A method and system for tracking the trajectory of a transport vehicle for road construction

Abstract

The present invention is applicable to the technical field of transport vehicle supervision, and provides a method and system for tracking the trajectory of transport vehicles used for road construction, including the following steps: determining the operation trajectory information of the transport vehicle based on a GPS positioning device, the operation trajectory information including time information, operation tracking trajectory and license plate number; receiving a loading monitoring video, analyzing the loading monitoring video to determine the license plate number and material information, and adding the material information to the operation trajectory information; determining the protection characteristics according to the material information, and determining whether the transport protection is qualified based on the protection characteristics and the loading monitoring video, and generating a first warning information when it is unqualified; determining whether a special vehicle is required according to the material information, and when required, determining whether the corresponding vehicle is compliant. The present invention realizes automatic monitoring and recording of material information by analyzing the loading monitoring video to determine the license plate number and material information, and determines the protection characteristics through the material information, thereby strengthening the supervision of transport protection.

Description

Technical Field

[0001] The present invention relates to the technical field of transport vehicle supervision, and in particular to a method and system for tracking the trajectory of a transport vehicle used in road construction. Background Art

[0002] In road construction projects, transport vehicles are responsible for delivering various materials such as sand, cement, asphalt, etc. to the construction site. However, traditional transport vehicle management methods have many shortcomings, especially in terms of material monitoring, transportation protection, and compliance inspections. At present, the loading process lacks effective monitoring, and the types of materials often rely on manual records and cannot be synchronized to the transportation trajectory in a timely manner. This not only increases labor costs, but also makes it difficult to ensure the accuracy and completeness of the data. In addition, due to the particularity of the nature of the materials (such as easy to spill, easy to contaminate, etc.), the protection requirements for transport vehicles are extremely high. However, traditional supervision methods often rely on on-site inspections, which makes it difficult to achieve comprehensive coverage and real-time early warning. Therefore, it is necessary to provide a method and system for tracking the trajectory of transport vehicles for road construction, aiming to solve the above problems. Summary of the invention

[0003] In view of the deficiencies in the prior art, the object of the present invention is to provide a method and system for tracking the trajectory of a transport vehicle for road construction, so as to solve the problems existing in the above-mentioned background technology.

[0004] The present invention is implemented as follows: a method for tracking the trajectory of a transport vehicle for road construction, the method comprising the following steps:

[0005] Determine the running track information of the transport vehicle based on the GPS positioning device, wherein the running track information includes time information, running tracking track and license plate number;

[0006] Receive the feeding monitoring video uploaded by the camera at the feeding point, analyze the feeding monitoring video to determine the license plate number and material information, and add the material information to the running track information;

[0007] Determine the protection features according to the material information, and determine whether the transportation protection is qualified based on the protection features and the loading monitoring video. If it is unqualified, generate the first warning information;

[0008] Determine whether a special vehicle is required based on the material information, and if a special vehicle is required, determine whether the corresponding vehicle is compliant based on the license plate number, and if it is not compliant, generate a second warning information;

[0009] Based on the operation trajectory information, the average daily transportation time of each type of transport vehicle is determined, and transport vehicle reminder information is generated.

[0010] As a further solution of the present invention: the step of analyzing the feeding monitoring video to determine the license plate number and material information specifically includes:

[0011] After detecting the license plate area of ​​the vehicle, use OCR technology to perform character recognition and determine the license plate number;

[0012] Perform image preprocessing on the loading materials in the feeding monitoring video to extract the color, texture and shape features of the materials;

[0013] The preprocessed material image is input into the trained material recognition model to output the material information.

[0014] As a further solution of the present invention: the step of determining the protection feature according to the material information and judging whether the transportation protection is qualified based on the protection feature and the loading monitoring video specifically includes:

[0015] Input material information into the material protection feature library, and output corresponding protection features, wherein the protection features include protection product features and installation features;

[0016] Perform feature recognition on the feeding monitoring video to determine whether the feeding monitoring video contains the protective product features. If not, directly determine that the transportation protection is unqualified; if contained, proceed to the next step;

[0017] Whether the protective equipment in the feeding monitoring video is correctly installed is determined based on the installation features, and the installation features are used to characterize the position information of the protective equipment.

[0018] As a further solution of the present invention: the step of determining whether a special vehicle is required based on the material information, and when a special vehicle is required, determining whether the corresponding vehicle is compliant based on the license plate number, specifically includes:

[0019] Input material information into the special vehicle library. When special vehicle information is output, it is required to use a special vehicle. The special vehicle information includes the special vehicle category.

[0020] Input the license plate number into the vehicle information database and output the specific vehicle information;

[0021] Determine whether the corresponding transport vehicle is compliant based on the vehicle-specific information and special vehicle category.

[0022] As a further solution of the present invention: the step of determining the average daily transportation time of each type of transportation vehicle based on the operation trajectory information and generating transportation vehicle reminder information specifically includes:

[0023] Determine the average daily transportation time of each vehicle based on the operation trajectory information, classify all transportation vehicles according to the license plate number, and determine the average daily transportation time of each type of transportation vehicle;

[0024] Retrieve the average daily transport limit range for each type of transport vehicle, and generate transport vehicle reminder information based on the average daily transport limit range and the average daily transport time.

[0025] Another object of the present invention is to provide a transport vehicle trajectory tracking system for road construction, the system comprising:

[0026] An operation track determination module is used to determine the operation track information of the transport vehicle based on the GPS positioning device, wherein the operation track information includes time information, operation tracking track and license plate number;

[0027] The material information adding module is used to receive the feeding monitoring video uploaded by the feeding point camera, analyze the feeding monitoring video to determine the license plate number and material information, and add the material information to the running track information;

[0028] A transport protection determination module is used to determine the protection characteristics according to the material information, and determine whether the transport protection is qualified based on the protection characteristics and the feeding monitoring video. If it is unqualified, a first warning information is generated;

[0029] A special vehicle determination module is used to determine whether a special vehicle is required according to the material information, and when a special vehicle is required, determine whether the corresponding vehicle is compliant based on the license plate number, and generate a second warning information when it is not compliant;

[0030] The transport vehicle reminder module is used to determine the average daily transport time of each type of transport vehicle based on the operation trajectory information and generate transport vehicle reminder information.

[0031] As a further solution of the present invention: the material information adding module includes:

[0032] The license plate number recognition unit is used to detect the license plate area of ​​the vehicle and use OCR technology to perform character recognition to determine the license plate number;

[0033] A material feature extraction unit is used to perform image preprocessing on the loaded materials in the loading monitoring video to extract the color features, texture features and shape features of the materials;

[0034] The material information output unit is used to input the preprocessed material image into the trained material recognition model and output the material information.

[0035] As a further solution of the present invention: the transportation protection determination module includes:

[0036] A protection feature determination unit, used to input material information into a material protection feature library and output corresponding protection features, wherein the protection features include protection product features and installation features;

[0037] The protective article recognition unit is used to perform feature recognition on the feeding monitoring video to determine whether the feeding monitoring video contains the protective article feature. If not, it is directly determined that the transportation protection is unqualified; if it is contained, the next step is executed;

[0038] The installation feature determination unit is used to determine whether the protective article in the feeding monitoring video is correctly installed according to the installation features, and the installation features are used to characterize the position information of the protective article.

[0039] As a further solution of the present invention: the dedicated vehicle determination module includes:

[0040] The special vehicle information unit is used to input material information into the special vehicle library. When there is special vehicle information output, it is required to use a special vehicle. The special vehicle information includes the special vehicle category.

[0041] The vehicle specific information unit is used to input the license plate number into the vehicle information database and output the vehicle specific information;

[0042] The vehicle compliance determination unit is used to determine whether the corresponding transport vehicle is compliant based on the vehicle specific information and the special vehicle category.

[0043] As a further solution of the present invention: the transport vehicle reminder module includes:

[0044] The daily average transportation time unit is used to determine the daily average transportation time of each vehicle based on the operation trajectory information, classify all transportation vehicles according to the license plate number, and determine the daily average transportation time of each type of transportation vehicle;

[0045] The reminder information generating unit is used to retrieve the daily average transportation limit range of each type of transportation vehicle, and generate transportation vehicle reminder information according to the daily average transportation limit range and the daily average transportation time.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] The present invention determines the license plate number and material information by analyzing the loading monitoring video, and adds the material information to the running track information to realize the automatic monitoring and recording of the material information. The protection characteristics are determined by the material information, and whether the transportation protection is qualified is determined based on the protection characteristics and the loading monitoring video. When it is unqualified, the management personnel are reminded to intervene in time, thereby strengthening the supervision of transportation protection. In addition, it is also determined whether a special vehicle is required based on the material information. When a special vehicle is required, it is determined whether the corresponding vehicle is compliant based on the license plate number to ensure the transportation safety and compliance of special materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The present invention is a flow chart of a method for tracking the trajectory of a transport vehicle used in road construction.

[0049] Figure 2 The present invention is a flow chart for determining the license plate number and material information in a method for tracking the trajectory of a transport vehicle used in road construction.

[0050] Figure 3 The present invention is a flow chart for determining whether transport protection is qualified in a method for tracking the trajectory of a transport vehicle used in road construction.

[0051] Figure 4 The present invention is a flow chart for determining whether a transport vehicle is compliant in a method for tracking the trajectory of a transport vehicle used for road construction.

[0052] Figure 5 The present invention is a flow chart of generating transport vehicle reminder information in a transport vehicle trajectory tracking method for road construction.

[0053] Figure 6 The figure is a structural schematic diagram of a transport vehicle trajectory tracking system for road construction. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0055] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0056] like Figure 1 As shown, an embodiment of the present invention provides a method for tracking the trajectory of a transport vehicle for road construction, the method comprising the following steps:

[0057] S100, determining the running track information of the transport vehicle based on the GPS positioning device, wherein the running track information includes time information, running tracking track and license plate number;

[0058] S200, receiving the feeding monitoring video uploaded by the camera at the feeding point, analyzing the feeding monitoring video to determine the license plate number and material information, and adding the material information to the running track information;

[0059] S300, determining protection features according to material information, determining whether transportation protection is qualified based on the protection features and the loading monitoring video, and generating first warning information when it is unqualified;

[0060] S400, determining whether a special vehicle is required based on the material information, and if a special vehicle is required, determining whether the corresponding vehicle is compliant based on the license plate number, and if it is not compliant, generating a second warning information;

[0061] S500, determining the average daily transportation time of each type of transportation vehicle based on the operation track information, and generating transportation vehicle reminder information.

[0062] It should be noted that in road construction projects, traditional means of transport vehicle management have many shortcomings, especially in terms of material monitoring, transport protection, and compliance inspections. At present, the loading process lacks effective monitoring, and the types of materials often rely on manual recording, which cannot be synchronized to the transportation track in a timely manner. This not only increases labor costs, but also makes it difficult to ensure the accuracy and completeness of the data. In addition, due to the particularity of the nature of the materials (such as easy to spill, easy to contaminate, etc.), the protection requirements for transport vehicles are extremely high. However, traditional supervision methods often rely on on-site inspections, which make it difficult to achieve comprehensive coverage and real-time early warning. The embodiments of the present invention are intended to solve the above problems.

[0063] In the embodiment of the present invention, a GPS positioning device is installed in each transport vehicle, and each GPS positioning device is bound to the license plate number of the corresponding transport vehicle. In this way, the operation track information of the transport vehicle can be automatically determined according to the GPS positioning device. The operation track information includes time information, operation tracking track and license plate number, so as to realize accurate and real-time track tracking. It is also necessary to install a loading point camera at the loading and feeding place of various materials. The loading point camera will upload the loading monitoring video in real time. The embodiment of the present invention will automatically analyze the loading monitoring video to determine the license plate number and material information, and add the material information to the operation track information to realize automatic monitoring and recording of material information. In addition, the present invention will also determine the protection characteristics according to the material information, and determine whether the transportation protection is qualified based on the protection characteristics and the loading monitoring video. When it is unqualified, the first warning information is generated to remind the management personnel to intervene in time, thereby strengthening the supervision of transportation protection. In addition, the embodiment of the present invention will determine whether a special vehicle is required based on the material information. When a special vehicle is required, it will determine whether the corresponding vehicle is compliant based on the license plate number. If it is not compliant, a second warning message will be generated to ensure the transportation safety and compliance of special materials. The average daily transportation time of each type of transport vehicle will be determined based on the operation trajectory information, and transport vehicle reminder information will be generated to provide managers with a scientific decision-making basis and optimize transportation efficiency.

[0064] like Figure 2 As shown, as a preferred embodiment of the present invention, the step of analyzing the feeding monitoring video to determine the license plate number and material information specifically includes:

[0065] S201, after detecting the license plate area of ​​the vehicle, use OCR technology to perform character recognition to determine the license plate number;

[0066] S202, performing image preprocessing on the loading material in the loading monitoring video to extract the color features, texture features and shape features of the material;

[0067] S203: input the preprocessed material image into the trained material recognition model, and output the material information.

[0068] In an embodiment of the present invention, after the license plate area is detected, the characters in the license plate are segmented out using methods such as vertical projection and horizontal projection, and then the segmented characters are recognized using OCR (optical character recognition) technology or a deep learning model (such as CNN) to combine into a license plate number. Then, the loaded materials in the feeding monitoring video are image preprocessed to extract the color features, texture features, and shape features of the materials. It is necessary to use deep learning technology (such as convolutional neural network CNN) to train a material recognition model in advance, input material images, and output material types. During the training process, a large number of images of different materials need to be collected as training data sets and annotated. Finally, the preprocessed material image is input into the trained material recognition model, and the material information will be automatically output.

[0069] As a preferred embodiment of the present invention, after detecting the license plate area of ​​the vehicle, using OCR technology to perform character recognition, the step of determining the license plate number specifically includes the following sub-steps:

[0070] Acquire a target vehicle image, convert the target vehicle image into a vehicle grayscale image, and apply Gaussian blur to the vehicle grayscale image to obtain a vehicle blurred image;

[0071] Calculate the median pixel value of the blurred vehicle image, and calculate the upper and lower thresholds of Canny edge detection based on the median pixel value;

[0072] According to the upper and lower thresholds of Canny edge detection, the Canny edge detection algorithm is used to perform multi-scale processing on the image using different Gaussian blur kernel sizes to obtain edge images processed at different scales.

[0073] The edge images processed at different scales are merged to obtain the maximum value of the edge images processed at all scales to obtain a comprehensive edge image;

[0074] The contour detection algorithm is used to search for contours in the integrated edge image to obtain several contours;

[0075] A polygonal approximation method is used for several contours to detect whether the contour is a quadrilateral; if so, a mask process is performed on the contour of the quadrilateral to extract the license plate area;

[0076] Calculate the bounding box of the license plate area to extract the license plate image;

[0077] Performing local thresholding processing on the license plate image to enhance the image, thereby obtaining an enhanced license plate image;

[0078] The enhanced license plate image is recognized by OCR to obtain the recognition result and output the license plate number text.

[0079] In an embodiment of the present invention, the upper and lower thresholds of Canny edge detection are calculated according to the median pixel value to achieve adaptive threshold adjustment, and then dynamically adjust the edge detection parameters, which can better handle the situation of uneven lighting, achieve recognition capabilities under different weather lighting, and enhance the ability to suppress image noise. In addition, combined with multi-scale Canny edge detection, more image details are retained. This is particularly important for detecting complex edges and subtle features, and helps to improve the accuracy of license plate area recognition. Contour detection and polygon approximation methods are used to effectively filter non-license plate areas. By assuming that the license plate is a rectangle, the accuracy of detection is improved and misidentification is reduced. Finally, after the license plate area is extracted, the local thresholding technology is applied to enhance the character contrast, provide a clear character image for OCR, and improve the accuracy of character recognition.

[0080] As a preferred embodiment of the present invention, the step of performing image preprocessing on the loaded material in the loading monitoring video to extract the color features, texture features and shape features of the material specifically includes the following sub-steps:

[0081] Extract the loading material image from the loading monitoring video and convert the loading material image into LAB color space;

[0082] The histograms of the L, A and B channels of the LAB color space are calculated respectively to obtain the histograms of three different channels;

[0083] Flatten the histograms of three different channels and then concatenate them to obtain the color feature vector;

[0084] The loading material image is converted into a material grayscale image, and Gaussian blur is applied to the material grayscale image to obtain a material blurred image;

[0085] The gray level co-occurrence matrix is ​​used to analyze the texture features of the material fuzzy image and obtain the texture features;

[0086] Binarize the fuzzy image of the material to obtain a binary image;

[0087] Get the connected region on the binary image, and calculate the area, perimeter, eccentricity and range of the connected region to obtain the shape feature;

[0088] The color feature vector, texture feature and shape feature are combined to obtain the key visual features.

[0089] In the embodiment of the present invention, the LAB color space is used for color feature extraction, which can better cope with the impact of illumination changes, because the LAB color space can effectively separate brightness and color information and improve the stability of color analysis. Gaussian blur and grayscale co-occurrence matrix (GLCM) are used for texture analysis, which can reduce the interference of noise on texture features and improve the robustness of feature extraction. The comprehensive use of multiple features can improve the accuracy of the material recognition model. Through the complementary characteristics of color, texture and shape, the risk of misidentification caused by a single feature is reduced.

[0090] As a preferred embodiment of the present invention, the step of inputting the preprocessed material image into the trained material recognition model and outputting the material information specifically includes the following sub-steps:

[0091] The key visual features are standardized and normalized in turn to obtain normalized features; standardization eliminates the impact of dimensions between different features so that they can be compared on the same scale, and then normalization is performed to compress the data into a fixed range, thereby maintaining the stability of the data in different algorithms.

[0092] The normalized features are subjected to dimensionality reduction using principal component analysis to obtain key features;

[0093] The normalized features are decomposed using independent component analysis to obtain independent features;

[0094] The key features and independent features are horizontally merged to form a new feature matrix and obtain the combined features;

[0095] Calculate the variance of each feature in the combined feature, and weight each feature in the combined feature according to the variance to obtain the weighted combined feature;

[0096] Construct a material recognition model using a pre-trained random forest model and pre-trained support vectors;

[0097] The weighted combined features are fed into the pre-trained random forest model to obtain the category probability of each material;

[0098] The weighted combined features are fed into the pre-trained support vector machine to obtain the decision function value of the probability of each material category;

[0099] The category probability of each material and the decision function value of each material category probability are weighted and summed to obtain the final classification probability, and the material information is output according to the final classification probability.

[0100] In an embodiment of the present invention, a combination of standardization and normalization is used to help eliminate dimensional differences between different features and improve data consistency. At the same time, normalization can improve the stability of algorithm processing, especially in some models that are sensitive to feature scales. By combining principal component analysis and independent component analysis, not only can the features with the largest variance in the data be extracted, but also statistically independent signals can be found. Such multiple dimensionality reduction methods make the features richer and more diverse, thereby providing more comprehensive information for subsequent models. By combining random forests and support vector machines, the advantages of different models are used for comprehensive prediction. Random forests provide robust probabilistic predictions, while support vector machines provide more accurate boundary judgments through decision functions. The strategy of integrated learning improves the robustness and accuracy of the model.

[0101] like Figure 3 As shown, as a preferred embodiment of the present invention, the step of determining the protection feature according to the material information and judging whether the transportation protection is qualified based on the protection feature and the loading monitoring video specifically includes:

[0102] S301, inputting material information into a material protection feature library, and outputting corresponding protection features, wherein the protection features include protection product features and installation features;

[0103] S302, feature recognition is performed on the feeding monitoring video to determine whether the feeding monitoring video contains protective product features. If not, the transportation protection is directly determined to be unqualified; if contained, the next step is executed;

[0104] S303, determining whether the protective article in the feeding monitoring video is correctly installed according to the installation feature, where the installation feature is used to characterize the position information of the protective article.

[0105] In the embodiment of the present invention, a material protection feature library is established in advance. The material protection feature library includes all material information. Each material information corresponds to a protection feature. The protection feature includes a protective product feature and an installation feature. The protective product feature includes the color, shape, texture, etc. of the protective product. The installation feature indicates the location information of the protective product, such as being laid flat on the roof. When the material information is input into the material protection feature library, the corresponding protection feature will be automatically output, and the feature recognition of the loading monitoring video will be performed to determine whether the protective product feature is included in the loading monitoring video. If it is not included, it means that no protection is performed, and the transportation protection is directly determined to be unqualified; if it is included, it is also necessary to determine whether the protective product in the loading monitoring video is correctly installed based on the installation feature, thereby further strengthening the supervision of transportation protection.

[0106] As a preferred embodiment of the present invention, the step of performing feature recognition on the feeding monitoring video to determine whether the feeding monitoring video contains protective product features specifically includes the following sub-steps:

[0107] Acquire a feeding monitoring image in a feeding monitoring video, and preprocess the feeding monitoring image to obtain a preprocessed feeding monitoring image;

[0108] The ResNet50 network is used to extract the preprocessed feeding monitoring images to obtain high-level features;

[0109] Apply the attention mechanism to weight the high-level features, and then sum the weighted results to obtain the attention feature vector;

[0110] Extracting the color features, texture features and shape features of the preprocessed feeding monitoring image, and fusing the color features, texture features, shape features and attention feature vectors of the preprocessed feeding monitoring image to form a multimodal feature;

[0111] Perform attention calculation on multimodal features to obtain weighted feature vectors;

[0112] Then send the weighted feature vector into the classifier to get the current prediction result;

[0113] The reinforcement learning algorithm is used for training based on the current prediction results. After the training is completed, the feature recognition results are output.

[0114] In the embodiment of the present invention, the information in the image can be described more comprehensively by combining color features, texture features and shape features. Different types of features complement each other and can make up for the shortcomings of a single feature, thereby improving the accuracy of feature recognition of protective products. In addition, in order to ensure better recognition results, the key areas are focused on through the attention mechanism. By focusing on the key areas, the attention mechanism helps to improve the accuracy and robustness of detection, especially in complex backgrounds and with more interference. By introducing reinforcement learning, the strategy can be continuously adjusted according to feedback, the detection process can be optimized, and the adaptability and robustness in different scenarios can be improved.

[0115] like Figure 4 As shown, as a preferred embodiment of the present invention, the step of determining whether a special vehicle is required according to the material information, and when a special vehicle is required, determining whether the corresponding vehicle is compliant based on the license plate number, specifically includes:

[0116] S401, inputting material information into the special vehicle library, when there is special vehicle information output, requiring the use of a special vehicle, the special vehicle information includes a special vehicle category;

[0117] S402, inputting the license plate number into the vehicle information database and outputting the vehicle specific information;

[0118] S403, determining whether the corresponding transport vehicle is compliant based on the vehicle specific information and the special vehicle category.

[0119] In an embodiment of the present invention, a special vehicle library is established in advance, and the special vehicle library contains a number of special vehicle information, and each special vehicle information corresponds to material information. The material information is input into the special vehicle library. When no special vehicle information is output, it means that a special transport vehicle is not required and an ordinary vehicle is sufficient. When special vehicle information is output, it means that a special vehicle is required, and the special vehicle information includes a special vehicle category. Then, in an embodiment of the present invention, the license plate number is automatically input into the vehicle information library, and the specific vehicle information is output, so that it can be determined whether the corresponding transport vehicle is compliant based on the specific vehicle information and the special vehicle category.

[0120] like Figure 5 As shown, as a preferred embodiment of the present invention, the step of determining the average daily transportation time of each type of transportation vehicle based on the operation trajectory information and generating the transportation vehicle reminder information specifically includes:

[0121] S501, determine the average daily transportation time of each vehicle according to the operation trajectory information, classify all transportation vehicles according to the license plate number, and determine the average daily transportation time of each type of transportation vehicle,

[0122] S502, retrieve the daily average transportation limit range of each type of transportation vehicle, and generate transportation vehicle reminder information according to the daily average transportation limit range and the daily average transportation time.

[0123] In an embodiment of the present invention, the average daily transportation time of each vehicle is determined according to the time information in the operation trajectory information, and all transportation vehicles are classified according to the license plate number. In this way, the average daily transportation time of each type of transportation vehicle can be determined, and then the average daily transportation limit range of each type of transportation vehicle is retrieved. The average daily transportation limit range of each type of transportation vehicle is set in advance, and transportation vehicle reminder information is generated according to the average daily transportation limit range and the average daily transportation time. When the average daily transportation time is higher than the average daily transportation limit range, the reminder information is that the load of this type of transportation vehicle is too high; when the average daily transportation time is lower than the average daily transportation limit range, the reminder information is that the load of this type of transportation vehicle is too low.

[0124] like Figure 6 As shown, an embodiment of the present invention further provides a transport vehicle trajectory tracking system for road construction, the system comprising:

[0125] The running track determination module 100 is used to determine the running track information of the transport vehicle based on the GPS positioning device, and the running track information includes time information, running tracking track and license plate number;

[0126] The material information adding module 200 is used to receive the feeding monitoring video uploaded by the feeding point camera, analyze the feeding monitoring video to determine the license plate number and material information, and add the material information to the running track information;

[0127] The transport protection determination module 300 is used to determine the protection characteristics according to the material information, and determine whether the transport protection is qualified based on the protection characteristics and the feeding monitoring video, and generate the first warning information when it is unqualified;

[0128] A special vehicle determination module 400 is used to determine whether a special vehicle is required according to the material information, and when a special vehicle is required, determine whether the corresponding vehicle is compliant based on the license plate number, and generate a second warning information when it is not compliant;

[0129] The transport vehicle reminder module 500 is used to determine the average daily transport time of each type of transport vehicle based on the operation track information and generate transport vehicle reminder information.

[0130] As a preferred embodiment of the present invention, the material information adding module 200 includes:

[0131] The license plate number recognition unit is used to detect the license plate area of ​​the vehicle and use OCR technology to perform character recognition to determine the license plate number;

[0132] A material feature extraction unit is used to perform image preprocessing on the loaded materials in the loading monitoring video to extract the color features, texture features and shape features of the materials;

[0133] The material information output unit is used to input the preprocessed material image into the trained material recognition model and output the material information.

[0134] As a preferred embodiment of the present invention, the transportation protection determination module 300 includes:

[0135] A protection feature determination unit, used to input material information into a material protection feature library and output corresponding protection features, wherein the protection features include protection product features and installation features;

[0136] The protective article recognition unit is used to perform feature recognition on the feeding monitoring video to determine whether the feeding monitoring video contains the protective article feature. If not, it is directly determined that the transportation protection is unqualified; if it is contained, the next step is executed;

[0137] The installation feature determination unit is used to determine whether the protective article in the feeding monitoring video is correctly installed according to the installation features, and the installation features are used to characterize the position information of the protective article.

[0138] As a preferred embodiment of the present invention, the dedicated vehicle determination module 400 includes:

[0139] The special vehicle information unit is used to input material information into the special vehicle library. When there is special vehicle information output, it is required to use a special vehicle. The special vehicle information includes the special vehicle category.

[0140] The vehicle specific information unit is used to input the license plate number into the vehicle information database and output the vehicle specific information;

[0141] The vehicle compliance determination unit is used to determine whether the corresponding transport vehicle is compliant based on the vehicle specific information and the special vehicle category.

[0142] As a preferred embodiment of the present invention, the transport vehicle reminder module 500 includes:

[0143] The daily average transportation time unit is used to determine the daily average transportation time of each vehicle based on the operation trajectory information, classify all transportation vehicles according to the license plate number, and determine the daily average transportation time of each type of transportation vehicle.

[0144] The reminder information generating unit is used to retrieve the daily average transportation limit range of each type of transportation vehicle, and generate transportation vehicle reminder information according to the daily average transportation limit range and the daily average transportation time.

[0145] The above only describes in detail the preferred embodiments of the present invention, which is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0146] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0147] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0148] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A method for tracking the trajectory of a transport vehicle for road construction, characterized in that: The method comprises the following steps: Determine the running track information of the transport vehicle based on the GPS positioning device, wherein the running track information includes time information, running tracking track and license plate number; Receive the feeding monitoring video uploaded by the camera at the feeding point, analyze the feeding monitoring video to determine the license plate number and material information, and add the material information to the running track information; Determine the protection features according to the material information, and determine whether the transportation protection is qualified based on the protection features and the loading monitoring video. If it is unqualified, generate the first warning information; Determine whether a special vehicle is required based on the material information, and if a special vehicle is required, determine whether the corresponding vehicle is compliant based on the license plate number, and if it is not compliant, generate a second warning information; Determine the average daily transportation time of each type of transport vehicle based on the operation trajectory information and generate transport vehicle reminder information; The step of analyzing the feeding monitoring video to determine the license plate number and material information specifically includes: After detecting the license plate area of ​​the vehicle, use OCR technology to perform character recognition and determine the license plate number; Perform image preprocessing on the loading materials in the feeding monitoring video to extract the color, texture and shape features of the materials; Input the preprocessed material image into the trained material recognition model and output the material information; The step of inputting the preprocessed material image into the trained material recognition model to output material information specifically includes the following sub-steps: The key visual features are standardized and normalized in sequence to obtain normalized features; The normalized features are subjected to dimensionality reduction using principal component analysis to obtain key features; The normalized features are decomposed using independent component analysis to obtain independent features; The key features and independent features are horizontally merged to form a new feature matrix and obtain the combined features; Calculate the variance of each feature in the combined feature, and weight each feature in the combined feature according to the variance to obtain the weighted combined feature; Construct a material recognition model using a pre-trained random forest model and pre-trained support vectors; The weighted combined features are fed into the pre-trained random forest model to obtain the category probability of each material; The weighted combined features are fed into the pre-trained support vector machine to obtain the decision function value of the probability of each material category; The category probability of each material and the decision function value of each material category probability are weighted and summed to obtain the final classification probability, and the material information is output according to the final classification probability.

2. The method for tracking the trajectory of a transport vehicle for road construction according to claim 1, characterized in that: After the license plate area of ​​the vehicle is detected, the step of using OCR technology to perform character recognition and determine the license plate number specifically includes the following sub-steps: Acquire a target vehicle image, convert the target vehicle image into a vehicle grayscale image, and apply Gaussian blur to the vehicle grayscale image to obtain a vehicle blurred image; Calculate the median pixel value of the blurred vehicle image, and calculate the upper and lower thresholds of Canny edge detection based on the median pixel value; According to the upper and lower thresholds of Canny edge detection, the Canny edge detection algorithm is used to perform multi-scale processing on the image using different Gaussian blur kernel sizes to obtain edge images processed at different scales. The edge images processed at different scales are merged to obtain the maximum value of the edge images processed at all scales to obtain a comprehensive edge image; The contour detection algorithm is used to search for contours in the integrated edge image to obtain several contours; A polygonal approximation method is used for several contours to detect whether the contour is a quadrilateral; if so, a mask process is performed on the contour of the quadrilateral to extract the license plate area; Calculate the bounding box of the license plate area to extract the license plate image; Performing local thresholding processing on the license plate image to enhance the image, thereby obtaining an enhanced license plate image; The enhanced license plate image is recognized by OCR to obtain the recognition result and output the license plate number text.

3. The method for tracking the trajectory of a transport vehicle for road construction according to claim 2, characterized in that: The step of performing image preprocessing on the loaded material in the loading monitoring video to extract the color features, texture features and shape features of the material specifically includes the following sub-steps: Extract the loading material image from the loading monitoring video and convert the loading material image into LAB color space; The histograms of the L, A and B channels of the LAB color space are calculated respectively to obtain the histograms of three different channels; Flatten the histograms of three different channels and then concatenate them to obtain the color feature vector; The loading material image is converted into a material grayscale image, and Gaussian blur is applied to the material grayscale image to obtain a material blurred image; The gray level co-occurrence matrix is ​​used to analyze the texture features of the material fuzzy image and obtain the texture features; Binarize the fuzzy image of the material to obtain a binary image; Get the connected region on the binary image, and calculate the area, perimeter, eccentricity and range of the connected region to obtain the shape feature; The color feature vector, texture feature and shape feature are combined to obtain the key visual features.

4. The method for tracking the trajectory of a transport vehicle for road construction according to claim 1, characterized in that: The step of determining the protection features according to the material information and judging whether the transportation protection is qualified based on the protection features and the loading monitoring video specifically includes: Input material information into the material protection feature library, and output corresponding protection features, wherein the protection features include protection product features and installation features; Perform feature recognition on the feeding monitoring video to determine whether the feeding monitoring video contains the protective product features. If not, directly determine that the transportation protection is unqualified; if contained, proceed to the next step; Whether the protective equipment in the feeding monitoring video is correctly installed is determined based on the installation features, and the installation features are used to characterize the position information of the protective equipment.

5. The method for tracking the trajectory of a transport vehicle for road construction according to claim 4, characterized in that: The step of performing feature recognition on the feeding monitoring video to determine whether the feeding monitoring video contains the protective product feature specifically includes the following sub-steps: Acquire a feeding monitoring image in a feeding monitoring video, and preprocess the feeding monitoring image to obtain a preprocessed feeding monitoring image; The ResNet50 network is used to extract the preprocessed feeding monitoring images to obtain high-level features; Apply the attention mechanism to weight the high-level features, and then sum the weighted results to obtain the attention feature vector; Extracting the color features, texture features and shape features of the preprocessed feeding monitoring image, and fusing the color features, texture features, shape features and attention feature vectors of the preprocessed feeding monitoring image to form a multimodal feature; Perform attention calculation on multimodal features to obtain weighted feature vectors; Then send the weighted feature vector into the classifier to get the current prediction result; The reinforcement learning algorithm is used for training based on the current prediction results. After the training is completed, the feature recognition results are output.

6. The method for tracking the trajectory of a transport vehicle for road construction according to claim 5, characterized in that: The step of determining whether a special vehicle is required based on the material information, and when a special vehicle is required, determining whether the corresponding vehicle is compliant based on the license plate number specifically includes: Input material information into the special vehicle library. When special vehicle information is output, it is required to use a special vehicle. The special vehicle information includes the special vehicle category. Input the license plate number into the vehicle information database and output the specific vehicle information; Determine whether the corresponding transport vehicle is compliant based on the vehicle-specific information and special vehicle category.

7. The method for tracking the trajectory of a transport vehicle for road construction according to claim 6, characterized in that: The step of determining the average daily transportation time of each type of transportation vehicle based on the operation trajectory information and generating transportation vehicle reminder information specifically includes: Determine the average daily transportation time of each vehicle based on the operation trajectory information, classify all transportation vehicles according to the license plate number, and determine the average daily transportation time of each type of transportation vehicle; Retrieve the average daily transport limit range for each type of transport vehicle, and generate transport vehicle reminder information based on the average daily transport limit range and the average daily transport time.

8. A transport vehicle trajectory tracking system for road construction, characterized in that: The system applies the method for tracking the trajectory of a transport vehicle for road construction as described in any one of claims 1 to 7 above, and the system comprises: An operation track determination module is used to determine the operation track information of the transport vehicle based on the GPS positioning device, wherein the operation track information includes time information, operation tracking track and license plate number; The material information adding module is used to receive the feeding monitoring video uploaded by the feeding point camera, analyze the feeding monitoring video to determine the license plate number and material information, and add the material information to the running track information; A transport protection determination module is used to determine the protection characteristics according to the material information, and determine whether the transport protection is qualified based on the protection characteristics and the feeding monitoring video. If it is unqualified, a first warning information is generated; A special vehicle determination module is used to determine whether a special vehicle is required according to the material information, and when a special vehicle is required, determine whether the corresponding vehicle is compliant based on the license plate number, and generate a second warning information when it is not compliant; The transport vehicle reminder module is used to determine the average daily transport time of each type of transport vehicle based on the operation trajectory information and generate transport vehicle reminder information.

9. The road construction transport vehicle trajectory tracking system according to claim 8, characterized in that: The material information adding module includes: The license plate number recognition unit is used to detect the license plate area of ​​the vehicle and use OCR technology to perform character recognition to determine the license plate number; A material feature extraction unit is used to perform image preprocessing on the loaded materials in the loading monitoring video to extract the color features, texture features and shape features of the materials; A material information output unit, used to input the preprocessed material image into the trained material recognition model and output material information; The transportation protection determination module includes: A protection feature determination unit, used to input material information into a material protection feature library and output corresponding protection features, wherein the protection features include protection product features and installation features; The protective article recognition unit is used to perform feature recognition on the feeding monitoring video to determine whether the feeding monitoring video contains the protective article feature. If not, it is directly determined that the transportation protection is unqualified; if it is contained, the next step is executed; An installation feature determination unit is used to determine whether the protective article in the feeding monitoring video is correctly installed according to the installation feature, and the installation feature is used to characterize the position information of the protective article; The special vehicle determination module comprises: The special vehicle information unit is used to input material information into the special vehicle library. When there is special vehicle information output, it is required to use a special vehicle. The special vehicle information includes the special vehicle category. The vehicle specific information unit is used to input the license plate number into the vehicle information database and output the vehicle specific information; A vehicle compliance determination unit, used to determine whether a corresponding transport vehicle is compliant based on vehicle specific information and special vehicle category; The transport vehicle reminder module includes: The daily average transportation time unit is used to determine the daily average transportation time of each vehicle based on the operation trajectory information, classify all transportation vehicles according to the license plate number, and determine the daily average transportation time of each type of transportation vehicle; The reminder information generating unit is used to retrieve the daily average transportation limit range of each type of transportation vehicle, and generate transportation vehicle reminder information according to the daily average transportation limit range and the daily average transportation time.

Images