Intelligent detection device and method for overloaded vehicles on expressway

Abstract

The invention discloses an intelligent detection device and method for overloaded vehicles on a highway. The device comprises a laser velocimeter, a laser scanner, a pressure power generation device, a video detection module and a data processing module. The laser velocimeter detects the appearance of a moving object and triggers the laser scanner to collect feature information of the moving object; the data processing module determines passing vehicles and vehicle type information of the passing vehicles at the entrance of the expressway according to the feature information of the moving objects; the pressure power generation device collects pressure power generation information when the detected passing vehicle passes through the pressure power generation belt; the data processing module judges whether the detected passing vehicle is overloaded or not according to the vehicle type information and the pressure power generation information of the detected passing vehicle, if the detected passing vehicle is overloaded, license plate information of the overloaded vehicle is shot and transmitted to a traffic management department to be intercepted, and meanwhile the license plate information is sent to a nearby display screen to conduct image and sound dual warning on a driver. According to the invention, overload vehicles can be effectively detected in time, and traffic tragedies caused by overload are greatly reduced.

Description

Technical Field

[0001] The invention relates to an intelligent detection device and method for overloaded vehicles on a highway, belonging to the technical field of traffic safety detection. Background Art

[0002] When an overloaded motor vehicle is on the road, it will endanger the driver's life and safety, and it is also a potential hidden danger for motor vehicles driving normally on the road. The overloaded weight can easily cause the vehicle to lose control, causing great harm to road safety and easily causing traffic accidents. According to incomplete statistics, the number of accidents caused by vehicle failure has increased every year, and has become a major risk factor threatening safe traffic. Among the statistics on vehicle failure accidents, a considerable number of accidents are related to overloading.

[0003] There is a lot of research on this, and most of the devices used in the market are mainly installed in the car, mainly through physical measurement of the car body, automatic shutdown of overload, etc. This kind of prevention method has cumbersome detection steps, and there are certain errors and driving dangers, and it is easy to cause damage and other problems.

[0004] At present, there is a lack of intelligent overload warning equipment installed on the road, which makes it impossible for the road safety management department to make corresponding early warning and law enforcement in time, affecting road safety. Therefore, the present invention proposes an intelligent detection device and method for overloaded vehicles on highways. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes an intelligent detection device and method for overloaded vehicles on highways, which can detect and warn vehicles with abnormal load weight, greatly improving the safety level of drivers driving on highways.

[0006] The technical solution adopted by the present invention to solve the technical problem is:

[0007] In a first aspect, an embodiment of the present invention provides an intelligent detection device for overloaded vehicles on a highway, comprising a laser speed meter, a laser scanner, a pressure power generation device, a video detection module and a data processing module;

[0008] The laser velocimeter detects the presence of a moving object and triggers a laser scanner, which collects characteristic information of the moving object and sends it to a data processing module;

[0009] The data processing module determines the passing vehicles and their vehicle models at the entrance of the expressway based on the characteristic information of the moving object;

[0010] The pressure power generation device collects the pressure power generation information of the detected passing vehicle when passing through the pressure power generation belt, and sends it to the data processing module;

[0011] The data processing module determines whether the detected passing vehicle is overloaded based on the vehicle model information and pressure generation information of the detected passing vehicle. If it is overloaded, the video detection module is triggered to capture the license plate information of the overloaded vehicle, and the license plate information of the overloaded vehicle is transmitted to the traffic management department for interception and processing, and sent to a nearby display screen for dual picture and sound warnings to the driver.

[0012] As a possible implementation of this embodiment, the specific process of the data processing module determining whether the detected passing vehicle is overloaded according to the vehicle type information and pressure power generation information of the detected passing vehicle is as follows:

[0013] The data processing module uses the DBSCAN algorithm and Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicles into a weight data range in real time, and compares it with the standard load weight threshold of the corresponding vehicle model retrieved from the database. If the threshold is exceeded, the detected passing vehicle is overloaded.

[0014] As a possible implementation of this embodiment, the data processing module uses the DBSCAN algorithm and the Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicle into the weight data interval in real time. The specific process is:

[0015] Eli Euclidean distance method A similar relationship is established between the pressure power generation and the power generation duration. After the vehicle passes through the power generation belt, n groups of data in this period are selected from the large amount of new data obtained for sample statistical calculation; the collected data are divided into A1(X1,Y1), A2(X2,Y2), A3(X3,Y3), A4(X4,Y4)...An(Xn,Yn), where X represents the instantaneous pressure power generation and Y represents the power generation duration;

[0016] Given a domain radius EPS, MinPts is calculated by the value of EPS;

[0017] Select the k value, take k as 2*dimension-1, calculate the distance from each point to the k point and arrange them from small to large, and draw a k-distance graph. The most obvious inflection point in the k-distance graph corresponds to the EPS value;

[0018] The value of MinPts is the above k value plus 1, that is, MinPts = k + 1. At this time, the values ​​of EPS and MinPts are determined. Next, the selected (x i ,y j ) number of cluster points;

[0019] From any point Pn(x i ,y j), mark it as allowed to access, check whether it is a core point within the nuclear load threshold range, if it is not a core point within the range, mark it as a noise point, otherwise it is a point Pn(x i ,y j ) Create a new cluster C, and put all objects in the Eps neighborhood of point Pn into the candidate set N temporarily used to store data for clustering;

[0020] Iteratively add objects that do not belong to other clusters from the candidate set N temporarily used to store data to the first cluster C. In this process, for the object p' marked as inaccessible in the set N, mark it as accessible and check its Eps neighborhood. If p' is also a core object, all objects in the Eps neighborhood of p' are added to the set N. Continue to add objects to cluster C until cluster C cannot be expanded, that is, until the set N is an empty set. At this time, cluster C is completely generated. After the first cluster is screened, randomly select the next unvisited object from the remaining clusters, and repeat the above process until all clusters are visited, and the dynamic clustering of the nuclear load data is completed.

[0021] After the dynamic clustering is completed, the Jaccard distance is used to perform similarity analysis on each clustered cluster: among the n clusters that have been established, two clusters are selected for similarity calculation to verify the accuracy of the clustering data of each cluster;

[0022] The two selected clusters are defined as cluster A1 and cluster B1, and the number of elements in the two clusters is initialized respectively:

[0023] |A1|=n,|B1|=m, and then calculate the intersection and union of the two sets respectively: |A∪B|=n+m-

[0024] |A∩B|, after obtaining the values ​​of intersection and union, the similarity of the two clusters is calculated and determined:

[0025]

[0026] When the similarity ratio value When the similarity meets the screening analysis requirements, the data cluster screening is qualified, and the above steps are repeated. The other clusters are selected for analysis. If there are unqualified clusters, each unqualified cluster will be re-clustered and screened for analysis until all are qualified.

[0027] As a possible implementation of this embodiment, the specific process of the data processing module determining whether the detected passing vehicle is overloaded according to the vehicle type information and pressure power generation information of the detected passing vehicle is as follows:

[0028] The data processing module retrieves the pressure generation threshold data corresponding to the vehicle model's rated load weight from the database based on the vehicle model information and pressure generation information of the detected passing vehicle, and compares it with the pressure generation information of the detected passing vehicle. If the threshold is exceeded, the detected passing vehicle is overloaded.

[0029] As a possible implementation method of this embodiment, the laser speed meter and the laser scanner are arranged on a column device, two column devices are grouped together, and the two column devices in each group are respectively arranged on both sides of the road. The pressure power generation device is arranged on the road surface at the column device to provide real-time feedback of the pressure power generation information of passing vehicles when passing through the pressure power generation belt.

[0030] As a possible implementation of this embodiment, there are three groups of column devices, one group is set at the intersection of the curved and straight roads of the highway, one group is set at the entrance of the toll station, and one group is set in the middle of the long straight road between the intersection of the curved and straight roads and the entrance of the toll station.

[0031] In a second aspect, an embodiment of the present invention provides an intelligent detection method for overloaded vehicles on a highway, comprising the following steps:

[0032] Step 1: Use a laser velocimeter to detect the presence of a moving object and trigger a laser scanner to collect feature information of the moving object;

[0033] Step 2, determining the passing vehicles and their vehicle types at the entrance of the expressway according to the characteristic information of the moving object;

[0034] Step 3, using a pressure power generation device to collect pressure power generation information when the detected passing vehicle passes through the pressure power generation belt;

[0035] Step 4: determine whether the detected passing vehicle is overloaded based on the vehicle model information and pressure power generation information of the detected passing vehicle. If it is overloaded, the video detection module is triggered to capture the license plate information of the overloaded vehicle;

[0036] Step 5: The license plate information of the overloaded vehicle is transmitted to the traffic management department for interception and processing, and is also sent to a nearby display screen to provide the driver with a dual warning of image and sound.

[0037] As a possible implementation of this embodiment, in step 4, the specific process of determining whether the detected passing vehicle is overloaded according to the vehicle type information and pressure power generation information of the detected passing vehicle is as follows:

[0038] The DBSCAN algorithm and Kalman filter algorithm are used to convert the pressure power generation information of the detected passing vehicles into a weight data range in real time, and compare it with the standard load weight threshold of the corresponding vehicle model retrieved from the database. If the threshold is exceeded, the detected passing vehicle is overloaded.

[0039] As a possible implementation of this embodiment, the specific process of using the DBSCAN algorithm and the Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicle into the weight data interval in real time is:

[0040] Eli Euclidean distance method A similar relationship is established between the pressure power generation and the power generation duration. After the vehicle passes through the power generation belt, n groups of data in this period are selected from the large amount of new data obtained for sample statistical calculation; the collected data are divided into A1(X1,Y1), A2(X2,Y2), A3(X3,Y3), A4(X4,Y4)...An(Xn,Yn), where X represents the instantaneous pressure power generation and Y represents the power generation duration;

[0041] Given a domain radius EPS, MinPts is calculated by the value of EPS;

[0042] Select the k value, take k as 2*dimension-1, calculate the distance from each point to the k point and arrange them from small to large, and draw a k-distance graph. The most obvious inflection point in the k-distance graph corresponds to the EPS value;

[0043] The value of MinPts is the above k value plus 1, that is, MinPts = k + 1. At this time, the values ​​of EPS and MinPts are determined. Next, the selected (x i ,y j ) number of cluster points;

[0044] From any point Pn(x i ,y j ), mark it as allowed to access, check whether it is a core point within the nuclear load threshold range, if it is not a core point within the range, mark it as a noise point, otherwise it is a point Pn(x i ,y j ) Create a new cluster C, and put all objects in the Eps neighborhood of point Pn into the candidate set N temporarily used to store data for clustering;

[0045] Iteratively add objects that do not belong to other clusters from the candidate set N temporarily used to store data to the first cluster C. In this process, for the object p' marked as inaccessible in the set N, mark it as accessible and check its Eps neighborhood. If p' is also a core object, all objects in the Eps neighborhood of p' are added to the set N. Continue to add objects to cluster C until cluster C cannot be expanded, that is, until the set N is an empty set. At this time, cluster C is completely generated. After the first cluster is screened, randomly select the next unvisited object from the remaining clusters, and repeat the above process until all clusters are visited, and the dynamic clustering of the nuclear load data is completed.

[0046] After the dynamic clustering is completed, the Jaccard distance is used to perform similarity analysis on each clustered cluster: among the n clusters that have been established, two clusters are selected for similarity calculation to verify the accuracy of the clustering data of each cluster;

[0047] The two selected clusters are defined as cluster A1 and cluster B1, and the number of elements in the two clusters is initialized respectively:

[0048] |A1|=n,|B1|=m, and then calculate the intersection and union of the two sets respectively: |A∪B|=n+m-

[0049] |A∩B|, after obtaining the values ​​of intersection and union, the similarity of the two clusters is calculated and determined:

[0050]

[0051] When the similarity ratio value When the similarity meets the screening analysis requirements, the data cluster screening is qualified, and the above steps are repeated. The other clusters are selected for analysis. If there are unqualified clusters, each unqualified cluster will be re-clustered and screened for analysis until all are qualified.

[0052] As a possible implementation of this embodiment, in step 4, the specific process of determining whether the detected passing vehicle is overloaded according to the vehicle type information and pressure power generation information of the detected passing vehicle is as follows:

[0053] According to the vehicle model information and pressure power generation information of the detected passing vehicle, the pressure power generation threshold data corresponding to the vehicle model's rated load weight is retrieved from the database and compared with the pressure power generation information of the detected passing vehicle. If the threshold is exceeded, the detected passing vehicle is overloaded.

[0054] The technical solution of the embodiment of the present invention may have the following beneficial effects:

[0055] The present invention greatly improves the safety level of drivers driving on highways, from vehicle pressure and power generation data detection, vehicle lane and license plate number detection, information processing, to activating a dual warning device to warn vehicles with abnormal load weight. The device is also applicable to various highway sections and has strong versatility.

[0056] The present invention processes and analyzes data through the DBSCAN algorithm and the Kalman filter algorithm, accurately screens and excludes bad data, makes the obtained data accurate and precise, and can intuitively present the intervals where data are densely distributed, which is intuitive and perceptible. The present invention effectively detects overloaded vehicles in a timely manner, realizes timely warning, greatly reduces traffic tragedies caused by overloading, and effectively guarantees the traffic safety of people's travel.

[0057] The present invention determines the similarity between clusters and re-clusters them by using the Jaccard similarity coefficient and the Jaccard distance, which greatly improves the accuracy of cluster data segmentation, further reduces errors, and makes the analyzed data more accurate and reliable.

[0058] The present invention combines laser scanning imaging with intelligent detection of overload devices for the first time, adopts three-primary color pixel comparison, converts three-dimensional objects into data matrices, intuitively displays the static and dynamic state of the object as well as the shape parameters of the object, and better identifies people and vehicles, which helps to accurately and timely detect overloaded vehicles and realize real-time early warning, greatly reducing traffic tragedies caused by overloading and effectively ensuring people's travel safety.

[0059] The present invention automatically collects data throughout the process, and the data is processed by algorithms, which greatly improves the precision and accuracy of judgment, eliminates a large number of interference factors, enhances data credibility, and ensures safe travel. At the same time, the heat map obtained by data processing can more intuitively display the areas, time periods, and sections where overloading problems are frequent, and reasonably allocates police forces based on the real-time vehicle overloading detection data on highways, saving police resources and reducing labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 is a schematic diagram of an intelligent detection device for overloaded vehicles on a highway according to an exemplary embodiment;

[0061] Figure 2 is a schematic diagram showing a k-distance graph and selection of EPS values ​​according to an exemplary embodiment;

[0062] Figure 3 is a schematic diagram of installing a column device according to an exemplary embodiment;

[0063] Figure 4 The present invention is a flow chart showing a method for intelligently detecting overloaded vehicles on a highway according to an exemplary embodiment. DETAILED DESCRIPTION

[0064] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0065] In order to clearly illustrate the technical features of the present solution, the present invention is described in detail below through specific implementation methods and in conjunction with the accompanying drawings. The disclosure below provides many different embodiments or examples for realizing different structures of the present invention. In order to simplify the disclosure of the present invention, the components and settings of specific examples are described below. In addition, the present invention may repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed. It should be noted that the components illustrated in the accompanying drawings are not necessarily drawn to scale. The present invention omits the description of known components and processing techniques and processes to avoid unnecessary limitations on the present invention.

[0066] Example 1

[0067] like Figure 1 As shown, an intelligent detection device for overloaded vehicles on highways provided by an embodiment of the present invention includes a laser speed meter, a laser scanner, a pressure power generation device, a video detection module and a data processing module;

[0068] The laser velocimeter detects the presence of a moving object and triggers a laser scanner, which collects characteristic information of the moving object and sends it to a data processing module;

[0069] The data processing module determines the passing vehicles and their vehicle models at the entrance of the expressway based on the characteristic information of the moving object;

[0070] The pressure power generation device collects the pressure power generation information of the detected passing vehicle when passing through the pressure power generation belt, and sends it to the data processing module;

[0071] The data processing module uses the DBSCAN algorithm and Kalman filtering algorithm to convert the pressure power generation information of the detected passing vehicles into a weight data range in real time, and compare it with the standard load weight threshold of the corresponding vehicle model retrieved from the database. If the threshold is exceeded, the video detection module is triggered to capture the license plate information of the overloaded vehicle, and the license plate information of the overloaded vehicle is transmitted to the traffic management department for interception and processing, and sent to a nearby display screen for dual picture and sound warnings to the driver.

[0072] As a possible implementation of this embodiment, the data processing module uses the DBSCAN algorithm and the Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicle into the weight data interval in real time. The specific process is:

[0073] Eli Euclidean distance method A similar relationship is established between the pressure power generation and the power generation duration. After the vehicle passes through the power generation belt, n groups of data in this period are selected from the large amount of new data obtained for sample statistical calculation; the collected data are divided into A1(X1,Y1), A2(X2,Y2), A3(X3,Y3), A4(X4,Y4)...An(Xn,Yn), where X represents the instantaneous pressure power generation and Y represents the power generation duration;

[0074] In order to establish data clusters, it is necessary to first determine the domain scope of each cluster in order to better screen and filter data. First, the domain radius Eps is given and MinPts is selected through it.

[0075] Given a neighborhood radius EPS (neighborhood radius value, full name Eps-neighborhood of a point), and calculate MinPts (a value to determine whether a point is a core point, edge point, or discrete irrelevant point, the minimum threshold of the number of points, full name minPoints) through the value of EPS;

[0076] Select the k value, take k as 2*dimension-1, calculate the distance from each point to the k point and arrange them from small to large, and draw a k-distance graph, such as Figure 2 As shown in the figure, the most obvious inflection point in the k-distance graph corresponds to the EPS value;

[0077] The value of MinPts is the above k value plus 1, that is, MinPts = k + 1. At this time, the values ​​of EPS and MinPts are determined. Next, the selected (x i ,y j ) number of cluster points;

[0078] From any point Pn(x i ,y j ), mark it as allowed to access, check whether it is a core point within the core load threshold (that is, the neighborhood of P with Eps as the radius has at least MinPts objects), if it is not a core point within the range, mark it as a noise point, otherwise it is a point Pn(x i ,y j ) Create a new cluster C, and put all objects in the Eps neighborhood of point Pn into the candidate set N temporarily used to store data for clustering;

[0079] Iteratively add objects that do not belong to other clusters from the candidate set N temporarily used to store data to the first cluster C. In this process, for the object p' marked as inaccessible in the set N, mark it as accessible and check its Eps neighborhood. If p' is also a core object, all objects in the Eps neighborhood of p' are added to the set N. Continue to add objects to cluster C until cluster C cannot be expanded, that is, until the set N is an empty set. At this time, cluster C is completely generated. After the first cluster is screened, randomly select the next unvisited object from the remaining clusters, and repeat the above process until all clusters are visited, and the dynamic clustering of the nuclear load data is completed.

[0080] After the dynamic clustering is completed, the Jaccard distance is used to perform similarity analysis on each clustered cluster: among the n clusters that have been established, two clusters are selected for similarity calculation to verify the accuracy of the clustering data of each cluster;

[0081] The two selected clusters are defined as cluster A1 and cluster B1, and the number of elements in the two clusters is initialized respectively:

[0082] |A1|=n,|B1|=m, and then calculate the intersection and union of the two sets respectively: |A∪B|=n+m-

[0083] |A∩B|, after obtaining the values ​​of intersection and union, the similarity of the two clusters is calculated and determined:

[0084]

[0085] When the similarity ratio value When the similarity meets the screening analysis requirements, the data cluster screening is qualified, and the above steps are repeated. The other clusters are selected for analysis. If there are unqualified clusters, each unqualified cluster will be re-clustered and screened for analysis until all are qualified.

[0086] A major advantage of the algorithm used in the present invention is that it can find clusters of any shape, not just convex clusters. At the same time, it can also find noise points and form them into an independent cluster. The multiple core points concentrated on the nuclear load threshold and the points connected by their density are given independent cluster labels, and the noise points are also given independent cluster labels.

[0087] As a possible implementation method of this embodiment, the laser speed meter and the laser scanner are arranged on a column device, two column devices are grouped together, and the two column devices in each group are respectively arranged on both sides of the road. The pressure power generation device is arranged on the road surface at the column device to provide real-time feedback of the pressure power generation information of passing vehicles when passing through the pressure power generation belt.

[0088] As a possible implementation of this embodiment, Figure 3 As shown, the column device has three groups A, B, and C. Group A is set at the intersection of the curved and straight roads of the expressway, group C is set at the entrance of the toll station, and group B is set in the middle of the long straight road between the intersection of the curved and straight roads and the entrance of the toll station.

[0089] As a possible implementation of this embodiment, the core of the data processing module is a 51 single-chip microcomputer and its corresponding database, and a Zigbee Internet of Things module is used for real-time information transmission. The laser scanner is a laser radar sensor (Beixing TF03).

[0090] The power supply unit of the intelligent detection device adopts a triple power supply mode of solar panels + mains access + battery pack for real-time power supply. Under normal circumstances, 220v mains power is used for power supply; in remote suburbs or areas with complex terrain and environment, solar panels and battery packs are used for power supply. The battery pack adopts 20v, 120Ah specifications. When fully charged, it can meet the battery life of 1 month. The maintenance cycle is once a year.

[0091] Example 2

[0092] like Figure 1 As shown, an intelligent detection device for overloaded vehicles on highways provided by an embodiment of the present invention includes a laser speed meter, a laser scanner, a pressure power generation device, a video detection module and a data processing module;

[0093] The laser velocimeter detects the presence of a moving object and triggers a laser scanner, which collects characteristic information of the moving object and sends it to a data processing module;

[0094] The data processing module determines the passing vehicles and their vehicle models at the entrance of the expressway based on the characteristic information of the moving object;

[0095] The pressure power generation device collects the pressure power generation information of the detected passing vehicle when passing through the pressure power generation belt, and sends it to the data processing module;

[0096] The data processing module retrieves the pressure generation threshold data corresponding to the vehicle model's rated load weight from the database based on the vehicle model information and pressure generation information of the detected passing vehicle, and compares it with the pressure generation information of the detected passing vehicle. If the threshold is exceeded, the video detection module is triggered to capture the license plate information of the overloaded vehicle, and the license plate information of the overloaded vehicle is transmitted to the traffic management department for interception and processing, and sent to a nearby display screen for dual picture and sound warnings to the driver.

[0097] As a possible implementation method of this embodiment, the laser speed meter and the laser scanner are arranged on a column device, two column devices are grouped together, and the two column devices in each group are respectively arranged on both sides of the road. The pressure power generation device is arranged on the road surface at the column device to provide real-time feedback of the pressure power generation information of passing vehicles when passing through the pressure power generation belt.

[0098] As a possible implementation of this embodiment, there are three groups of column devices, one group is set at the intersection of the curved and straight roads of the highway, one group is set at the entrance of the toll station, and one group is set in the middle of the long straight road between the intersection of the curved and straight roads and the entrance of the toll station.

[0099] Example 3

[0100] like Figure 4 As shown, an intelligent detection method for overloaded vehicles on a highway provided by an embodiment of the present invention comprises the following steps:

[0101] Step 1: Use a laser velocimeter to detect the presence of a moving object and trigger a laser scanner to collect feature information of the moving object;

[0102] Step 2, determining the passing vehicles and their vehicle types at the entrance of the expressway according to the characteristic information of the moving object;

[0103] Step 3, using a pressure power generation device to collect pressure power generation information when the detected passing vehicle passes through the pressure power generation belt;

[0104] Step 4: Use the DBSCAN algorithm and Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicle into a weight data interval in real time, and compare it with the standard load weight threshold of the corresponding vehicle model retrieved from the database. If it exceeds the threshold, the detected passing vehicle is overloaded; if it is overloaded, the video detection module is triggered to capture the license plate information of the overloaded vehicle;

[0105] Step 5: The license plate information of the overloaded vehicle is transmitted to the traffic management department for interception and processing, and is also sent to a nearby display screen to give the driver a dual warning of image and sound, prompting the driver to exit the expressway at the nearest intersection. The data processing module marks the vehicle data and detects whether the vehicle has exited the expressway. If it is detected that the vehicle has not exited the expressway, its data information is reported to the traffic police department for processing.

[0106] As a possible implementation of this embodiment, the specific process of using the DBSCAN algorithm and the Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicle into the weight data interval in real time is:

[0107] Eli Euclidean distance method A similar relationship is established between the pressure power generation and the power generation duration. After the vehicle passes through the power generation belt, n groups of data in this period are selected from the large amount of new data obtained for sample statistical calculation; the collected data are divided into A1(X1,Y1), A2(X2,Y2), A3(X3,Y3), A4(X4,Y4)...An(Xn,Yn), where X represents the instantaneous pressure power generation and Y represents the power generation duration;

[0108] Given a domain radius EPS, MinPts is calculated by the value of EPS;

[0109] Select the k value, take k as 2*dimension-1, calculate the distance from each point to the k point and arrange them from small to large, and draw a k-distance graph. The most obvious inflection point in the k-distance graph corresponds to the EPS value;

[0110] The value of MinPts is the above k value plus 1, that is, MinPts = k + 1. At this time, the values ​​of EPS and MinPts are determined. Next, the selected (x i ,y j ) number of cluster points;

[0111] From any point Pn9x i ,y j ), mark it as allowed to access, check whether it is a core point within the core load threshold (that is, the neighborhood of P with Eps as the radius has at least MinPts objects), if it is not a core point within the range, mark it as a noise point, otherwise it is a point Pn(x i ,y j ) Create a new cluster C, and put all objects in the Eps neighborhood of point Pn into the candidate set N temporarily used to store data for clustering;

[0112] Iteratively add objects that do not belong to other clusters from the candidate set N temporarily used to store data to the first cluster C. In this process, for the object p' marked as inaccessible in the set N, mark it as accessible and check its Eps neighborhood. If p' is also a core object, all objects in the Eps neighborhood of p' are added to the set N. Continue to add objects to cluster C until cluster C cannot be expanded, that is, until the set N is an empty set. At this time, cluster C is completely generated. After the first cluster is screened, randomly select the next unvisited object from the remaining clusters, and repeat the above process until all clusters are visited, and the dynamic clustering of the nuclear load data is completed.

[0113] After the dynamic clustering is completed, the Jaccard distance is used to perform similarity analysis on each clustered cluster: among the n clusters that have been established, two clusters are selected for similarity calculation to verify the accuracy of the clustering data of each cluster;

[0114] The two selected clusters are defined as cluster A1 and cluster B1, and the number of elements in the two clusters is initialized respectively:

[0115] |A1|=n,|B1|=m, and then calculate the intersection and union of the two sets respectively: |A∪B|=n+m-

[0116] |A∩B|, after obtaining the values ​​of intersection and union, the similarity of the two clusters is calculated and determined:

[0117]

[0118] When the similarity ratio value When the similarity meets the screening analysis requirements, the data cluster screening is qualified, and the above steps are repeated. The other clusters are selected for analysis. If there are unqualified clusters, each unqualified cluster will be re-clustered and screened for analysis until all are qualified.

[0119] After clustering and screening all data, the current vehicle load is P1. The basic vehicle information and vehicle type can be obtained by high-speed continuous shooting and laser scanning of the video detection device. After obtaining the basic information of the vehicle, the load weight corresponding to the nuclear load power generation power of this type of vehicle is searched in the database as P0. P1 is compared with P0. If the P1 value is greater than P0, the warning module starts the audio-visual dual warning to remind the driver to leave the highway at the nearest fork. After the alarm, the vehicle feature information is marked. If the subsequent device still detects that the vehicle is still driving on the highway, the information processing module reports the relevant information of the vehicle to the traffic police department to notify the stop. The highway vehicle load threshold is shown in Table 1. For example, the normal load of a truck in a two-axle vehicle is less than or equal to 18 tons (t), and it is a dangerous load if it exceeds 18 tons; and the highway vehicle load threshold of a truck in a four-axle vehicle is 31 tons.

[0120] Table 1: Highway vehicle load thresholds:

[0121]

[0122] Example 4

[0123] like Figure 4 As shown, an intelligent detection method for overloaded vehicles on a highway provided by an embodiment of the present invention comprises the following steps:

[0124] Step 1: Use a laser velocimeter to detect the presence of a moving object and trigger a laser scanner to collect feature information of the moving object;

[0125] Step 2, determining the passing vehicles and their vehicle types at the entrance of the expressway according to the characteristic information of the moving object;

[0126] Step 3, using a pressure power generation device to collect pressure power generation information when the detected passing vehicle passes through the pressure power generation belt;

[0127] Step 4: According to the vehicle model information and pressure power generation information of the detected passing vehicle, the pressure power generation threshold data corresponding to the vehicle model's rated load weight is retrieved from the database, and compared with the pressure power generation information of the detected passing vehicle. If the threshold is exceeded, the detected passing vehicle is overloaded; if overloaded, the video detection module is triggered to capture the license plate information of the overloaded vehicle;

[0128] Step 5: The license plate information of the overloaded vehicle is transmitted to the traffic management department for interception and processing, and is also sent to a nearby display screen to provide the driver with a dual warning of image and sound.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. An intelligent detection device for overloaded vehicles on highways, characterized in that: It includes a laser speed meter, a laser scanner, a pressure power generation device, a video detection module and a data processing module; The laser velocimeter detects the presence of a moving object and triggers a laser scanner, which collects characteristic information of the moving object and sends it to a data processing module; The data processing module determines the passing vehicles and their vehicle models at the entrance of the expressway based on the characteristic information of the moving object; The pressure power generation device collects the pressure power generation information of the detected passing vehicle when passing through the pressure power generation belt, and sends it to the data processing module; The data processing module determines whether the detected passing vehicle is overloaded based on the vehicle model information and pressure generation information of the detected passing vehicle. If it is overloaded, the video detection module is triggered to capture the license plate information of the overloaded vehicle, and the license plate information of the overloaded vehicle is transmitted to the traffic management department for interception and processing, and sent to a nearby display screen for dual picture and sound warnings to the driver.

2. The intelligent detection device for overloaded vehicles on highways according to claim 1 is characterized in that: The specific process of the data processing module determining whether the detected passing vehicle is overloaded according to the vehicle type information and pressure power generation information of the detected passing vehicle is as follows: The data processing module uses the DBSCAN algorithm and Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicles into a weight data range in real time, and compares it with the standard load weight threshold of the corresponding vehicle model retrieved from the database. If the threshold is exceeded, the detected passing vehicle is overloaded.

3. The intelligent detection device for overloaded vehicles on highways according to claim 2 is characterized in that: The specific process of the data processing module using the DBSCAN algorithm and the Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicle into the weight data interval in real time is as follows: Eli Euclidean distance method A similar relationship is established between the pressure power generation and the power generation duration. After the vehicle passes through the power generation belt, n groups of data in this period are selected from the large amount of new data obtained for sample statistical calculation; the collected data are divided into A1(X1,Y1), A2(X2,Y2), A3(X3,Y3), A4(X4,Y4)...An(Xn,Yn), where X represents the instantaneous pressure power generation and Y represents the power generation duration; Given a domain radius EPS, MinPts is calculated by the value of EPS; Select the k value, take k as 2*dimension-1, calculate the distance from each point to the k point and arrange them from small to large, and draw a k-distance graph. The most obvious inflection point in the k-distance graph corresponds to the EPS value; The value of MinPts is the above k value plus 1, that is, MinPts = k + 1. At this time, the values ​​of EPS and MinPts are determined. Next, the selected (x i ,y j ) number of cluster points; From any point Pn(x i ,y j ), mark it as allowed to access, check whether it is a core point within the nuclear load threshold range, if it is not a core point within the range, mark it as a noise point, otherwise it is a point Pn(x i ,y j ) Create a new cluster C, and put all objects in the Eps neighborhood of point Pn into the candidate set N temporarily used to store data for clustering; Iteratively add objects that do not belong to other clusters from the candidate set N temporarily used to store data to the first cluster C. In this process, for the object p' marked as inaccessible in the set N, mark it as accessible and check its Eps neighborhood. If p' is also a core object, all objects in the Eps neighborhood of p' are added to the set N. Continue to add objects to cluster C until cluster C cannot be expanded, that is, until the set N is an empty set. At this time, cluster C is completely generated. After the first cluster is screened, randomly select the next unvisited object from the remaining clusters, and repeat the above process until all clusters are visited, and the dynamic clustering of the nuclear load data is completed. After the dynamic clustering is completed, the Jaccard distance is used to perform similarity analysis on each clustered cluster: among the n clusters that have been established, two clusters are selected for similarity calculation to verify the accuracy of the clustering data of each cluster; The two selected clusters are defined as cluster A1 and cluster B1, and the number of elements in the two clusters is initialized respectively: |A1|=n,|B1|=m, and then calculate the intersection and union of the two sets respectively: |A∪B|=n+m-|A∩B|. After obtaining the values ​​of the intersection and union, the similarity of the two clusters is calculated and determined: When the similarity ratio value When the similarity meets the screening analysis requirements, the data cluster screening is qualified, and the above steps are repeated. The other clusters are selected for analysis. If there are unqualified clusters, each unqualified cluster will be re-clustered and screened for analysis until all are qualified.

4. The intelligent detection device for overloaded vehicles on highways according to claim 1 is characterized in that: The specific process of the data processing module determining whether the detected passing vehicle is overloaded according to the vehicle type information and pressure power generation information of the detected passing vehicle is as follows: The data processing module retrieves the pressure generation threshold data corresponding to the vehicle model's rated load weight from the database based on the vehicle model information and pressure generation information of the detected passing vehicle, and compares it with the pressure generation information of the detected passing vehicle. If the threshold is exceeded, the detected passing vehicle is overloaded.

5. The intelligent detection device for overloaded vehicles on highways according to any one of claims 1 to 4, characterized in that: The laser speed meter and laser scanner are arranged on column devices, two column devices are arranged in a group, and the two column devices in each group are arranged on both sides of the road respectively. The pressure power generation device is arranged on the road surface at the column device to provide real-time feedback of pressure power generation information when passing vehicles pass through the pressure power generation belt.

6. The intelligent detection device for overloaded vehicles on highways according to claim 5, characterized in that: There are three groups of column devices, one group is set at the intersection of the curved and straight roads of the expressway, one group is set at the entrance of the toll station, and one group is set in the middle of the long straight road between the intersection of the curved and straight roads and the entrance of the toll station.

7. An intelligent detection method for overloaded vehicles on highways, characterized in that: The following steps are involved: Step 1: Use a laser velocimeter to detect the presence of a moving object and trigger a laser scanner to collect feature information of the moving object; Step 2, determining the passing vehicles and their vehicle types at the entrance of the expressway according to the characteristic information of the moving object; Step 3, using a pressure power generation device to collect pressure power generation information when the detected passing vehicle passes through the pressure power generation belt; Step 4: determine whether the detected passing vehicle is overloaded based on the vehicle model information and pressure power generation information of the detected passing vehicle. If it is overloaded, the video detection module is triggered to capture the license plate information of the overloaded vehicle; Step 5: The license plate information of the overloaded vehicle is transmitted to the traffic management department for interception and processing, and is also sent to a nearby display screen to provide the driver with a dual warning of image and sound.

8. The intelligent detection method for overloaded vehicles on highways according to claim 7 is characterized in that: In step 4, the specific process of determining whether the detected passing vehicle is overloaded according to the vehicle type information and pressure power generation information of the detected passing vehicle is as follows: The DBSCAN algorithm and Kalman filter algorithm are used to convert the pressure power generation information of the detected passing vehicles into a weight data range in real time, and compare it with the standard load weight threshold of the corresponding vehicle model retrieved from the database. If the threshold is exceeded, the detected passing vehicle is overloaded.

9. The intelligent detection method for overloaded vehicles on highways according to claim 8, characterized in that: The specific process of using the DBSCAN algorithm and the Kalman filter algorithm to convert the pressure power generation information of the detected passing vehicle into the weight data interval in real time is as follows: Eli Euclidean distance method A similar relationship is established between the pressure power generation and the power generation duration. After the vehicle passes through the power generation belt, n groups of data in this period are selected from the large amount of new data obtained for sample statistical calculation; the collected data are divided into A1(X1,Y1), A2(X2,Y2), A3(X3,Y3), A4(X4,Y4)...An(Xn,Yn), where X represents the instantaneous pressure power generation and Y represents the power generation duration; Given a domain radius EPS, MinPts is calculated by the value of EPS; Select the k value, take k as 2*dimension-1, calculate the distance from each point to the k point and arrange them from small to large, and draw a k-distance graph. The most obvious inflection point in the k-distance graph corresponds to the EPS value; The value of MinPts is the above k value plus 1, that is, MinPts = k + 1. At this time, the values ​​of EPS and MinPts are determined. Next, the selected (x i ,y j ) number of cluster points; From any point Pn(x i ,y j ), mark it as allowed to access, check whether it is a core point within the nuclear load threshold range, if it is not a core point within the range, mark it as a noise point, otherwise it is a point Pn(x i ,y j ) Create a new cluster C, and put all objects in the Eps neighborhood of point Pn into the candidate set N temporarily used to store data for clustering; Iteratively add objects that do not belong to other clusters from the candidate set N temporarily used to store data to the first cluster C. In this process, for the object p' marked as inaccessible in the set N, mark it as accessible and check its Eps neighborhood. If p' is also a core object, all objects in the Eps neighborhood of p' are added to the set N. Continue to add objects to cluster C until cluster C cannot be expanded, that is, until the set N is an empty set. At this time, cluster C is completely generated. After the first cluster is screened, randomly select the next unvisited object from the remaining clusters, and repeat the above process until all clusters are visited, and the dynamic clustering of the nuclear load data is completed. After the dynamic clustering is completed, the Jaccard distance is used to perform similarity analysis on each clustered cluster: among the n clusters that have been established, two clusters are selected for similarity calculation to verify the accuracy of the clustering data of each cluster; The two selected clusters are defined as cluster A1 and cluster B1, and the number of elements in the two clusters is initialized respectively: |A1|=n,|B1|=m, and then calculate the intersection and union of the two sets respectively: |A∪B|=n+m-|A∩B|. After obtaining the values ​​of the intersection and union, the similarity of the two clusters is calculated and determined: When the similarity ratio value When the similarity meets the screening analysis requirements, the data cluster screening is qualified, and the above steps are repeated. The other clusters are selected for analysis. If there are unqualified clusters, each unqualified cluster will be re-clustered and screened for analysis until all are qualified.

10. The intelligent detection method for overloaded vehicles on highways according to claim 7, characterized in that: In step 4, the specific process of determining whether the detected passing vehicle is overloaded according to the vehicle type information and pressure power generation information of the detected passing vehicle is as follows: According to the vehicle model information and pressure power generation information of the detected passing vehicle, the pressure power generation threshold data corresponding to the vehicle model's rated load weight is retrieved from the database and compared with the pressure power generation information of the detected passing vehicle. If the threshold is exceeded, the detected passing vehicle is overloaded.

Images