Image-based vehicle overload judgment method and device, equipment and storage medium

Abstract

The invention relates to the technical field of transportation, and discloses an image-based vehicle overload judgment method and device, computer equipment and a storage medium. Comprising the following steps: acquiring a target image containing a target vehicle; inputting the target image into an attribute extraction model for attribute extraction to obtain a target attribute of the target vehicle; according to the target attribute, the no-load mass and the actual mass of the target vehicle are determined; and according to the no-load mass and the actual mass, whether the target vehicle is overloaded or not is judged. It can be seen that whether the target vehicle is overloaded or not is judged only according to the target image containing the target vehicle, compared with the mode that whether the target vehicle is overloaded or not is judged through various sensors, the influence of natural factors is smaller, and the purpose of improving the vehicle overload detection reliability can be achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of transportation, and particularly to an image-based vehicle overloading determination method, device, computer device, and storage medium. Background Art

[0002] The development of the economy is inseparable from transportation. As an important part of transportation, road transportation has always received extensive attention.

[0003] In order to ensure the safety during transportation and prevent truck drivers from overloading, vehicle overloading detection systems have been widely used on highway sections in various regions.

[0004] Currently, common vehicle overloading detection systems usually require the cooperation of multiple sensors and cameras to detect vehicle overloading. This leads to the detection accuracy of vehicle overloading detection systems being easily affected by bad weather. As long as one sensor is affected by the weather, such as the sensor being corroded by rain and snow, resulting in errors when the sensor obtains data, the accuracy rate of the vehicle overloading detection system will decrease.

[0005] Therefore, the reliability of existing vehicle overloading detection methods is relatively low. Summary of the Invention

[0006] Embodiments of the present invention provide an image-based vehicle overloading determination method, device, computer device, and storage medium to solve the problem of relatively low reliability of existing vehicle overloading detection systems.

[0007] An image-based vehicle overloading determination method, the method includes:

[0008] Obtain a target image including a target vehicle;

[0009] Input the target image into an attribute extraction model for attribute extraction to obtain target attributes of the target vehicle;

[0010] According to the target attributes, determine the unladen mass and actual mass of the target vehicle;

[0011] According to the unladen mass and the actual mass, determine whether the target vehicle is overloaded.

[0012] In the above method, optionally, the target attributes at least include vehicle attributes and tire attributes;

[0013] The determining the unladen mass and actual mass of the target vehicle according to the target attributes includes:

[0014] According to the vehicle attributes, determine the unladen mass of the target vehicle in a stationary state;

[0015] Determine the actual mass of the target vehicle in the current driving state according to the tire attributes.

[0016] In the above method, optionally, the tire attributes include brand attributes and height attributes;

[0017] The step of determining the actual mass of the target vehicle in the current driving state according to the tire attributes includes:

[0018] Determine the tire spring coefficient corresponding to the brand attribute;

[0019] Calculate the actual mass of the target vehicle in the current driving state according to the tire spring coefficient and the height attribute.

[0020] In the above method, optionally, the target attribute includes the rated load mass of the vehicle;

[0021] The step of determining whether the target vehicle is overloaded according to the unladen mass and the actual mass includes:

[0022] Calculate the mass difference between the unladen mass and the actual mass;

[0023] Determine whether the mass difference is greater than the rated load mass of the vehicle;

[0024] If the mass difference is less than the rated load mass of the vehicle, determine that the target vehicle is not overloaded;

[0025] If the mass difference is greater than the rated load mass of the vehicle, determine that the target vehicle is overloaded.

[0026] In the above method, optionally, the target image includes a thermal imaging image; the thermal imaging image is used to determine whether the vehicle is overloaded with goods in the case of determining that the target vehicle is overloaded.

[0027] In the above method, optionally, the attribute extraction model is trained in the following manner:

[0028] Obtain images of vehicles of different brands as training samples;

[0029] Divide the training samples into a test set and a training set;

[0030] Train the attribute extraction model based on the training samples in the training set and the test set.

[0031] An image-based vehicle overloading determination device includes:

[0032] An image acquisition unit, configured to acquire a target image including a target vehicle;

[0033] An attribute extraction unit, configured to input the target image into an attribute extraction model for attribute extraction to obtain the target attributes of the target vehicle;

[0034] A mass determination unit, configured to determine the unladen mass and the actual mass of the target vehicle according to the target attributes;

[0035] An overloading determination unit, configured to determine whether the target vehicle is overloaded according to the unladen mass and the actual mass.

[0036] For the above device, optionally, the target attributes at least include vehicle attributes and tire attributes; the mass determination unit is configured to:

[0037] Determine the unladen mass of the target vehicle in a stationary state according to the vehicle attributes;

[0038] Determine the actual mass of the target vehicle in the current driving state according to the tire attributes.

[0039] A computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where when the processor executes the computer program, the method for determining vehicle overloading based on an image as described in any one of the above is implemented.

[0040] A computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for determining vehicle overloading based on an image as described in any one of the above is implemented.

[0041] For the above method, device, computer device, and storage medium for determining vehicle overloading based on an image, by obtaining a target image including a target vehicle, extracting the target attributes of the target vehicle from the target image, and then determining the unladen mass and the actual mass of the target vehicle according to the target attributes, and determining whether the target vehicle is overloaded according to the unladen mass and the actual mass. It can be seen that the present invention only needs to determine whether the target vehicle is overloaded according to the target image including the target vehicle, and is less affected by natural factors compared with using various sensors to determine whether the target vehicle is overloaded, and can achieve the purpose of improving the reliability of vehicle overloading detection. Description of the Drawings

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required to be used in the description of the embodiments of the present invention. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0043] Figure 1It is a flowchart of an implementation of the method for determining vehicle overloading based on images in an embodiment of the present invention;

[0044] Figure 2 It is a partial flowchart of an implementation of the method for determining vehicle overloading based on images in an embodiment of the present invention;

[0045] Figure 3 It is a partial flowchart of an implementation of the method for determining vehicle overloading based on images in an embodiment of the present invention;

[0046] Figure 4 It is a partial flowchart of an implementation of the method for determining vehicle overloading based on images in an embodiment of the present invention;

[0047] Figure 5 It is a partial flowchart of an implementation of the method for determining vehicle overloading based on images in an embodiment of the present invention;

[0048] Figure 6 It is a schematic structural diagram of a vehicle overloading determination device based on images in an embodiment of the present invention;

[0049] Figure 7 It is a schematic structural diagram of a computer device in an embodiment of the present invention. Detailed implementation manners

[0050] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0051] It should be understood that when used in the specification and the appended claims of the present invention, the term "comprising" indicates the presence of the described features, wholes, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0052] It should also be understood that the term " / and / or" used in the specification and the appended claims of the present invention refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0053] As used in the specification of the present invention and the appended claims, the term "if" can be interpreted as "when" or "once" or "in response to determining" or "in response to detecting" depending on the context. Similarly, the phrase "if it is determined" or "if [the described condition or event] is detected" can be interpreted as meaning "once it is determined" or "in response to determining" or "once [the described condition or event] is detected" or "in response to detecting [the described condition or event]" depending on the context.

[0054] In addition, in the description of the specification of the present invention and the appended claims, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0055] The reference to "one embodiment" or "some embodiments" or the like described in the specification of the present invention means that a specific feature, structure, or characteristic described in connection with that embodiment is included in one or more embodiments of the present invention. Thus, the statements "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments", etc. that appear in different places in this specification do not necessarily all refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "comprising", "including", "having", and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0056] The present invention discloses an image-based vehicle overloading determination method, device, computer device, and storage medium. By acquiring a target image including a target vehicle, extracting target attributes of the target vehicle from the target image, and then determining the empty load mass and actual mass of the target vehicle according to the target attributes, it is determined whether the target vehicle is overloaded based on the empty load mass and the actual mass. It can be seen that the present invention only needs to determine whether the target vehicle is overloaded according to the target image including the target vehicle. Compared with the need for various sensors to determine whether the target vehicle is overloaded, it is less affected by natural factors and can achieve the purpose of improving the reliability of vehicle overloading detection.

[0057] As Figure 1 shown, an image-based vehicle overloading determination method disclosed in an embodiment of the present invention is applicable to electronic devices with image acquisition capabilities and image processing capabilities, such as mobile phones, computers connected to cameras, servers, and the like.

[0058] S101: Acquire a target image including a target vehicle.

[0059] Among them, in the target image including the target vehicle, at least a complete tire image of the lower half of the target vehicle and an image capable of characterizing the specific vehicle brand and vehicle model of the target vehicle are included.

[0060] Specifically, in this embodiment, cameras can be installed on the roads where vehicles travel to obtain target images of target vehicles. For example, high-speed capture cameras can be established beside highways, or detection cameras can be set up at vehicle checkpoints. The target images containing the target vehicles are obtained through these cameras at specific locations, and then the target images are transmitted to an electronic device with image processing capabilities, such as a data processing server in a high-speed data center, to complete the overloading determination of the target vehicle based on the target images. Accordingly, the purpose of obtaining the target images can be achieved.

[0061] S102: Input the target image into an attribute extraction model for attribute extraction to obtain the target attributes of the target vehicle.

[0062] Input the target image into the trained attribute extraction model to extract features of the target image to obtain the image features of the target image, and then determine the target attributes of the target vehicle according to the extracted image features.

[0063] In specific implementation, the attribute extraction model in this embodiment includes, but is not limited to, a convolutional neural network model (CNN). Input the target image into the convolutional neural network model, and the convolutional neural network model extracts features of the target image to obtain the image features in the target image, and then determines the target attributes of the target vehicle according to the image features.

[0064] S103: Determine the unladen mass and actual mass of the target vehicle according to the target attributes.

[0065] Among them, the unladen mass refers to the mass of the target vehicle when it leaves the factory and there are no items placed on the vehicle. For example, when there is no driver in the target vehicle and no goods are loaded; the actual mass refers to the mass of the target vehicle when it is loaded with goods and a driver. For example, the mass of a truck when it is fully loaded with goods, or the mass of a bus when it is fully loaded with passengers, etc.

[0066] Specifically, the target attributes in this embodiment at least include vehicle attributes and tire attributes. Determine the unladen mass and actual mass of the target vehicle according to the vehicle attributes and tire attributes respectively. The specific steps are as Figure 2 shown:

[0067] S201: Determine the unladen mass of the target vehicle in a stationary state according to the vehicle attributes.

[0068] Establish a correspondence relationship between the vehicle attributes and the unladen mass of the target vehicle. Then, when the vehicle attributes are determined, determine the unladen mass of the target vehicle in a stationary state or the unladen mass of the target vehicle in a uniform motion state according to the mapping relationship between the vehicle attributes and the unladen mass.

[0069] Specifically, in this embodiment, a physical model of vehicle attributes and the unladen mass of the target vehicle can be constructed. By inputting the vehicle attributes into the physical model, the unladen mass of the target vehicle corresponding to the vehicle attributes can be obtained. Among them, the vehicle attributes include but are not limited to vehicle brand, vehicle model, etc. The unladen mass of the target vehicle corresponding to the vehicle brand and vehicle model is queried in the database. For example, the unladen mass of the vehicles with the same vehicle brand is screened out according to the vehicle brand, and then the unladen mass of the vehicles with the same vehicle brand is secondarily screened according to the vehicle model to obtain the unladen mass of the vehicle with the specified vehicle brand and vehicle model. Thus, the unladen mass of the target vehicle in the stationary state can be determined. It can be seen that in this embodiment, two identifiers, namely the vehicle brand and the vehicle model, are used successively to determine the unladen mass of the target vehicle step by step. Compared with screening out the unladen mass of the target vehicle by traversing the unladen mass of all vehicles through only one identifier, such as the unique number of the vehicle, the screening efficiency can be effectively improved.

[0070] Among them, the vehicle brand refers to the brand of the automobile manufacturer to which the target vehicle belongs, and the vehicle model refers to the model of the specific vehicle under the automobile manufacturer. For example, FAW Jiefang, where FAW refers to the vehicle brand and Jiefang refers to the vehicle model. Thus, the specific model of the vehicle can be specified, and the unladen mass of the target vehicle in the unladen state can be determined.

[0071] It should be understood that there are significant differences in the quality of vehicles from different vehicle brand manufacturers, and there are also significant differences in the quality of different vehicle models of the same manufacturer. Therefore, by determining the unladen mass of the target vehicle in the stationary state through the target attributes, more accurate unladen mass data can be obtained, which is beneficial to improving the accuracy of vehicle overloading determination.

[0072] S202: Determine the actual mass of the target vehicle in the current driving state according to the tire attributes.

[0073] Through the tire attributes, determine the deformation degree of the tires of the target vehicle during driving or in the stationary state, and then determine the actual mass of the target vehicle in the current driving state according to the deformation degree of the tires.

[0074] Specifically, the tire attributes in this embodiment at least include brand attributes and height attributes. The actual mass of the target vehicle in the current driving state is determined according to the brand attributes and height attributes through the following steps, as shown below:

[0075] S301: Determine the tire spring coefficient corresponding to the brand attributes.

[0076] Among them, the brand attributes at least include the tire brand and tire model of the tires installed on the wheels of the target vehicle. It should be understood that the tire spring coefficients of different brand tires are different, and the tire spring coefficients of different tire models of the same brand are also different.

[0077] Specifically, in this embodiment, a mapping relationship between different tire brands and different tire models and the tire spring coefficient of the tire can be established. Then, when determining the brand attributes of the tire, according to the tire brand and tire model included in the brand attributes, the corresponding tire spring coefficient can be determined. It should be understood that for tires of different tire brands, there may be the same tire model that is similar or the same. Then, when determining the tire spring coefficient corresponding to the brand attributes, the screening range can be determined first according to the tire brand, that is, the tire spring coefficient corresponding to the tires of the specified tire brand is screened out from the tire spring coefficient corresponding to the tire, and then according to the tire model, the tire spring coefficient corresponding to the tire of the specified tire model is screened out from the tire spring coefficient corresponding to the tires of the specified tire brand. Based on this, the tire spring coefficient corresponding to the brand attributes can be determined.

[0078] S302: Calculate the actual mass of the target vehicle in the current driving state according to the tire spring coefficient and the height attribute.

[0079] Among them, the height attribute includes the height of the tire deformation of the target vehicle during driving or in a stationary state. For example, taking the tire radius as 50 cm as an example, after the vehicle loads a large amount of goods, the tire deforms, and the height from the tire center to the ground becomes 45 cm. Then, the height attribute of the tire deformation is 5 cm.

[0080] The support force between the target vehicle and the ground is calculated through the tire spring coefficient and the height attribute, and then this support force is the actual mass of the target vehicle in the current driving state.

[0081] Specifically, in this embodiment, the tire spring coefficient and the height attribute can be input into the support force calculation formula to obtain the support force given by the ground to the target vehicle. Then, according to the equal relationship between the support force of the ground to the target vehicle and the gravity of the vehicle, the actual mass of the target vehicle in the current driving state can be calculated.

[0082] Among them, the actual mass calculation formula is as follows:

[0083] F = KH

[0084] Among them, K represents the tire spring coefficient, H represents the height attribute, and F represents the support force between the target vehicle and the ground.

[0085] W = mg

[0086] Among them, m represents the mass of the target vehicle, g represents the acceleration due to gravity (m / s 2 ), and W represents the gravity of the target vehicle.

[0087] According to the equality relationship between the supporting force and the gravity, the following actual mass calculation formula can be obtained:

[0088]

[0089] Among them, g represents the acceleration due to gravity (m / s 2 ), K represents the tire spring coefficient, H represents the height attribute, and m represents the actual mass of the target vehicle.

[0090] That is to say, in the case of determining the tire spring coefficient and the height attribute, the supporting force calculation formula and the gravity calculation formula can be combined to calculate the actual mass of the target vehicle, and the subsequent steps of vehicle overload detection can be executed according to the obtained actual mass.

[0091] It should be noted that in this embodiment, not only can the actual mass of the target vehicle be calculated according to the height attribute of the tire deformation after the target vehicle is loaded with goods, but also the actual mass of the target vehicle can be calculated according to the contact area and the contact length between the tire and the ground, etc. In this embodiment, the method for calculating the actual mass of the target vehicle is not specifically limited.

[0092] S104: Determine whether the target vehicle is overloaded according to the unladen mass and the actual mass.

[0093] According to the unladen mass and the actual mass, determine the actual load of the target vehicle, and determine whether the target vehicle is overloaded according to the actual load. Among them, the actual load refers to the mass of the goods actually loaded on the target vehicle, or the remaining mass obtained by subtracting the vehicle curb mass from the total mass of the target vehicle.

[0094] Specifically, the target attribute in this embodiment further includes the vehicle rated mass. Determine whether the target vehicle is overloaded according to the vehicle rated mass, the unladen mass and the actual mass, as follows:

[0095] 401: Calculate the mass difference between the unladen mass and the actual mass.

[0096] Specifically, in this embodiment, the mass difference can be obtained by inputting the unladen mass and the actual mass into the difference calculation formula for difference calculation.

[0097] Among them, the difference calculation formula can be as follows:

[0098] F = |m - W|

[0099] Among them, m represents the actual mass, W represents the unladen mass, and F represents the mass difference between the unladen mass and the actual mass.

[0100] Calculate the difference between the unladen mass and the actual mass, then take the absolute value of the difference calculation result, and determine that the obtained mass difference is positive, so as to facilitate the determination of vehicle overloading in subsequent steps.

[0101] 402: Determine whether the mass difference is greater than the vehicle's rated load mass.

[0102] If the mass difference is greater than the vehicle's rated load mass, determine that the target vehicle is not overloaded; if the mass difference is less than the vehicle's rated load mass, determine that the target vehicle is overloaded.

[0103] Among them, the vehicle's rated load mass refers to the maximum load mass of the target vehicle. When the mass of the goods loaded on the target vehicle is greater than the vehicle's rated load mass, it is determined that the target vehicle is overloaded. When the mass of the goods loaded on the target vehicle is less than or equal to the vehicle's rated load mass, it is determined that the target vehicle is not overloaded.

[0104] 403: Determine that the target vehicle is not overloaded.

[0105] Specifically, if the mass difference is less than the vehicle's rated load mass, it means that the mass of the target vehicle after loading goods is less than the vehicle's rated load mass specified for the target vehicle. That is to say, the mass of the goods loaded on the target vehicle is within the specified mass range, and it is determined that the target vehicle is not overloaded.

[0106] 404: Determine that the target vehicle is overloaded.

[0107] Specifically, if the mass difference is greater than the vehicle's rated load mass, it means that the mass of the target vehicle after loading goods has exceeded the vehicle's rated load mass specified for the target vehicle, and it is determined that the target vehicle is overloaded.

[0108] In one implementation, the target image includes a thermal imaging image. Among them, the thermal imaging image is used to determine whether the target vehicle is overloaded with goods or overloaded with passengers when it is determined that the target vehicle is overloaded. That is to say, when the target image is obtained, not only can the target image including the target vehicle be obtained, but also the thermal imaging image of the target vehicle can be obtained synchronously. Then, when it is determined that the target vehicle is overloaded, it is determined whether the target vehicle is overloaded with goods or overloaded with passengers according to the thermal imaging image.

[0109] It should be understood that when a vehicle is loaded with too much cargo, there will be an overloading phenomenon. Similarly, when a vehicle carries too many passengers, there will also be an overloading of passengers. Through the thermal imaging image of the target vehicle, it is possible to assist in determining whether the target vehicle is overloaded with cargo or passengers based on the shape of the thermal imaging image. For example, when a large number of human-shaped images are shown in the thermal imaging image and the vehicle is overloaded, it is determined that the target vehicle is overloaded with passengers. Another example is when no human-shaped images are shown in the thermal imaging image and the vehicle is overloaded, it is determined that the target vehicle is overloaded with cargo.

[0110] Specifically, in this embodiment, a thermal imaging night vision device can be synchronously installed at the high-speed capture camera established beside the highway, or at the detection camera established at the vehicle inspection station, so that when the target image is obtained, the thermal imaging image can be synchronously obtained. Then, based on the target image and the thermal imaging image, it is determined whether the target vehicle is overloaded, and when it is determined that the target vehicle is overloaded, it is determined whether the target vehicle is overloaded with passengers or cargo. In addition, in this embodiment, it can also be determined whether the target vehicle is a freight vehicle used for passengers through the thermal imaging image. For example, through the target attributes of the target vehicle, such as the vehicle brand and vehicle model, the transportation type of the target vehicle, such as passenger transportation, cargo transportation, etc., is determined. When the transportation type of the target vehicle is cargo transportation, and a large number of passengers are found in the target vehicle through the thermal imaging image, it can be determined that the target vehicle is a freight vehicle used for passengers. It should be understood that different technical purposes achieved according to the technical solutions of the present invention are also within the protection scope of the present invention.

[0111] In one implementation, the attribute extraction model in this embodiment is trained in the following manner, as Figure 5 shown:

[0112] S501: Obtain images of vehicles of different brands as training samples.

[0113] According to actual needs, select images of vehicles of different brands as training samples. For example, when it is necessary to detect overloading of freight vehicles, only select images of vehicles of different freight vehicle brands as training samples.

[0114] Specifically, in this embodiment, sample images of vehicles of different brands can be selected according to actual needs, and then each image is added with a sample label to obtain training samples. Among them, the sample label includes but is not limited to the brand and model of the vehicle in the image, the brand and model of the tires, the spring coefficient, and the height of the tire deformation after the vehicle is loaded with cargo, etc. Accordingly, after adding the sample label to the sample image, the training samples required for training are obtained. Among them, adding the sample label to the sample image in this embodiment includes but is not limited to manual methods.

[0115] In addition, in the sample images of this embodiment, for vehicles of the same vehicle brand and vehicle type, images at least with different shooting angles and different shooting distances are acquired as sample images, so as to improve the accuracy and efficiency of attribute extraction of the trained attribute extraction model.

[0116] S502: Divide the training samples into a test set and a training set.

[0117] Specifically, in this embodiment, some training samples can be randomly selected as the test set, and the remaining training samples are used as the training set. For example, 70% of the training samples are randomly selected as the training set, and the remaining 30% are used as the test set. Another example is that 75% of the training samples are randomly selected as the training set, and the remaining 25% are used as the test set. In this embodiment, the ratio of the test set to the training set is not limited.

[0118] In addition, in this embodiment, several images can also be separately selected from the images of different vehicle brands and different vehicle types for training. In this embodiment, the selection methods of the test set and the training set are not limited.

[0119] S503: Based on the training samples in the training set and the test set, train to obtain an attribute extraction model.

[0120] Input the training samples in the training set into the model to be trained for training. Then, after completing one round of training, input the training samples in the test set into the model to be trained for testing, so as to determine whether the model to be trained has been fully trained. When it is determined that the model to be trained has been fully trained, train to obtain an attribute extraction model.

[0121] Specifically, in this embodiment, by separately inputting the training samples in the test set into the model to be trained, the attribute extraction results of each training sample in the test set are obtained. According to the accuracy rate of the attribute extraction results, it is determined whether the model to be trained has been fully trained. For example, taking a preset accuracy rate of 90% as an example, if the consistency between the attribute extraction results of the training samples in the test set and the sample labels reaches 90% or more, that is, the proportion of the attribute extraction results of the training samples in the test set being exactly the same as the sample labels reaches 90% or more, it is determined that the model to be trained has been fully trained, and an attribute extraction model is trained. If the consistency between the attribute extraction results of the training samples in the test set and the sample labels is less than 90%, then the training samples in the training set are again separately input into the model to be trained for iterative training. After completing the second iterative training, the training effect of the model to be trained is again judged through the training samples in the test set, and so on, until the consistency between the attribute extraction results of the training samples in the test set and the sample labels reaches 90% or more, and it is determined that the model to be trained has been fully trained, and the required attribute extraction model is trained.

[0122] In addition, in this embodiment, not only the unladen mass and the actual mass of the target vehicle are used to determine whether the target vehicle is overloaded, but also the total mass and the actual mass of the target vehicle can be used to determine whether the target vehicle is overloaded. Specifically, it can be determined whether the target vehicle is overloaded by judging whether the total mass of the target vehicle is less than the actual mass. If the total mass of the target vehicle is less than or equal to the actual mass, it is determined that the target vehicle is overloaded. If the total mass of the target vehicle is greater than the actual mass, it is determined that the target vehicle is not overloaded.

[0123] In summary, a method for determining vehicle overload based on an image disclosed in the present invention obtains a target image including a target vehicle, extracts target attributes of the target vehicle from the target image, and then determines the unladen mass and the actual mass of the target vehicle according to the target attributes, and determines whether the target vehicle is overloaded according to the unladen mass and the actual mass. It can be seen that the present invention only needs to determine whether the target vehicle is overloaded according to the target image including the target vehicle. Compared with the need for various sensors to determine whether the target vehicle is overloaded, it is less affected by natural factors and can achieve the purpose of improving the reliability of vehicle overload detection.

[0124] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The order of execution of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0125] As Figure 6 shown, it is a structural schematic diagram of a device for determining vehicle overload based on an image disclosed in an embodiment of the present invention. The device is applicable to electronic devices with image acquisition capabilities and image processing capabilities, such as mobile phones, computers connected to cameras, servers, and so on.

[0126] In specific implementation, the device in this embodiment may specifically include the following units:

[0127] An image acquisition unit 601, configured to acquire a target image including a target vehicle;

[0128] Among them, in the target image including the target vehicle, at least a complete tire image of the lower half of the target vehicle and an image capable of characterizing the specific vehicle brand and vehicle model of the target vehicle are included.

[0129] Specifically, in this embodiment, a camera may be installed on the road where the vehicle travels to acquire the target image of the target vehicle, such as establishing a high-speed capture camera beside a highway or a detection camera established at a vehicle checkpoint. The target image including the target vehicle is acquired through the cameras at these specific positions, and then the target image is transmitted to an electronic device with image processing capabilities, such as a data processing server in a high-speed data center, to complete the overload determination of the target vehicle according to the target image. Accordingly, the purpose of acquiring the target image can be achieved.

[0130] An attribute extraction unit 602 is configured to input a target image into an attribute extraction model for attribute extraction to obtain target attributes of the target vehicle.

[0131] Input the target image into a trained attribute extraction model to perform feature extraction on the target image to obtain image features of the target image, and then determine the target attributes of the target vehicle according to the extracted image features.

[0132] In a specific implementation, the attribute extraction model in this embodiment includes, but is not limited to, a convolutional neural network model (CNN). Input the target image into the convolutional neural network model, perform feature extraction on the target image through the convolutional neural network model to obtain image features in the target image, and further determine the target attributes of the target vehicle according to the image features.

[0133] A quality determination unit 603 is configured to determine the unladen mass and the actual mass of the target vehicle according to the target attributes.

[0134] Among them, the unladen mass refers to the mass of the target vehicle when no items are placed on the vehicle at the time of leaving the factory, such as the mass when there is no driver in the target vehicle and no cargo is loaded; the actual mass refers to the mass of the target vehicle when loading cargo and having a driver on board, such as the mass when a truck is fully loaded with cargo, or the mass when a bus is fully loaded with passengers, etc.

[0135] The target attributes in this embodiment at least include vehicle attributes and tire attributes, and the unladen mass and the actual mass of the target vehicle are determined according to the vehicle attributes and the tire attributes respectively.

[0136] Specifically, determine the unladen mass of the target vehicle in a stationary state according to the vehicle attributes. Establish a correspondence relationship between the vehicle attributes and the unladen mass of the target vehicle, and then, when the vehicle attributes are determined, determine the unladen mass of the target vehicle in a stationary state or the unladen mass of the target vehicle in a uniform motion state according to the mapping relationship between the vehicle attributes and the unladen mass.

[0137] Specifically, in this embodiment, a physical model of vehicle attributes and the unladen mass of the target vehicle can be constructed. By inputting the vehicle attributes into the physical model, the unladen mass of the target vehicle corresponding to the vehicle attributes can be obtained. Among them, vehicle attributes include but are not limited to vehicle brand, vehicle model, etc. The unladen mass of the target vehicle corresponding to the vehicle brand and vehicle model is queried in the database. For example, the unladen masses of vehicles with the same vehicle brand are screened out according to the vehicle brand, and then the unladen masses of vehicles with the same vehicle brand are secondarily screened according to the vehicle model to obtain the unladen mass of the vehicle with the specified vehicle brand and vehicle model. Thus, the unladen mass of the target vehicle in the stationary state can be determined. It can be seen that in this embodiment, two identifiers, namely the vehicle brand and the vehicle model, are used successively to determine the unladen mass of the target vehicle step by step. Compared with screening the unladen mass of the target vehicle by traversing the unladen masses of all vehicles through only one identifier, such as the unique number of the vehicle, the screening efficiency can be effectively improved.

[0138] Among them, the vehicle brand refers to the brand of the automobile manufacturer to which the target vehicle belongs, and the vehicle model refers to the model of a specific vehicle under the automobile manufacturer. For example, FAW Jiefang, where FAW refers to the vehicle brand and Jiefang refers to the vehicle model. Thus, the specific model of the vehicle can be specified, and the unladen mass of the target vehicle in the unladen state can be determined.

[0139] It should be understood that there are significant differences in the quality of vehicles of different vehicle brand manufacturers, and there are also significant differences in the quality of different vehicle models of the same manufacturer. Therefore, the unladen mass of the target vehicle determined by the target attributes can obtain more accurate unladen mass data, which is beneficial to improving the accuracy of vehicle overloading determination.

[0140] Specifically, according to the tire attributes, the actual mass of the target vehicle in the current driving state is determined. Through the tire attributes, the deformation degree of the tires of the target vehicle during driving or in a stationary state is determined, and then the actual mass of the target vehicle in the current driving state is determined according to the deformation degree of the tires.

[0141] The overloading determination unit 604 is used to determine whether the target vehicle is overloaded according to the unladen mass and the actual mass.

[0142] According to the unladen mass and the actual mass, the actual load of the target vehicle is determined, and according to the actual load, it is determined whether the target vehicle is overloaded. Among them, the actual load refers to the mass of the goods actually loaded on the target vehicle, or the remaining mass obtained by subtracting the vehicle curb mass from the total mass of the target vehicle.

[0143] In summary, a vehicle overloading determination device based on images disclosed by the present invention obtains a target image including a target vehicle, extracts target attributes of the target vehicle from the target image, and then determines the unladen mass and the actual mass of the target vehicle according to the target attributes, and determines whether the target vehicle is overloaded according to the unladen mass and the actual mass. It can be seen that the present invention only needs to determine whether the target vehicle is overloaded according to the target image including the target vehicle. Compared with using various sensors to determine whether the target vehicle is overloaded, it is less affected by natural factors and can achieve the purpose of improving the reliability of vehicle overloading detection.

[0144] In one implementation, the target attributes at least include vehicle attributes and tire attributes; the mass determination unit 603 can be used for:

[0145] Determine the unladen mass of the target vehicle in a stationary state according to the vehicle attributes;

[0146] Determine the actual mass of the target vehicle in the current driving state according to the tire attributes.

[0147] In one implementation, the tire attributes include brand attributes and height attributes;

[0148] The mass determination unit 603 can be used for:

[0149] Determine the tire spring coefficient corresponding to the brand attribute;

[0150] Calculate the actual mass of the target vehicle in the current driving state according to the tire spring coefficient and the height attribute.

[0151] In one implementation, the target attributes include the vehicle's permitted load mass;

[0152] The overloading determination unit 604 can be used for:

[0153] Calculate the mass difference between the unladen mass and the actual mass;

[0154] Judge whether the mass difference is greater than the vehicle's permitted load mass;

[0155] If the mass difference is less than or equal to the vehicle's permitted load mass, determine that the target vehicle is not overloaded;

[0156] If the mass difference is greater than the vehicle's permitted load mass, determine that the target vehicle is overloaded.

[0157] In one implementation, the target image includes a thermal imaging image; the thermal imaging image is used to determine whether the vehicle is overloaded with goods in the case of determining that the target vehicle is overloaded.

[0158] In one implementation, the attribute extraction model is trained in the following manner:

[0159] Obtain images of vehicles of different brands as training samples;

[0160] Divide the training samples into a test set and a training set;

[0161] Based on the training samples in the training set and the test set, train an attribute extraction model.

[0162] For the specific limitations of the image-based vehicle overload determination device, reference can be made to the relevant limitations of the image-based vehicle overload determination method in the above text, which will not be elaborated here. Each module in the above image-based vehicle overload determination device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to the above modules.

[0163] In one implementation manner, an embodiment of the present application discloses a computer device, which may be a server, and its internal structure diagram may be as Figure 7 shown. The computer device includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements an image-based vehicle overload determination method.

[0164] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented:

[0165] Obtain a target image including a target vehicle;

[0166] Input the target image into the attribute extraction model for attribute extraction to obtain the target attributes of the target vehicle;

[0167] According to the target attributes, determine the unladen mass and the actual mass of the target vehicle;

[0168] According to the unladen mass and the actual mass, determine whether the target vehicle is overloaded.

[0169] In summary, a computer device disclosed by the present invention obtains a target image including a target vehicle, extracts target attributes of the target vehicle from the target image, and then determines the unladen mass and actual mass of the target vehicle according to the target attributes, and determines whether the target vehicle is overloaded according to the unladen mass and the actual mass. It can be seen that the present invention only needs to determine whether the target vehicle is overloaded according to the target image including the target vehicle. Compared with using various sensors to determine whether the target vehicle is overloaded, it is less affected by natural factors and can achieve the purpose of improving the reliability of vehicle overload detection.

[0170] In one implementation, the target attributes at least include vehicle attributes and tire attributes;

[0171] Determining the unladen mass and actual mass of the target vehicle according to the target attributes includes:

[0172] Determining the unladen mass of the target vehicle in a stationary state according to the vehicle attributes;

[0173] Determining the actual mass of the target vehicle in the current driving state according to the tire attributes.

[0174] In one implementation, the tire attributes include brand attributes and height attributes;

[0175] Determining the actual mass of the target vehicle in the current driving state according to the tire attributes includes:

[0176] Determining the tire spring coefficient corresponding to the brand attributes;

[0177] Calculating the actual mass of the target vehicle in the current driving state according to the tire spring coefficient and the height attributes.

[0178] In one implementation, the target attributes include the vehicle's permitted load mass;

[0179] Determining whether the target vehicle is overloaded according to the unladen mass and the actual mass includes:

[0180] Calculating the mass difference between the unladen mass and the actual mass;

[0181] Judging whether the mass difference is greater than the vehicle's permitted load mass;

[0182] If the mass difference is less than or equal to the vehicle's permitted load mass, it is determined that the target vehicle is not overloaded;

[0183] If the mass difference is greater than the vehicle's permitted load mass, it is determined that the target vehicle is overloaded.

[0184] In one implementation, the target image includes a thermal imaging image; the thermal imaging image is used to determine whether the vehicle is overloaded with goods in the case of determining that the target vehicle is overloaded.

[0185] In one implementation, the attribute extraction model is trained as follows:

[0186] Obtain images of vehicles of different brands as training samples;

[0187] Divide the training samples into a test set and a training set;

[0188] Based on the training samples in the training set and the test set, train the attribute extraction model.

[0189] In one implementation, an embodiment of the present application discloses a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by a processor in a computer device, the computer device can execute each step of any embodiment of a vehicle overloading determination method based on images disclosed in the present invention. The computer-readable storage medium can be non-volatile or volatile.

[0190] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0191] Obtain a target image including a target vehicle;

[0192] Input the target image into the attribute extraction model for attribute extraction to obtain the target attributes of the target vehicle;

[0193] According to the target attributes, determine the unladen mass and the actual mass of the target vehicle;

[0194] According to the unladen mass and the actual mass, determine whether the target vehicle is overloaded.

[0195] In summary, for a computer-readable storage medium disclosed in the present invention, when the instructions in the computer-readable storage medium are executed by a processor in a computer device, by obtaining a target image including a target vehicle, extracting the target attributes of the target vehicle from the target image, and then according to the target attributes, determining the unladen mass and the actual mass of the target vehicle, and determining whether the target vehicle is overloaded according to the unladen mass and the actual mass. It can be seen that the present invention only needs to determine whether the target vehicle is overloaded according to the target image including the target vehicle. Compared with using various sensors to determine whether the target vehicle is overloaded, it is less affected by natural factors and can achieve the purpose of improving the reliability of vehicle overloading detection.

[0196] In one implementation, the target attributes at least include vehicle attributes and tire attributes;

[0197] According to the target attributes, determining the unladen mass and the actual mass of the target vehicle includes:

[0198] Determine the unladen mass of the target vehicle in a stationary state according to the vehicle attributes;

[0199] Determine the actual mass of the target vehicle in the current driving state according to the tire attributes.

[0200] In one implementation, the tire attributes include brand attributes and height attributes;

[0201] Determine the actual mass of the target vehicle in the current driving state according to the tire attributes, including:

[0202] Determine the tire spring coefficient corresponding to the brand attributes;

[0203] Calculate the actual mass of the target vehicle in the current driving state according to the tire spring coefficient and height attributes.

[0204] In one implementation, the target attributes include the vehicle's rated load mass;

[0205] Determine whether the target vehicle is overloaded according to the unladen mass and the actual mass, including:

[0206] Calculate the mass difference between the unladen mass and the actual mass;

[0207] Judge whether the mass difference is greater than the vehicle's rated load mass;

[0208] If the mass difference is less than or equal to the vehicle's rated load mass, determine that the target vehicle is not overloaded;

[0209] If the mass difference is greater than the vehicle's rated load mass, determine that the target vehicle is overloaded.

[0210] In one implementation, the target image includes a thermal imaging image; the thermal imaging image is used to determine whether the vehicle is overloaded with goods in the case of determining that the target vehicle is overloaded.

[0211] In one implementation, the attribute extraction model is trained in the following way:

[0212] Obtain images of vehicles of different brands as training samples;

[0213] Divide the training samples into a test set and a training set;

[0214] Train the attribute extraction model based on the training samples in the training set and the test set.

[0215] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. This computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in this application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of 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 (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0216] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0217] The above embodiments are only used to illustrate the technical solutions of the present invention, not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An image-based method for determining vehicle overloading, characterized in that, The method includes: Obtaining a target image including a target vehicle; Inputting the target image into an attribute extraction model for attribute extraction to obtain target attributes of the target vehicle; Determining the unladen mass and the actual mass of the target vehicle according to the target attributes; Determining whether the target vehicle is overloaded according to the unladen mass and the actual mass; 2. The method for determining vehicle overloading based on images according to claim 1, wherein The target attributes at least include vehicle attributes and tire attributes; The determining the unladen mass and the actual mass of the target vehicle according to the target attributes includes: Determining the unladen mass of the target vehicle in a stationary state according to the vehicle attributes; Determining the actual mass of the target vehicle in the current driving state according to the tire attributes; 3. The method for determining vehicle overloading based on images according to claim 2, wherein The tire attributes include brand attributes and height attributes; The determining the actual mass of the target vehicle in the current driving state according to the tire attributes includes: Determining the tire spring coefficient corresponding to the brand attributes; Calculating the actual mass of the target vehicle in the current driving state according to the tire spring coefficient and the height attributes; 4. The method for determining overloading of a vehicle based on an image according to claim 1, characterized in that, The target attributes include the vehicle's rated load mass; The determining whether the target vehicle is overloaded according to the unladen mass and the actual mass includes: Calculating the mass difference between the unladen mass and the actual mass; Judging whether the mass difference is greater than the vehicle's rated load mass; If the mass difference is less than or equal to the vehicle's rated load mass, determining that the target vehicle is not overloaded; If the mass difference is greater than the vehicle's rated load mass, determining that the target vehicle is overloaded; 5. The method for determining vehicle overloading based on images according to claim 1, wherein The target image includes a thermal imaging image; the thermal imaging image is used to determine whether the vehicle is overloaded with goods in the case of determining that the target vehicle is overloaded; 6. The method for determining vehicle overloading based on images according to claim 1, wherein, The attribute extraction model is trained in the following manner: Obtaining images of vehicles of different brands as training samples; Dividing the training samples into a test set and a training set; Training the attribute extraction model based on the training samples in the training set and the test set; 7. An image-based vehicle overloading determination device, characterized in that, Including: An image acquisition unit for obtaining a target image including a target vehicle; An attribute extraction unit for inputting the target image into an attribute extraction model for attribute extraction to obtain target attributes of the target vehicle; A mass determination unit for determining the unladen mass and the actual mass of the target vehicle according to the target attributes; An overload determination unit for determining whether the target vehicle is overloaded according to the unladen mass and the actual mass; 8. The image-based vehicle overloading determination device according to claim 7, wherein, The target attributes at least include vehicle attributes and tire attributes; the mass determination unit is used for: Determining the unladen mass of the target vehicle in a stationary state according to the vehicle attributes; Determining the actual mass of the target vehicle in the current driving state according to the tire attributes; 9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the image-based vehicle overload determination method according to any one of claims 1 to 6; 10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the image-based vehicle overload determination method according to any one of claims 1 to 6.

Images