine-tuning calibration formula increases when When increasing from 2 degrees to 3 degrees, the calibration results corresponding to the low, medium, and high adjustment amounts will become:

[0256] Low Adjustment:

[0257]

[0258] Medium adjustment amount:

[0259]

[0260] High Adjustment:

[0261]

[0262] when When reducing:

[0263] The corresponding fine-tuning calibration formula decreases; when When decreasing from 2 degrees to 1 degree, the calibration results corresponding to the low, medium, and high adjustments will become:

[0264] Low Adjustment:

[0265]

[0266] Medium adjustment amount:

[0267]

[0268] High Adjustment:

[0269]

[0270] In the high adjustment range, if If the change is less than 1%, and The changes will also be small and remain within a reasonable range;

[0271] Further explanation: according to the adjusted angle and scanning parameters, the vehicle is scanned by LiDAR i and LiDAR j to obtain vehicle information, and the vehicle model recognition is completed by the pre-trained neural network algorithm. The original data is matched and recognized to generate output content. The output content includes vehicle model, vehicle height, vehicle length, and vehicle speed data, specifically including the following contents:

[0272] The vehicle type recognition work is completed through the pre-trained neural network algorithm. Please refer to the "Vehicle Type Recognition Method and System Implementation Based on LiDAR Technology" released by Beijing Benewake Photonics Technology Co., Ltd.

[0273] The steps for vehicle model identification are as follows:

[0274] 1) Set the adjusted angle and scanning parameters:

[0275] Get the adjusted angle , and scanning parameters including scanning frequency, scanning angle and transmitting power, to configure LiDAR i and LiDAR j;

[0276] LiDAR i and LiDAR j constitute a LiDAR vehicle type recognition module. By installing structural parts, two TF03 LiDARs, model TF02Pro, are fixed at an angle of θ. LiDAR i is vertically downward, and LiDAR j is tilted downward. The entire module is directly installed on the top of the toll station, under the overpass, or under the traffic light pole through the mounting hole, and connected to the main system through a data cable.

[0277] 2) Vehicle scanning:

[0278] The configured laser radar i and laser radar j are used to scan the vehicles passing through the highway toll station in real time, and the two-dimensional contour data of the vehicles are collected respectively.

[0279] The data fusion algorithm is used to fuse the two-dimensional data collected by the two laser radars to generate the three-dimensional contour data of the vehicle. The fusion formula is:

[0280]

[0281] It is the fused three-dimensional contour data;

[0282] , are the two-dimensional data collected by two laser radars respectively; f is the data fusion function;

[0283] Acquire three-dimensional contour data of the vehicle;

[0284] The scanning frequency is set to 200Hz to ensure complete 3D data acquisition of high-speed vehicles;

[0285] The scanning range is set to 0-50 meters to cover all lanes before the toll booth;

[0286] 3) Data preprocessing:

[0287] Preprocess the raw point cloud data collected by the LiDAR, including denoising, filtering, and data format conversion, and input the processed data into the pre-trained neural network model for recognition;

[0288] The denoising formula is:

[0289]

[0290] in, It is the denoised point cloud data; is the original point cloud data; is the noisy data;

[0291] 4) Input of the neural network model:

[0292] The pre-processed point cloud data is fed into a pre-trained neural network model; this neural network model combines a convolutional neural network (CNN) and a long short-term memory network (LSTM) to effectively identify and classify different vehicle models;

[0293] The input data format is:

[0294]

[0295] is the input data of the neural network model;

[0296] is the three-dimensional coordinate of each point cloud data point;

[0297] 5) Vehicle type identification:

[0298] The neural network model processes the input point cloud data and generates vehicle identification results, including vehicle type, vehicle height, vehicle length and vehicle speed;

[0299] 6) Output result processing and storage:

[0300] The output results of the neural network model are processed, converted into a format suitable for storage and display, and stored in a central database for subsequent query and analysis.

[0301] Embodiment 4:

[0302] See also Figure 2 A front laser radar vehicle type recognition system for a high-speed toll collection system, the system is used to execute the front laser radar vehicle type recognition method for the high-speed toll collection system, comprising:

[0303] Angle marking module: used to determine the angle between the two laser radars in the laser radar vehicle identification module as θ, and mark the vertically downward laser radar as i, and the tilted downward laser radar as j;

[0304] Data processing module: used to collect the environmental data around the two laser radars in real time, and monitor the pollutant characteristic data on the detection end surface of laser radar i and laser radar j respectively, transmit the environmental data and pollutant characteristic data to the central processing unit through the wireless network, and perform data preprocessing;

[0305] Environmental impact coefficient generation module: used to obtain pre-processed environmental data for analysis and generate environmental impact coefficients, which are used to evaluate the impact of environmental data on the accuracy of the lidar;

[0306] Cleanliness influence coefficient generation module: used to obtain the pre-processed pollutant characteristic data on the monitoring laser radar i, and analyze and process it to generate a first cleanliness influence coefficient; obtain the pre-processed pollutant characteristic data on the laser radar j, and analyze and process it to generate a second cleanliness influence coefficient;

[0307] A first adjustment index generation module: used to obtain the environmental impact coefficient and the first cleanliness impact coefficient, and perform analysis and processing to construct a first adjustment index, which is used to generate a first adjustment strategy for dynamically adjusting the scanning parameters of the laser radar i;

[0308] A second adjustment index generation module: used to obtain the environmental impact coefficient and the second cleanliness impact coefficient, and perform analysis and processing to construct a second adjustment index, which is used to generate a second adjustment strategy for dynamically adjusting the scanning parameters of the laser radar j;

[0309] Comparison module: used for setting adjustment thresholds of the first adjustment strategy and the second adjustment strategy, which are the first threshold and the second threshold respectively, and comparing and analyzing the first adjustment strategy with the first threshold to generate a first comparison result; and comparing and analyzing the second adjustment strategy with the second threshold to generate a second comparison result;

[0310] Angle adjustment module: used to obtain the first comparison result and the second comparison result, perform comprehensive analysis, and finally determine the angle after the angle θ is adjusted;

[0311] Fine-tuning calibration module: used to classify the adjustment amount of the angle θ into grades. The larger the adjustment amount, the higher the grade. According to different adjustment amount grades, the first threshold and the second threshold for the next vehicle model recognition are fine-tuned and calibrated to the corresponding values.

[0312] Vehicle type recognition module: It is used to scan the vehicle through LiDAR i and LiDAR j according to the adjusted angle and scanning parameters to obtain vehicle information, and complete the vehicle type recognition work through the pre-trained neural network algorithm, match and identify the original data, and generate output content. The output content includes vehicle type, vehicle height, vehicle length, and vehicle speed data.

[0313] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0314] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0315] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0316] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for identifying vehicle type using a front laser radar in a high-speed toll collection system, the method being applied to a laser radar vehicle type identification module, and characterized in that: The specific steps include: Step S1: Determine that the angle between the two laser radars in the laser radar vehicle type recognition module is θ, and mark the vertically downward laser radar as i, and mark the obliquely downward laser radar as j; Step S2: collect environmental data around the two laser radars in real time, and monitor pollutant characteristic data on the detection end surfaces of laser radar i and laser radar j respectively, transmit the environmental data and pollutant characteristic data to the central processing unit through a wireless network, and perform data preprocessing; The environmental data include temperature, humidity, visibility, and atmospheric optical thickness, and the temperature, humidity, visibility, and atmospheric optical thickness are marked as T, H, V, and AOt in sequence; The pollutant characteristic data include particle deposition, transmittance, and adhesion coefficient. The particle deposition, transmittance, and adhesion coefficient of the monitoring lidar i are marked as PCsi, PTci, and PAni respectively; The particle deposition, transmittance, and adhesion coefficient of the monitoring lidar j are marked as PCsj, PTcj, and PAnj, respectively; Step S3: Obtain the pre-processed environmental data for analysis and generate an environmental impact coefficient, which is used to evaluate the impact of the environmental data on the accuracy of the laser radar; The environmental impact coefficient is defined as EIC, and the calculation formula is as follows: Among them, α is the adjustment coefficient, which controls the amplitude of the overall influence coefficient, β is the adjustment parameter of the temperature influence coefficient, and γ is the adjustment parameter of the humidity influence coefficient. The value ranges of α, β, and γ are all (0, 1), and the value range of EIC is set to (0, 1); When EIC is closer to 0, it means that the environment has a greater impact on the recognition accuracy, and the laser radar recognition accuracy is lower; When EIC is closer to 1, it means that the environment has less impact on the recognition accuracy, and the laser radar recognition accuracy is higher; Step S4: obtaining the pre-processed pollutant characteristic data on the monitoring laser radar i, and analyzing and processing it to generate a first cleanliness influence coefficient; obtaining the pre-processed pollutant characteristic data on the laser radar j, and analyzing and processing it to generate a second cleanliness influence coefficient; The first cleanliness influence coefficient is defined as CICi, and the calculation formula is as follows: The second cleanliness influence coefficient is defined as CICj, and the calculation formula is as follows: Among them, a1 and b1 are adjustment coefficients to control the amplitude of the overall influence coefficient, a2 and b2 are adjustment parameters of the particle deposition influence coefficient, a3 and b3 are adjustment parameters of the transmittance influence coefficient, and the value ranges of a1, a2, a3, b1, b2, and b3 are all (0,1); The value range of CICi and CICj is set to (0,1), and the influence of cleanliness is expressed according to the value range; When CICi or CICj is closer to 0, it means that the pollutant characteristic data has a greater impact on the cleanliness of the lidar, and the cleanliness of the lidar is lower; When CICi or CICj is closer to 1, it means that the influence of pollutant characteristic data on the cleanliness of the lidar is smaller, and the cleanliness of the lidar is higher; Step S5: Obtaining the environmental impact coefficient and the first cleanliness impact coefficient, analyzing and processing them, and constructing a first adjustment index, where the first adjustment index is used to generate a first adjustment strategy for dynamically adjusting the scanning parameters of the laser radar i; The first adjustment index corresponding to the laser radar i is defined as AIi, and the calculation formula is as follows: Among them, k1 and k2 are the weights of EIC and CICi respectively. EIC represents the environmental impact coefficient; CICi represents the first cleanliness impact coefficient; AIi represents the first adjustment index, which is used to generate a first adjustment strategy for dynamically adjusting the scanning parameters of the laser radar i; When EIC decreases and CICi increases, k1>k2 is set to ensure that the change in EIC has a more prominent effect on AIi; When EIC increases and CICi decreases, k2>k1 is set to ensure that the change in CICi has a more prominent impact on AIi; Set the value range of AIi to (0,1); When AIi is closer to 0, it means that the environmental impact and cleanliness impact are greater, and the amount of adjustment of the scanning parameters of the laser radar i is greater; When AIi is closer to 1, it means that the environmental impact and cleanliness impact are smaller, and the need to adjust the scanning parameters of LiDAR i is smaller; The first adjustment strategy includes: The scanning parameters include scanning frequency, scanning angle and transmission power. The comparison threshold of the first adjustment index is set to Q1, the value range of Q1 is 0.35≤Q1≤0.68, and the initial reference value of Q1 is set to 0.46; 1) When the first adjustment index is greater than the threshold value Q1, the scanning frequency increases by 20%, the scanning angle increases by 15%, and the transmission power increases by 25%; 2) If EIC < 0.3, based on step 1), the scanning frequency is increased by 5% and the transmission power is increased by 10%; If CICi<0.3, increase the scanning angle by 5% based on step 1); When the first adjustment index is equal to the threshold Q1, the scanning frequency, scanning angle and transmission power remain unchanged; 3) When the first adjustment index is less than the threshold value Q1, the scanning frequency is reduced by 15%; the scanning angle is reduced by 10%; and the transmission power is reduced by 20%; 4) If EIC>0.7, based on step 3), the scanning frequency is reduced by 5% and the transmission power is reduced by 5%; If CICi>0.7, reduce the scanning angle in step 3) by 5%; Step S6: Obtain the environmental impact coefficient and the second cleanliness impact coefficient, analyze and process them, and construct a second adjustment index, where the second adjustment index is used to generate a second adjustment strategy for dynamically adjusting the scanning parameters of the laser radar j; The second adjustment index corresponding to laser radar j is defined as AIj, and the calculation formula is as follows: Among them, k3 and k4 are the weights of EIC and CICj respectively. When EIC decreases and CICj increases, k3>k4 is set to ensure that the change in EIC has a more prominent effect on AIj; When EIC increases and CICj decreases, k4>k3 is set to ensure that the change of CICj has a more prominent effect on AIj; Set the value range of AIj to (0,1); When AIj is closer to 0, it means that the environment and cleanliness have a greater impact, and the amount of adjustment of the scanning parameters of the laser radar j is greater; When AIj is closer to 1, it means that the impact of environment and cleanliness is smaller, and the need to adjust the scanning parameters of LiDAR j is smaller; The second adjustment strategy includes the following: The comparison threshold of the second adjustment index is set to Q2, the value range of Q2 is 0.38≤Q2≤0.75, and the initial reference value of Q2 is set to 0.52; 1.1) When the second adjustment index is greater than the threshold value Q2, the scanning frequency is reduced by 10%, the scanning angle is reduced by 5%, and the transmission power is reduced by 15%; 1.2) When EIC>0.8, based on step 1.1), the scanning frequency is reduced by 5% and the transmission power is reduced by 6%; When CICj>0.8, the scanning angle is reduced by 4% based on step 1.1); 2.1) When the second adjustment index is equal to the threshold Q2, the scanning frequency, scanning angle and transmission power remain unchanged; 3.1) When the second adjustment index is less than the threshold value Q2, the scanning frequency increases by 20%, the scanning angle increases by 15%, and the transmission power increases by 25%; 3.2) When EIC<0.2, based on step 3.1), the scanning frequency is increased by 5% and the transmission power is increased by 6%; When CICj<0.2, the scanning angle is increased by 6% based on step 3.1); Step S7: setting adjustment thresholds of the first adjustment strategy and the second adjustment strategy, which are the first threshold and the second threshold respectively, and comparing and analyzing the first adjustment strategy with the first threshold to generate a first comparison result; and comparing and analyzing the second adjustment strategy with the second threshold to generate a second comparison result; The first comparison result includes the following: Set the first threshold of the comparison threshold Q1 to r1; 1) When the comparison threshold Q1 is greater than the first threshold r1, the fine-tuning formula for Q1 is as follows: Q1new=Q1+ΔQ1 Wherein, ΔQ1=0.05×(Q1-r1); Q1new is the adjusted threshold value, and 0.35≤Q1new≤0.68; The adjustment formula of the angle θ between laser radar i and laser radar j is: θ1=θ+Δθ Wherein, Δθ=2°×(Q1-r1), Δθ is the angle adjustment amount; θ1 is the adjusted angle; 2) When the comparison threshold Q1 is equal to the fine-tuning threshold r1, Q1 remains unchanged; The angle θ between LiDAR i and LiDAR j remains unchanged; 3) When the comparison threshold Q1 is less than the first threshold r1, the fine-tuning formula for Q1 is as follows: Q1new=Q1-ΔQ1 Where, ΔQ1=0.05×(r1-Q1); Q1new is the adjusted threshold; The adjustment formula of the angle θ between laser radar i and laser radar j is: θ1=θ-Δθ Wherein, Δθ = 2° × (r1-Q1), θ1 is the adjusted angle; The second comparison result includes the following: Set the second threshold of the comparison threshold Q2 to r2; 1) When the comparison threshold Q2 is greater than the second threshold r2, the fine-tuning formula for Q2 is as follows: Q2new=Q2+ΔQ2 Wherein, ΔQ2=0.05×(Q2-r2), Q2new is the adjusted threshold, and 0.38≤Q2new≤0.75; The angle adjustment formula between laser radar i and laser radar j is: θ2=θ+Δθ Wherein, Δθ=2°×(Q2-r2); θ2 is the adjusted angle; 2) When the comparison threshold Q2 is equal to the second threshold r2, Q2 remains unchanged; the angle between the laser radar i and the laser radar j remains unchanged; 3) When the comparison threshold Q2 is less than the second threshold r2, the fine-tuning formula for Q2 is as follows: Q2new=Q2-ΔQ2 Where, ΔQ2 = 0.05 × (r2-Q2), Q2new is the adjusted threshold; The formula for adjusting the angle between LiDAR i and LiDAR j is: θ2=θ-Δθ Wherein, Δθ=2°×(r2-Q2); θ2 is the adjusted angle; Step S8: after obtaining the first comparison result and the second comparison result, a comprehensive analysis is performed, and finally the angle after the included angle θ is adjusted is determined; The angle θ after comprehensive adjustment is defined as θnew, and the calculation formula of θnew is: Among them, θ1 is the adjusted angle calculated according to the first comparison result, θ2 is the adjusted angle calculated according to the second comparison result, and η is the adjustment coefficient used to adjust Adjust the positive and negative of Set the value of η to When θ1 and θ2 are both reducing angles, the value range of η is -1<η<0; When θ1 and θ2 both increase the angle, the value range of η is 0<η<1; When either θ1 or θ2 increases its angle, the other decreases its angle; When the weight of laser radar i is greater than that of laser radar j, -1<η<0; When the weight of laser radar i is less than that of laser radar j, 0<η<1; When lidar i is equal to the weight of lidar j, -0.1<η<0.1; Step S9: the adjustment amount of the angle θ is graded, the larger the adjustment amount, the higher the grade, and the first threshold and the second threshold are fine-tuned and calibrated to the corresponding values ​​for the next vehicle model recognition according to different adjustment amount grades; The adjustment amount of the angle θ is divided into the following levels, and the current adjustment amount of the angle θ is set to Δθ′; Low adjustment means that the angle adjustment is less than 5%; Medium adjustment means that the angle adjustment is between 5%-20%; A high adjustment means that the angle adjustment is greater than 20%; 1) When the adjustment level is low, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle type identification. The fine-tuning formula is: Wherein, r1new is the adjusted first threshold; r2new is the adjusted second threshold; Δθ′ is the current adjustment amount; 2) When the adjustment level is the medium adjustment level, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle type identification. The fine-tuning formula is: 3) When the adjustment level is a high adjustment, the first threshold r1 and the second threshold r2 are fine-tuned and calibrated to the corresponding values ​​for the next vehicle type identification. The fine-tuning formula is: Step S10: According to the adjusted angle and scanning parameters, the vehicle is scanned by laser radar i and laser radar j to obtain vehicle information, and the vehicle model recognition is completed through the pre-trained neural network algorithm, the original data is matched and recognized, and output content is generated. The output content includes vehicle model, vehicle height, vehicle length, and vehicle speed data.

2. The method for identifying vehicle type using a front laser radar of a high-speed toll collection system according to claim 1, characterized in that: The method of determining that the angle between the two laser radars in the laser radar vehicle type recognition module is θ, and marking the vertically downward laser radar as i, and marking the obliquely downward laser radar as j, specifically includes the following contents: 1) Install laser radar equipment; 2) Measure the angle θ; 3) Adjust the installation angle of the laser radar; 4) Label lidar i and lidar j.

3. A front laser radar vehicle type recognition system for a high-speed toll collection system, characterized in that: The system is used to execute the front laser radar vehicle type recognition method of the high-speed toll collection system according to any one of claims 1-2, comprising: Angle marking module: used to determine the angle between the two laser radars in the laser radar vehicle identification module as θ, and mark the vertically downward laser radar as i, and the tilted downward laser radar as j; Data processing module: used to collect the environmental data around the two laser radars in real time, and monitor the pollutant characteristic data on the detection end surface of laser radar i and laser radar j respectively, transmit the environmental data and pollutant characteristic data to the central processing unit through the wireless network, and perform data preprocessing; Environmental impact coefficient generation module: used to obtain pre-processed environmental data for analysis and generate environmental impact coefficients, which are used to evaluate the impact of environmental data on the accuracy of the lidar; Cleanliness influence coefficient generation module: used to obtain the pre-processed pollutant characteristic data on the monitoring laser radar i, and analyze and process it to generate a first cleanliness influence coefficient; obtain the pre-processed pollutant characteristic data on the laser radar j, and analyze and process it to generate a second cleanliness influence coefficient; A first adjustment index generation module: used to obtain the environmental impact coefficient and the first cleanliness impact coefficient, and perform analysis and processing to construct a first adjustment index, which is used to generate a first adjustment strategy for dynamically adjusting the scanning parameters of the laser radar i; A second adjustment index generation module: used to obtain the environmental impact coefficient and the second cleanliness impact coefficient, and perform analysis and processing to construct a second adjustment index, which is used to generate a second adjustment strategy for dynamically adjusting the scanning parameters of the laser radar j; Comparison module: used for setting adjustment thresholds of the first adjustment strategy and the second adjustment strategy, which are the first threshold and the second threshold respectively, and comparing and analyzing the first adjustment strategy with the first threshold to generate a first comparison result; and comparing and analyzing the second adjustment strategy with the second threshold to generate a second comparison result; Angle adjustment module: used to obtain the first comparison result and the second comparison result, perform comprehensive analysis, and finally determine the angle after the angle θ is adjusted; Fine-tuning calibration module: used to classify the adjustment amount of the angle θ into grades. The larger the adjustment amount, the higher the grade. According to different adjustment amount grades, the first threshold and the second threshold for the next vehicle model recognition are fine-tuned and calibrated to the corresponding values. Vehicle type recognition module: It is used to scan the vehicle through LiDAR i and LiDAR j according to the adjusted angle and scanning parameters to obtain vehicle information, and complete the vehicle type recognition work through the pre-trained neural network algorithm, match and identify the original data, and generate output content. The output content includes vehicle type, vehicle height, vehicle length, and vehicle speed data.

Images