A driver portrait pre-classification method, image classification device and storage medium

Abstract

The present application relates to the technical field of image clustering, and discloses a driver portrait pre-classification method, an image classification device, and a storage medium. The method comprises: obtaining a plurality of driver images; wherein the driver images are acquired by monitoring equipment, and the driver images include a driver portrait and vehicle information of the corresponding vehicle; determining the monitoring equipment corresponding to the driving path of the same vehicle according to the vehicle information; determining the association relationship between the monitoring equipment according to the monitoring equipment corresponding to the driving path of the vehicle; and pre-classifying the plurality of driver images according to the association relationship between the monitoring equipment. Through the above method, the association relationship between the monitoring equipment is determined according to the driving path of the vehicle, and then the driver images are pre-classified according to the association relationship between the monitoring equipment, which can enhance the correlation between the driver images in the driver image set after pre-classification and improve the effect of driver portrait clustering.

Description

Technical Field

[0001] The present application relates to the technical field of image clustering, and in particular to a driver portrait pre-classification method, an image classification device and a storage medium. Background Art

[0002] The number of motor vehicles in cities is growing rapidly every year, and driving has become a common way for people to travel. In order to improve the level of urban security and management, it is very important to monitor motor vehicles and drivers. With the full coverage of intelligent monitoring equipment at urban intersections, a large amount of vehicle passing data will be accumulated in a few hours. Through distributed computing, portrait images in the massive vehicle passing data are clustered, and identities are classified into clusters based on people, so that drivers can be monitored.

[0003] In order to prevent data skew in distributed computing, data needs to be pre-bucketed to balance the load on computing nodes. The geographical locations of monitoring devices are relatively far apart, and the pre-bucketing method that directly divides the massive amount of captured portrait image data evenly will result in weak correlation between the data after bucketing, which will lead to poor results in clustering. Summary of the invention

[0004] The main technical problem solved by the present application is to provide a driver portrait pre-classification method, an image classification device and a storage medium, which can solve the problem that the existing driver portrait data is classified in an average division manner, resulting in weak correlation between the classified data and poor clustering effect.

[0005] In order to solve the above technical problems, a technical solution adopted by the present application is: to provide a driver portrait pre-classification method, the method comprising:

[0006] Acquire multiple driver images; wherein the driver images are collected by monitoring equipment, and the driver images include a portrait of the driver and vehicle information of the corresponding vehicle; determine the monitoring equipment corresponding to the driving path of the same vehicle based on the vehicle information; determine the association relationship between the monitoring equipment based on the monitoring equipment corresponding to the driving path of the vehicle; pre-classify the multiple driver images based on the association relationship between the monitoring equipment.

[0007] Optionally, in the above method, determining the monitoring device corresponding to the driving path of the same vehicle according to the vehicle information includes:

[0008] According to the vehicle information, multiple driver images are grouped to obtain a driver image group corresponding to the target vehicle; according to the driver image group corresponding to the target vehicle, the driving path of the target vehicle is determined; according to the driving path of the target vehicle, the monitoring device corresponding to the driving path of the target vehicle is determined.

[0009] Optionally, in the above method, the vehicle information includes license plate information; and grouping the plurality of driver images according to the vehicle information to obtain a driver image group corresponding to the target vehicle includes:

[0010] Based on the license plate information corresponding to the multiple driver images, the multiple driver images are grouped to obtain a first driver image group corresponding to the target vehicle; based on the acquisition time of the driver images, the driver images in the first driver image group are sorted to obtain a second driver image group; the second driver image group is deduplicated to obtain a driver image group corresponding to the target vehicle.

[0011] Optionally, in the above method, determining the driving path of the target vehicle according to the driver image group corresponding to the target vehicle includes:

[0012] Obtain the location information of the monitoring device corresponding to each driver image in the driver image group corresponding to the target vehicle; determine the driving path corresponding to the target vehicle according to the location information of the monitoring device corresponding to each driver image in the driver image group.

[0013] Optionally, in the above method, determining the association relationship between the monitoring devices according to the monitoring devices corresponding to the driving path of the vehicle includes:

[0014] The FP-growth algorithm is used to calculate the association rules of the monitoring devices corresponding to the vehicle's driving path to determine the association relationship between the monitoring devices.

[0015] Optionally, in the above method, before using the FP-growth algorithm to calculate the association rules of the monitoring devices corresponding to the driving path of the vehicle, the method further includes:

[0016] Remove driving paths that contain only one monitoring device.

[0017] Optionally, in the above method, pre-classifying the plurality of driver images according to the association relationship between the monitoring devices includes:

[0018] According to the association relationship between the monitoring devices, the monitoring devices are divided into a plurality of monitoring device sets; and the driver images in the same monitoring device set are pre-classified into the same category.

[0019] Optionally, in the above method, the monitoring devices are divided into a plurality of monitoring device sets according to association relationships between the monitoring devices, including:

[0020] According to the association relationship between monitoring devices, a monitoring device dictionary is established with a monitoring device as the key value and another monitoring device that satisfies a strong association rule of positive correlation with the monitoring device as the value value; according to the monitoring device dictionary, the monitoring devices are divided into multiple monitoring device sets.

[0021] Optionally, in the above method, the monitoring devices are divided into a plurality of monitoring device sets according to the monitoring device dictionary, including:

[0022] According to the monitoring device dictionary, establish the monitoring device graph structure;

[0023] Obtain the monitoring devices in the monitoring device graph structure that satisfy the strong association rule of positive correlation with the value;

[0024] The monitoring devices satisfying the strong association rule of positive correlation are deduplicated to obtain a set of monitoring devices satisfying the strong association rule of positive correlation.

[0025] Optionally, in the above method, the method further comprises:

[0026] If there are monitoring devices that are not classified into any monitoring device set, the unclassified monitoring devices are classified into an existing monitoring device set.

[0027] Optionally, in the above method, the method further comprises:

[0028] When the number of driver images in the pre-classified target driver image set is greater than a first preset threshold, the target driver image set is segmented; or when the number of driver images in the pre-classified target driver image set is less than a second preset threshold, the target driver image set is merged into the remaining driver image set except the target driver images.

[0029] In order to solve the above technical problem, another technical solution adopted by the present application is: to provide an image classification device, which includes a processor and a memory.

[0030] Specifically, the memory is used to store program instructions, and the processor is used to execute the program instructions to implement the above-mentioned driver portrait pre-classification method.

[0031] In order to solve the above technical problem, another technical solution adopted by the present application is: providing a storage medium, which stores program instructions, and the program instructions can be executed to implement the above driver portrait pre-classification method.

[0032] Different from the prior art, the present application provides a driver portrait pre-classification method, an image classification device, and a storage medium, the method comprising: obtaining multiple driver images; wherein the driver images are acquired by monitoring equipment, and the driver images include driver portraits and vehicle information of corresponding vehicles; determining the monitoring equipment corresponding to the driving path of the same vehicle based on the vehicle information; determining the association relationship between the monitoring equipment based on the monitoring equipment corresponding to the driving path of the vehicle; and pre-classifying multiple driver images based on the association relationship between the monitoring equipment. Through the above method, the association relationship between the monitoring equipment is determined based on the driving path of the vehicle, and then the driver images are pre-classified based on the association relationship between the monitoring equipment, which can enhance the correlation between the driver images in the driver image set after pre-classification and improve the effect of driver portrait clustering. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0034] Figure 1 It is a flowchart of an embodiment of a driver portrait pre-classification method provided by the present application;

[0035] Figure 2 yes Figure 1 Schematic diagram of the process of S12;

[0036] Figure 3 yes Figure 2 Schematic diagram of the process of S121;

[0037] Figure 4 yes Figure 2 Schematic diagram of the process of S122;

[0038] Figure 5 yes Figure 1 Schematic diagram of the process of S13;

[0039] Figure 6 yes Figure 1 Schematic diagram of the process of S14;

[0040] Figure 7 yes Figure 6 Schematic diagram of the process of S141;

[0041] Figure 8 yes Figure 7 Schematic diagram of the process of S1412;

[0042] Fig. 9It is a structural schematic diagram of an embodiment of an image classification device provided by the present application;

[0043] Fig.10 It is a structural diagram of an embodiment of a storage medium provided by the present application. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It will be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only some but not all structures related to the present application are shown in the drawings. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0045] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0046] The steps in the embodiments of the present application do not necessarily have to be processed in the described order of steps. The steps can be selectively rearranged, or the steps in the embodiments can be deleted, or the steps in the embodiments can be added as needed. The step descriptions in the embodiments of the present application are only optional sequence combinations and do not represent all step sequence combinations in the embodiments of the present application. The order of steps in the embodiments cannot be considered as a limitation of the present application.

[0047] The term "and / or" in the embodiments of the present application refers to any and all possible combinations of one or more of the associated enumerated items. It should also be noted that when used in this specification, "include / comprise" specifies the presence of stated features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements and / or components and / or their groups.

[0048] The terms "first", "second", etc. in this application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0049] In the process of urban security and management, distributed computing is used to cluster the driver images collected by road monitoring equipment, and the identities are classified into clusters of people, so that the monitoring of drivers can be realized. In the prior art, the driver images collected by the monitoring equipment are directly classified in an average division manner, which often leads to weak correlation of the classified data and poor clustering effect.

[0050] Based on this, the present application proposes a driver portrait pre-classification method, which determines the correlation between monitoring devices according to the vehicle's driving path, and then pre-classifies the driver images according to the correlation between the monitoring devices. This method can enhance the correlation between driver images in the driver image set after pre-classification and improve the effect of driver portrait clustering.

[0051] See also Figure 1 , Figure 1 : is a flowchart of an embodiment of a driver portrait pre-classification method provided by the present application, the method comprising:

[0052] S11: Acquire multiple driver images; wherein the driver images are acquired by monitoring equipment, and the driver images include a driver portrait and vehicle information of a corresponding vehicle.

[0053] Optionally, the geographical scope for acquiring driver images is the size of a prefecture-level city, that is, the driver images collected by monitoring equipment within the scope of a prefecture-level city are acquired; according to the travel time analysis, the time period for acquiring driver images is 3 hours, that is, a batch of driver images collected by monitoring equipment are acquired every 3 hours. It can be understood that the geographical scope and time period for acquiring driver images can be set according to the specific requirements of urban security and management, and are not specifically limited here.

[0054] Optionally, the acquired driver image may include a variety of information, such as the monitoring device number, longitude and latitude coordinates, the driver's portrait, the vehicle's license plate, color, model, etc., which are not specifically limited here.

[0055] S12: Determine the monitoring device corresponding to the driving path of the same vehicle according to the vehicle information.

[0056] Optionally, the vehicle information includes the vehicle's license plate, color, model and other information. The acquired driver images can be grouped according to the vehicle information such as the license plate to determine the vehicle's driving path, and then the monitoring device corresponding to the vehicle's driving path can be determined based on the vehicle's driving path.

[0057] S13: Determine the association relationship between the monitoring devices according to the monitoring devices corresponding to the driving path of the vehicle.

[0058] Optionally, according to the monitoring devices corresponding to the vehicle's driving path, an association rule mining algorithm is used to calculate the association rules of the monitoring devices corresponding to the vehicle's driving path, so that the association relationship between the monitoring devices can be determined.

[0059] Optionally, commonly used association rule mining algorithms mainly include Apriori algorithm and FP-growth algorithm, wherein the FP-growth algorithm has a great improvement in efficiency compared with the Apriori algorithm. This embodiment mainly adopts the FP-growth algorithm to calculate association rules.

[0060] S14: Pre-classifying the plurality of driver images according to the association relationship between the monitoring devices.

[0061] It can be understood that in order to improve the effect of driver portrait clustering, when pre-classifying driver images, driver images belonging to the same driver should be grouped together as much as possible. If two monitoring devices meet the strong association rule of positive correlation, then the possibility that the two monitoring devices have collected driver images of the same driver is the greatest. For example, if driver Xiao Ming is photographed by monitoring device A while driving, and monitoring device A and monitoring device B meet the strong association rule of positive correlation, then the possibility that Xiao Ming is photographed by monitoring device B while driving is also very high. Therefore, the driver images collected by monitoring devices that meet the strong association rule of positive correlation also have a strong correlation.

[0062] Optionally, based on the association relationship between monitoring devices, driver images collected by monitoring devices that satisfy the strong association rule of positive correlation are pre-classified into the same category, which can enhance the association between driver images in the obtained driver image set.

[0063] Compared with the prior art, this embodiment provides a driver portrait pre-classification method, which includes: acquiring multiple driver images; wherein the driver images are acquired by monitoring equipment, and the driver images include driver portraits and vehicle information of corresponding vehicles; determining the monitoring equipment corresponding to the driving path of the same vehicle based on the vehicle information; determining the association relationship between the monitoring equipment based on the monitoring equipment corresponding to the driving path of the vehicle; and pre-classifying multiple driver images based on the association relationship between the monitoring equipment. Through the above method, the association relationship between the monitoring equipment is determined based on the driving path of the vehicle, and then the driver images are pre-classified based on the association relationship between the monitoring equipment, which can enhance the correlation between the driver images in the driver image set after pre-classification and improve the effect of driver portrait clustering.

[0064] See also Figure 2 , Figure 2 yes Figure 1 In the flowchart of S12, S12 may further include:

[0065] S121: Grouping the plurality of driver images according to the vehicle information to obtain a driver image group corresponding to a target vehicle.

[0066] Optionally, the vehicle information also includes license plate information of the vehicle, and the plurality of driver images are grouped according to the license plate information to obtain a driver image group corresponding to each vehicle.

[0067] S122: Determine the driving path of the target vehicle according to the driver image group corresponding to the target vehicle.

[0068] Optionally, based on the driver image group corresponding to the vehicle, the intersections equipped with monitoring equipment that the vehicle passes through during driving can be determined, and then multiple optional driving paths for the vehicle can be determined.

[0069] S123: Determine the monitoring device corresponding to the driving path of the target vehicle according to the driving path of the target vehicle.

[0070] Optionally, the corresponding monitoring device is determined according to the location information of the monitoring device included in the vehicle's driving path, for example, the corresponding monitoring device number is determined according to the longitude and latitude coordinates of the monitoring device, wherein the longitude and latitude coordinates of the monitoring device and the number are in one-to-one correspondence.

[0071] See also Figure 3 , Figure 3 yes Figure 2 In the flowchart of S121, S121 may further include:

[0072] S1211: Grouping the multiple driver images based on the license plate information corresponding to the multiple driver images to obtain a first driver image group corresponding to a target vehicle.

[0073] Optionally, according to the license plate information, the driver images of the same vehicle are divided into the same group to obtain a first driver image group.

[0074] S1212: Sort the driver images in the first driver image group based on the acquisition time of the driver images to obtain a second driver image group.

[0075] It can be understood that the acquisition time of the multiple driver images has a chronological order, corresponding to the time when the driver drives the vehicle through the intersection equipped with the monitoring equipment.

[0076] Optionally, according to the acquisition time of the driver images, the driver images in the first driver image group are sorted in the order of acquisition time to obtain the second driver image group.

[0077] S1213: Deduplication processing is performed on the second driver image group to obtain a driver image group corresponding to the target vehicle.

[0078] It is understandable that when the same driver drives a vehicle through an intersection equipped with monitoring equipment, the monitoring equipment generally collects multiple driver images, and these driver images are repeated; when two monitoring devices are installed at the same intersection, the driver images collected by the two monitoring devices are also repeated. In the above grouping process, these repeated driver images are grouped into the same group.

[0079] Optionally, the second driver image group is deduplicated to obtain a driver image group of the target vehicle.

[0080] See also Figure 4 , Figure 4 yes Figure 2 In the flowchart of S122, S122 may further include:

[0081] S1221: Obtain location information of a monitoring device corresponding to each driver image in the driver image group corresponding to the target vehicle.

[0082] Optionally, the driver image includes location information of the monitoring device, such as longitude and latitude coordinates, so the longitude and latitude coordinates of the monitoring device corresponding to each driver image in the driver image group of the vehicle can be obtained.

[0083] S1222: Determine the driving path corresponding to the target vehicle according to the location information of the monitoring device corresponding to each driver image in the driver image group.

[0084] Optionally, for a vehicle, the location information of the monitoring device corresponding to each driver image in the corresponding driver image group, such as the latitude and longitude coordinates of the monitoring device, is input into a map search tool to obtain multiple optional driving routes for the vehicle.

[0085] Optionally, the map search tool presets multiple recommendation strategies, such as shortest distance, shortest time, least traffic lights passed and least fees, etc. Based on these recommendation strategies, multiple driving routes corresponding to the vehicle can be obtained.

[0086] See also Figure 5 , Figure 5 yes Figure 1 In the flowchart of S13, S13 may further include:

[0087] S131: Remove the driving path that only includes one monitoring device.

[0088] Optionally, when the driving path of the vehicle includes only one monitoring device, it is useless for calculating the association relationship between the monitoring devices, so the driving path including only one monitoring device is removed.

[0089] It is understandable that this step may not be performed when the driving paths of all vehicles include two or more monitoring devices.

[0090] S132: Using the FP-growth algorithm, calculate the association rules of the monitoring devices corresponding to the vehicle's driving path to determine the association relationship between the monitoring devices.

[0091] Optionally, the minimum support (min_support) and the minimum confidence (min_confidence) are set, and the FP-growth algorithm is used to calculate the association rules of the monitoring devices corresponding to the vehicle's driving path, obtain frequent items, and sort them from large to small.

[0092] Specifically, the FP-growth algorithm is an efficient association rule mining algorithm that can be used to mine the connections between item sets from a large amount of data. A typical example of association rule mining is shopping basket analysis, which analyzes customers' purchasing habits by discovering the connections between different items placed in their shopping baskets. By understanding which items are frequently bought by customers at the same time, it can help retailers develop reasonable marketing strategies.

[0093] Specifically, the smallest unit of information that cannot be divided in the data is called an item, and a collection of items is called an itemset.

[0094] Specifically, the support of an association rule refers to the ratio of the number of item sets that contain both X and Y to the number of all item sets, denoted as support(X=>Y). The support reflects the probability that an item set that contains both X and Y appears in all item sets.

[0095] Specifically, the confidence of an association rule is the ratio of the number of item sets that contain both X and Y to the number of item sets that contain X, denoted as confidence(X=>Y). Confidence reflects the conditional probability of Y appearing in the item set containing X.

[0096] Specifically, if the minimum confidence threshold and the minimum support threshold are satisfied at the same time, that is, support(X=>Y) and confidence(X=>Y) are both greater than or equal to the set threshold, then the association rule X=>Y is called a strong association rule. In addition, if the lift is greater than 1, it indicates a strong association rule with positive correlation; if the lift is less than 1, it indicates a strong association rule with negative correlation. The lift represents the ratio of the probability of containing Y under the condition of containing X to the probability of Y occurring overall.

[0097] Optionally, the size of the minimum support and the minimum confidence determines the number of monitoring devices with strong association rules, so the minimum support and the minimum confidence can be set according to the specific requirements of the driver portrait pre-classification.

[0098] Optionally, monitoring devices with strong association rules are obtained, rules with a lift greater than 1 are selected, and monitoring devices that satisfy the strong association rules with positive correlation are further determined to obtain the association relationship between the monitoring devices.

[0099] See also Figure 6 , Figure 6 yes Figure 1 In the flowchart of S14, S14 may further include:

[0100] S141: According to the association relationship between the monitoring devices, the monitoring devices are divided into a plurality of monitoring device sets.

[0101] Optionally, according to the association relationship between the monitoring devices, the monitoring devices that satisfy the strong association rule of positive correlation are divided into the same monitoring device set to obtain multiple monitoring device sets.

[0102] Optionally, in some other embodiments, if there are monitoring devices that are not classified into any monitoring device set, the unclassified monitoring devices are classified into an existing monitoring device set.

[0103] It can be understood that in the above S131, the driving path containing only one monitoring device is removed, and some monitoring devices do not satisfy the strong association rule of positive correlation with other devices, so these monitoring devices need to be merged into the monitoring device set.

[0104] Optionally, filter out this part of monitoring devices from all monitoring devices, obtain the location information corresponding to this part of monitoring devices, such as longitude and latitude coordinates, calculate the distance between this part of monitoring devices and the monitoring devices in the monitoring device set according to the location information, and merge this part of monitoring devices into the monitoring device set with the closest distance according to the distance.

[0105] S142: Pre-classify driver images in the same monitoring device set into the same category.

[0106] Optionally, the driver images collected by the monitoring devices in the same monitoring device set have a strong correlation, and the driver images in the same monitoring device set are pre-classified into the same category to obtain a corresponding driver image set.

[0107] S143: When the number of driver images in the pre-classified target driver image set is greater than a first preset threshold, segmenting the target driver image set.

[0108] Optionally, based on the business requirements of driver portrait pre-classification and the processing speed of the hardware equipment, a threshold range for the number of driver images in a driver image set is set to ensure that the data volume of the driver image set obtained by pre-classification is more balanced; wherein the upper limit of the threshold range is the first preset threshold and the lower limit is the second preset threshold.

[0109] Optionally, when the number of driver images in the pre-classified driver image set is greater than a first preset threshold, the driver image set is split so that the number of driver images therein meets the requirement of the threshold range.

[0110] Optionally, when the number of driver images in the pre-classified driver image set is less than a second preset threshold, S144 is executed.

[0111] S144: When the number of driver images in the pre-classified target driver image set is less than a second preset threshold, merging the target driver image set into a remaining driver image set except the target driver image set.

[0112] Optionally, when the number of driver images in the pre-classified driver image set is less than a second preset threshold, the driver image set is merged into other driver image sets except the driver set, so that the number of driver images in all driver image sets meets the requirements of the threshold range.

[0113] See also Figure 7 , Figure 7 yes Figure 6In the flowchart of S141, S141 may further include:

[0114] S1411: According to the association relationship between the monitoring devices, a monitoring device dictionary is established with a monitoring device as a key value and another monitoring device that satisfies a strong association rule of positive correlation with the monitoring device as a value value.

[0115] Optionally, the monitoring devices are mapped according to the association relationship between the monitoring devices, and a monitoring device dictionary is established with a monitoring device as a key value and another monitoring device that satisfies a strong association rule of positive correlation with the monitoring device as a value.

[0116] In an optional embodiment, it is assumed that the positively correlated strong association rules involving the monitoring device A include the following situations: A=>B; A, B=>C; A=>D, C; E=>A; E=>A, B.

[0117] First select A=>B; A=>D, C. Select B, D, C, and sort them in descending order of confidence value.

[0118] Next, select E=>A; E=>A, B. Here, choose E instead of B, because the purpose is to find A. Starting from E, you can find A, starting from E, you can also find B, or starting from E, you can know the result of going from A to B. This situation may not be good for A, so don't choose B here.

[0119] Finally, select A, B=>C. Select B, C, and sort them in descending order according to the confidence value.

[0120] Therefore, the following data A=>(B,D,C),(E,B),(B,C) is generated. Then, the related monitoring devices are deduplicated to obtain A=>B,D,C,E. Finally, the monitoring device corresponding to A is stored in the monitoring device dictionary.

[0121] S1412: According to the monitoring device dictionary, the monitoring devices are divided into a plurality of monitoring device sets.

[0122] Optionally, according to the monitoring device dictionary, monitoring devices that satisfy the strong association rule of positive correlation are divided into the same monitoring device set.

[0123] See also Figure 8 , Figure 8 yes Figure 7 In the flowchart of S1412, S1412 may further include:

[0124] S14121: Establish a monitoring device graph structure according to the monitoring device dictionary.

[0125] Optionally, a graph is used to deduplicate the labeling results. Establishing a monitoring device graph structure will add an edge between monitoring devices.

[0126] S14122: Obtain monitoring devices in the monitoring device graph structure that satisfy a strong association rule of positive correlation with the value.

[0127] Optionally, the value corresponding to the key value in the monitoring device dictionary is taken out, that is, another monitoring device that satisfies the strong association rule of positive correlation with the monitoring device, and the monitoring device that satisfies the strong association rule of positive correlation with the value value is obtained by deep traversal. When performing deep traversal, the monitoring device that satisfies the strong association rule of positive correlation with the value value is edge-connected.

[0128] Specifically, the method of deep traversal includes: starting from a vertex v in the graph structure, visiting vertex v; starting from the unvisited adjacent points of v in turn, performing deep traversal of the graph; until all vertices in the graph structure that have paths connecting to v are visited; if there are still vertices in the graph structure that have not been visited at this time, starting from an unvisited vertex, re-performing deep traversal until all vertices in the graph structure have been visited.

[0129] S14123: De-duplicate the monitoring devices that satisfy the strong association rule of positive correlation to obtain a set of monitoring devices that satisfy the strong association rule of positive correlation.

[0130] Optionally, when duplicate monitoring devices appear, the corresponding edge in the monitoring device graph structure is deleted. In an optional embodiment, assuming A=>B,D,C,E; B=>F,E,H, the edge between E and F is deleted.

[0131] Optionally, after deduplication processing is performed on monitoring devices that satisfy the strong association rule of positive correlation in the monitoring device graph structure, a set of monitoring devices that satisfy the strong association rule of positive correlation can be obtained.

[0132] See also Fig. 9 , Fig. 9 2 is a schematic diagram of the structure of an embodiment of an image classification device provided in the present application. The image classification device 200 includes a processor 201 and a memory 202 .

[0133] Specifically, the memory 202 is used to store program instructions, and the processor 201 is used to execute the program instructions to implement the method provided by any one or any non-conflicting combination of the above embodiments.

[0134] Optionally, processor 201 is a central processing unit (CPU), which is one of the main devices of an electronic computer and a core component in a computer. Its main function is to interpret computer instructions and process data in computer software. CPU is the core component in a computer responsible for reading instructions, decoding instructions and executing instructions. The central processing unit mainly includes two parts, namely a controller and an arithmetic unit, which also includes a cache memory and a data and control bus that realizes the connection between them. The functions of the central processing unit are mainly to process instructions, perform operations, control time, and process data. In computer architecture, the CPU is a core hardware unit that controls and allocates all hardware resources of the computer (such as memory, input and output units) and performs general operations. The CPU is the computing and control core of the computer. The operations of all software layers in the computer system will eventually be mapped to the operations of the CPU through the instruction set.

[0135] Optionally, the memory 202 is a read-only memory (ROM) or a random access memory (RAM), which is a memory device in a computer system, mainly used to store programs and data. All information in the computer, including input raw data, computer programs, intermediate operation results and final operation results, are stored in the memory. It stores and retrieves information according to the location specified by the controller.

[0136] In an optional embodiment, the image classification device 200 is a server in the background of the city security management system, which is used to periodically obtain driver images collected by road monitoring equipment in a certain area of ​​the city, and pre-classify the driver's portraits according to the method provided by any one or any non-conflicting combination of the above embodiments, so as to keep the load of each computing node balanced during the distributed computing process.

[0137] See also Fig.10 , Fig.10 300 includes program instructions 301, which can be executed to implement the method provided by any one or any non-conflicting combination of the above embodiments. The capacity of the storage medium 300 can meet the requirements of storing the program instructions 301.

[0138] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more storage media 300 (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer usable program code.

[0139] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by the storage medium 300. These storage media 300 can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the program instructions 301 executed by the processor of the computer or other programmable data processing device generate the instructions for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0140] These storage media 300 may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the program instructions 301 stored in the storage media 300 produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0141] These storage media 300 can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing program instructions 301 executed on the computer or other programmable device for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0142] The above description is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made according to the description and drawings of the present application, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A driver portrait pre-classification method, characterized in that: The method comprises: Acquire multiple driver images; wherein the driver images are acquired by monitoring equipment, and the driver images include driver portraits and vehicle information of corresponding vehicles; Determine the monitoring device corresponding to the driving path of the same vehicle according to the vehicle information; Determine the association relationship between the monitoring devices according to the monitoring devices corresponding to the vehicle's driving path; Pre-classifying the plurality of driver images according to the association relationship between the monitoring devices; The pre-classifying the plurality of driver images according to the association relationship between the monitoring devices comprises: According to the association relationship between the monitoring devices, the monitoring devices are divided into a plurality of monitoring device sets; Driver images in the same monitoring device set are pre-classified into the same class.

2. The method according to claim 1, characterized in that The step of determining the monitoring device corresponding to the driving path of the same vehicle according to the vehicle information includes: According to the vehicle information, the plurality of driver images are grouped to obtain a driver image group corresponding to the target vehicle; Determining a driving path of the target vehicle according to the driver image group corresponding to the target vehicle; According to the driving path of the target vehicle, a monitoring device corresponding to the driving path of the target vehicle is determined.

3. The method according to claim 2, characterized in that The vehicle information includes license plate information; The step of grouping the plurality of driver images according to the vehicle information to obtain a driver image group corresponding to the target vehicle includes: Based on the license plate information corresponding to the multiple driver images, grouping the multiple driver images to obtain a first driver image group corresponding to the target vehicle; sorting the driver images in the first driver image group based on the acquisition time of the driver images to obtain a second driver image group; The second driver image group is deduplicated to obtain a driver image group corresponding to the target vehicle.

4. The method according to claim 2, characterized in that: Determining the driving path of the target vehicle according to the driver image group corresponding to the target vehicle includes: Acquire location information of a monitoring device corresponding to each driver image in a driver image group corresponding to the target vehicle; The driving path corresponding to the target vehicle is determined according to the position information of the monitoring device corresponding to each driver image in the driver image group.

5. The method according to claim 1, characterized in that: The determining the association relationship between the monitoring devices according to the monitoring devices corresponding to the driving path of the vehicle includes: The FP-growth algorithm is used to calculate the association rules of the monitoring devices corresponding to the vehicle's driving path to determine the association relationship between the monitoring devices.

6. The method according to claim 5, characterized in that Before the FP-growth algorithm is used to calculate the association rules of the multiple monitoring devices corresponding to the driving path of the vehicle, the method further includes: Remove driving paths that contain only one monitoring device.

7. The method according to claim 1, characterized in that The monitoring devices are divided into a plurality of monitoring device sets according to the association relationship between the monitoring devices, including: According to the association relationship between the monitoring devices, a monitoring device dictionary is established with a monitoring device as a key value and another monitoring device that satisfies a strong association rule of positive correlation with the monitoring device as a value value; According to the monitoring device dictionary, the monitoring devices are divided into a plurality of monitoring device sets.

8. The method according to claim 7, characterized in that The monitoring devices are divided into a plurality of monitoring device sets according to the monitoring device dictionary, including: According to the monitoring device dictionary, a monitoring device graph structure is established; Obtaining monitoring devices in the monitoring device graph structure that satisfy a strong association rule of positive correlation with the value; The monitoring devices satisfying the strong association rule of positive correlation are deduplicated to obtain a set of monitoring devices satisfying the strong association rule of positive correlation.

9. The method according to claim 1, characterized in that: The method further comprises: If there are monitoring devices that are not classified into any monitoring device set, the unclassified monitoring devices are classified into an existing monitoring device set.

10. The method according to claim 1, characterized in that The method further comprises: When the number of driver images in the pre-classified target driver image set is greater than a first preset threshold, segmenting the target driver image set; or When the number of driver images in the pre-classified target driver image set is less than a second preset threshold, the target driver image set is merged into a remaining driver image set except the target driver image set.

11. An image classification device, characterized in that: The image classification device comprises a processor and a memory, wherein the memory is used to store program instructions, and the processor is used to execute the program instructions to implement the method according to any one of claims 1 to 10.

12. A storage medium, characterized in that: Program instructions are stored, and the program instructions can be executed to implement the method according to any one of claims 1 to 10.

Images