CN114494994A

CN114494994A - Vehicle abnormal aggregation monitoring method and device, computer equipment and storage medium

Info

Publication number: CN114494994A
Application number: CN202111660556.XA
Authority: CN
Inventors: 蔡志东
Original assignee: Shenzhen Intellifusion Technologies Co Ltd
Current assignee: Shenzhen Intellifusion Technologies Co Ltd
Priority date: 2021-12-30
Filing date: 2021-12-30
Publication date: 2022-05-13

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a vehicle abnormal aggregation monitoring method, a device, computer equipment and a storage medium, wherein the method comprises the following steps: acquiring a plurality of image data shot in a region to be monitored within a preset time period; determining the number of target vehicles of a specific vehicle type in a region to be monitored within a preset time period and owner information of the target vehicles according to the plurality of image data; determining the owner ratio of the abnormal aggregation records based on the number of vehicles, the owner information of the target vehicle and a preset abnormal aggregation blacklist; and determining an abnormal aggregation detection result of the area to be monitored in a preset time period according to the number of the vehicles and the vehicle owner occupation ratio. Therefore, the invention realizes the automatic determination of the vehicle number and the owner information of the target vehicle of the specific vehicle type, automatically identifies the abnormal aggregation detection result of the area to be monitored based on the abnormal aggregation blacklist, and improves the accuracy of the abnormal aggregation monitoring of the vehicle.

Description

Vehicle abnormal aggregation monitoring method and device, computer equipment and storage medium

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a vehicle abnormal aggregation monitoring method and device, computer equipment and a storage medium.

Background

At present, with the progress of society and the rapid development of economy, the quality of life of people is improved, the use and the mobility of vehicles are rapidly increased, abnormal gathering conditions of people or vehicles, such as illegal behaviors like collective fighting, are easy to occur, alarm calls which are only received by public security related departments often occur after the conditions, a hysteresis phenomenon exists, serious dangerous events are easy to occur, in the existing early warning/monitoring method for abnormal gathering, whether high-density gathering conditions of vehicles in a certain area occur or not is mostly determined through the positioning information of a vehicle-mounted GPS terminal, and therefore early warning/monitoring is carried out according to the density conditions, but misjudgment conditions are easy to occur in the scheme, such as the conditions of peak periods of going to and going to work, a certain exhibition or a gathered vehicle is not a concerned object (also called a vehicle owner with illegal behaviors or a vehicle with illegal records), therefore, the early warning/monitoring accuracy and the accuracy of the abnormal aggregation are not high, and the resource waste of the working personnel for processing the abnormal aggregation related affairs which are early warned/monitored is caused.

Disclosure of Invention

The invention provides a vehicle abnormal aggregation monitoring method, a device, computer equipment and a storage medium, which can determine the vehicle number and the owner information of a target vehicle of a specific vehicle type from massive image data, and objectively, accurately and automatically identify whether abnormal aggregation exists in an abnormal aggregation detection result of an area to be monitored based on an abnormal aggregation blacklist.

A vehicle abnormal aggregation monitoring method, comprising:

acquiring a plurality of image data shot in a region to be monitored within a preset time period;

determining the number of target vehicles of a specific vehicle type in a region to be monitored within a preset time period and owner information of the target vehicles according to the image data;

determining the owner ratio value with abnormal aggregation records based on the number of the vehicles, the owner information of the target vehicle and a preset abnormal aggregation blacklist;

and determining an abnormal aggregation detection result of the area to be monitored in a preset time period according to the number of the vehicles and the vehicle owner ratio.

A vehicle abnormality gathering monitoring device comprising:

the system comprises an acquisition module, a monitoring module and a monitoring module, wherein the acquisition module is used for acquiring a plurality of image data which are obtained by shooting in a to-be-monitored area within a preset time period;

the statistical module is used for determining the number of target vehicles of a specific vehicle type in a region to be monitored within a preset time period and the owner information of the target vehicles according to the image data;

the determining module is used for determining the owner ratio value of the abnormal aggregation records based on the number of the vehicles, the owner information of the target vehicle and a preset abnormal aggregation blacklist;

and the output module is used for determining an abnormal aggregation detection result of the area to be monitored in a preset time period according to the number of the vehicles and the vehicle owner occupation ratio.

A computer device comprising a memory, a processor and a computer program stored in said memory and executable on said processor, said processor implementing the steps of the above-mentioned vehicle anomaly aggregate monitoring method when executing said computer program.

A computer-readable storage medium, which stores a computer program that, when executed by a processor, implements the steps of the above-described vehicle abnormal gathering monitoring method.

According to the method, the device, the computer equipment and the storage medium for monitoring the abnormal aggregation of the vehicle, a plurality of image data which are obtained by shooting in a region to be monitored within a preset time period are obtained; determining the number of target vehicles of a specific vehicle type in a region to be monitored within a preset time period and owner information of the target vehicles according to the image data; determining the owner ratio value with abnormal aggregation records based on the number of the vehicles, the owner information of the target vehicle and a preset abnormal aggregation blacklist; according to the number of the vehicles and the owner ratio, determining the abnormal aggregation detection result of the area to be monitored in the preset time period, therefore, the invention automatically determines the owner ratio through the image data shot in the area to be monitored in the preset time period, automatically determines the number of the vehicles and the owner information of the target vehicles with the specific vehicle types from all the image data, automatically outputs the abnormal aggregation detection result of the area to be monitored based on the abnormal aggregation blacklist, integrates the number of the vehicles and the owner ratio, thus obtaining the image data from massive shot data, determining the number of the vehicles and the owner information of the target vehicles with the specific vehicle types, objectively, accurately and automatically identifying whether the abnormal aggregation detection result of the area to be monitored has the abnormal aggregation condition or not based on the abnormal aggregation blacklist, whether the image data are gathered or not does not need to be identified manually, labor cost and energy are greatly saved, and the accuracy and the correctness of monitoring abnormal gathering of the vehicle are improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.

FIG. 1 is a schematic diagram of an application environment of a vehicle abnormal aggregation monitoring method according to an embodiment of the present invention;

FIG. 2 is a flow chart of a vehicle anomaly aggregation monitoring method in one embodiment of the present invention;

FIG. 3 is a flowchart of step S10 of the vehicle abnormal collection monitoring method according to one embodiment of the present invention;

FIG. 4 is a schematic block diagram of a vehicle anomaly aggregate monitoring apparatus in an embodiment of the present invention;

FIG. 5 is a schematic diagram of a computer device in an embodiment of the invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The vehicle abnormal aggregation monitoring method provided by the invention can be applied to the application environment shown in fig. 1, wherein a client (computer equipment or terminal) is communicated with a server through a network. The client (computer device or terminal) includes, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices. The server may be implemented as a stand-alone server or as a server cluster consisting of a plurality of servers.

In an embodiment, as shown in fig. 2, a method for monitoring abnormal aggregation of vehicles is provided, which mainly includes the following steps S10-S40:

and S10, acquiring a plurality of image data shot in the area to be monitored within a preset time period.

Understandably, the image data is data which is extracted from original data collected by an area to be monitored within a preset time period, the preset time period can be set according to requirements, for example, 14:00 to 14:59 in 24 hours a day, the area to be monitored is an area to be deployed and needs to be monitored, the monitoring area can be shot through deployed image collection equipment, the original data is an image shot through image collection equipment of the area to be monitored which is deployed on a large data platform and data of relevant shooting information, the image shot in the original data can contain a vehicle or does not contain a vehicle, the original data comprises an equipment identifier and a snapshot image, the equipment identifier is a unique identifier of the image collection equipment which is currently shot, one equipment identifier is associated with one monitoring area identifier, the area to be monitored which is shot by the image collection equipment can be obtained through the equipment identifier, each area to be monitored is given a unique identifier, the image acquisition device can be a camera or a timing acquisition device for shooting images at preset time intervals, the snap-shot images are one frame of images shot by the image acquisition device, and the process of extracting image data to be monitored (including vehicle related information) from massive original data can be as follows: firstly, acquiring a plurality of original data of an area to be monitored in a preset time period; secondly, carrying out vehicle identification on the snapshot image in each original data, identifying which original data contain a first identification result of the vehicle, and screening out the original data containing the vehicle according to the identified first identification result; and finally, extracting the vehicle snapshot image from the screened original data, identifying the vehicle sub-attribute of the extracted vehicle snapshot image, identifying a second identification result of the vehicle attention attribute, and recording the first identification result and the second identification result of the original data as the image data of the original data, wherein the vehicle attention attribute can be set according to requirements, such as the vehicle attention attribute comprises a vehicle license plate, a vehicle type (such as a taxi, a family car, a vehicle type and the like) and the like.

In one embodiment, as shown in fig. 3, the step S10 of acquiring a plurality of image data captured in the area to be monitored within the preset time period includes:

s101, acquiring a plurality of original data.

Understandably, the original data is an image shot by image acquisition equipment of the area to be monitored deployed through a big data platform and data of shooting related information, the image shot in the original data can contain vehicles or not, the original data also comprises monitoring time, the monitoring time is the shooting time of corresponding image acquisition equipment, the original data in a preset time period can be obtained through the monitoring time, the equipment identification is a unique identification code of the currently shot image acquisition equipment, one equipment identification is associated with one monitoring area identification, the area to be monitored shot by the image acquisition equipment can be obtained through the equipment identification, each area to be monitored is endowed with a unique identification, the image acquisition equipment can be a camera or a timing collector for shooting images at intervals of preset time, and the captured image is a frame of image shot by the image acquisition equipment, the method comprises the steps that a monitoring area identifier of an area to be monitored is associated with a risk level, the risk level reflects the risk degree of aggregation of the monitoring area, the risk level can be updated continuously along with the time, for example, the number of times of abnormal aggregation detection results with abnormal aggregation is accumulated to determine the risk level, namely, the risk level of the monitoring area can be analyzed along with the related historical data of the monitoring area, the acquired video is subjected to frame processing through video acquisition equipment to obtain original data, the frame processing is to extract each frame image in the acquired video, and the processing process of marking the monitoring time and the equipment identifier for each frame image.

S102, vehicle identification is carried out on the snapshot image in the original data aiming at each original data to obtain a first identification result corresponding to the original data; the first recognition result shows whether the snapshot image contains a vehicle or not.

Understandably, the process of vehicle identification of the captured image in the original data can be realized by a vehicle identification network completed through training, the process of vehicle identification is a process of extracting vehicle features of the captured image, determining a region containing the vehicle features according to the extracted vehicle features, determining the region as a vehicle region, and determining whether a first identification result of the vehicle is contained according to the number of the vehicle regions, wherein the vehicle attribute result comprises the first identification result and identification results output by other vehicle related attribute networks, the first identification result shows whether the captured image contains the vehicle, and the vehicle features are features of the vehicle, such as: the vehicle front end has a license plate characteristic, the vehicle bottom has a tire characteristic, the vehicle front end has a portrait characteristic and the like, the network structure of the vehicle identification network can be set according to requirements, for example, the network structure of the vehicle identification network can be a network structure such as CenterNet, Faster R-CNN, SSD and YOLO, preferably, the network structure of the vehicle identification network is a network structure of YOLO V3, because the position of the vehicle can be quickly located from different scales by using the network structure of YOLO V3, the training process of the vehicle identification network is to input a sample image containing a vehicle marking area, extract (convolve) the vehicle characteristic of the sample image, obtain a small-scale vehicle characteristic map according to the extracted vehicle characteristic, perform up-sampling and tensor stitching on the small-scale vehicle characteristic map to obtain a medium-scale vehicle characteristic map, and perform up-sampling and tensor stitching on the medium-scale vehicle characteristic map, and obtaining a large-scale vehicle characteristic map.

And further, carrying out multi-scale detection on the small-scale, medium-scale and large-scale vehicle characteristic graphs, predicting a rectangular region with vehicle characteristics, calculating a loss value between the rectangular region and a vehicle labeling region, and iteratively updating parameters of a vehicle identification network according to the loss value until the loss value reaches a convergence condition, so as to stop the training process, wherein the small-scale, medium-scale and large-scale are 1/32 scale, 1/16 scale and 1/8 scale relative to the size of the snapshot image, and the vehicle characteristics with finer granularity can be detected through the vehicle characteristic graphs with different scales. The method comprises the steps of performing convolution on a vehicle identification network, wherein the convolution is performed on a captured image, the vehicle identification network comprises a vehicle identification process and a tensor splicing process, the upsampling process is reverse propagation of convolution, namely, deconvolution or forward propagation transposition, the tensor splicing process is a process of splicing an upsampling result of a middle layer and an adjacent next layer in the convolution process, the vehicle identification process of the vehicle identification network is a process of extracting vehicle characteristics from the captured image through the vehicle identification network, predicting vehicle areas according to the extracted vehicle characteristics, predicting results of rectangular vehicle areas of zero or at least one vehicle, if zero vehicle areas are predicted, determining that a first identification result is not provided with the vehicle, and if at least one vehicle area is predicted, determining that the first identification result is provided with the vehicle and comprises each predicted vehicle area.

And S103, screening all the original data to screen out the original data containing the vehicles based on the first identification results corresponding to the original data.

Understandably, screening all the original data including the vehicle according to the first identification result including the vehicle to screen out the original data including the vehicle.

And S104, extracting the snapshot image in the screened original data based on the vehicle area in the first identification result to obtain at least one vehicle snapshot image, and performing vehicle sub-attribute identification on the vehicle snapshot image to obtain a second identification result corresponding to the original data.

Understandably, for one screened original data, according to at least one vehicle area in the first recognition result, the vehicle area is a rectangular area (coordinate area in the snapshot image) framing a vehicle in the snapshot image, vehicle snapshot images corresponding to the vehicle areas one by one are copied from the snapshot image corresponding to the first recognition result, that is, one vehicle area corresponds to one vehicle snapshot image, the vehicle sub-attribute recognition is a recognition process of recognizing a result of the vehicle attention attribute in the input vehicle snapshot image through a network of the vehicle attention attribute, networks of other vehicle attention attributes include a target detection network, a character recognition network and a vehicle type detection network, and the vehicle sub-attribute recognition process may be: carrying out owner area positioning on the vehicle snapshot image to obtain an owner area image of the vehicle snapshot image; carrying out license plate and vehicle type attribute identification on the vehicle snapshot image to obtain a license plate number and attribute result of the vehicle snapshot image; and finally, summarizing the owner area image, the license plate number and the attribute result of the vehicle snapshot image to obtain the result of each sub-attribute of the vehicle, so that the result of each sub-attribute of the vehicle is determined as a second identification result.

In one embodiment, the vehicle owner area of the vehicle snapshot image is located to obtain a vehicle owner area image of the vehicle snapshot image.

Understandably, the process of vehicle owner zone location is: firstly, carrying out image preprocessing on an input vehicle snapshot image, wherein the image preprocessing comprises an image zooming technology and an image enhancement technology so as to enable the size of the vehicle snapshot image to be in line with the subsequent positioning of a vehicle owner area, enable the vehicle snapshot image to be clearer and enable image information contained in the vehicle snapshot image to be more obvious, and obtain a to-be-processed image with a preset size (for example 512 multiplied by 512) after the image preprocessing; secondly, performing owner region identification on the image to be processed, wherein the owner region identification process can be realized through a trained target detection network, the network structure of the target detection network can be set according to requirements, for example, the network structure of the target detection network can be VGG, CenterNet, Faster R-CNN, SSD, YOLO and the like, preferably, the network structure of the target detection network is a network structure based on VGG19, the owner feature is extracted from the image to be processed through the target detection network, the owner feature is a feature related to an owner of a driving position, for example, the owner feature is a feature with a human face, a steering wheel and a driving state with the human face adjacent to the steering wheel, and classification identification is performed according to the extracted owner feature to identify the detection result of the main region range of the vehicle; and finally, extracting the vehicle owner area image from the image to be processed according to the vehicle owner area range in the detection result.

The image scaling technology is a method for making the size of an image reach a preset size by using a down-sampling or/and bilinear interpolation method, the down-sampling method is a method for carrying out down-sampling on a vehicle snapshot image through convolution check of a preset down-sampling unit and continuously and circularly carrying out down-sampling until the size of the image reaches the preset size, the down-sampling method also comprises a method for calculating a proportional value according to the preset size and the current size of the vehicle snapshot image, then carrying out down-sampling on the current vehicle snapshot image according to the size of a down-sampling convolution kernel corresponding to the proportional value, and enabling the size of the vehicle snapshot image of a gun to reach the preset size, the bilinear interpolation method is a method for respectively carrying out linear interpolation in two directions (the X-axis direction and the Y-axis direction) aiming at a target pixel point needing to be inserted, namely firstly using linear interpolation in one direction, the method comprises the steps of obtaining a first interpolation value, then obtaining a second interpolation value by using linear interpolation in the other direction, obtaining a pixel value of a target pixel point according to the weighted sum of the first interpolation value and the second interpolation value, linearly amplifying the pixel position of each coordinate point according to a linear amplification function between the current size and the preset size after the pixel value of each target pixel is inserted by a bilinear interpolation method so as to expand the pixel position to the preset size, freely zooming the vehicle snapshot image by a downsampling method and the bilinear interpolation method, and linearly calculating the numerical value according to two data which are close to the left and right of a point needing to be interpolated in a one-dimensional data sequence to obtain the linear interpolation value.

The image enhancement processing method is set according to requirements, for example, the image enhancement method may include a method of removing noise, enhancing contrast, enhancing color, and the like, the noise removal is a processing process of transforming each pixel in an input image by using a fourier transform algorithm, the contrast enhancement is a processing process of enhancing a black-and-white contrast degree to make a boundary or an edge in the image more obvious, the color enhancement is a processing process of converting an image of each pixel in the image into a color space of another spatial domain, subtracting a minimum value of all pixels in the spatial domain from a pixel value of each pixel, and converting the pixel value into a space of an original domain, and the gray scale processing is a method of calculating a pixel value of each pixel in the input image by using a graying algorithm to obtain the image.

The method comprises the steps of inputting a trained sample image containing an owner area label in the training process of a target detection network, extracting owner characteristics of the sample image through the network containing parameters, obtaining the parameters of the network through transfer learning, classifying according to the extracted owner characteristics, classifying sample identification results of the sample image containing an owner area range, calculating a loss value between the sample identification results and the owner area label related to the sample image, iteratively updating the parameters of the target detection network according to the loss value until the loss value reaches a convergence condition, stopping training when the convergence condition can be that the loss value is not changed any more or the iteration times reach preset times, and recording the converged target detection network as the trained target detection network.

In an embodiment, the vehicle sub-attribute recognition is performed on the vehicle snapshot image to obtain a second recognition result corresponding to the original data, and the method includes:

and carrying out license plate recognition on the vehicle snapshot image to obtain the license plate number of the vehicle snapshot image.

Understandably, the license plate recognition is a recognition process for recognizing the license plate number of the vehicle in the input vehicle snapshot image, and the license plate recognition process is as follows: firstly, the vehicle license plate positioning is carried out on the vehicle snapshot image, because the characters of the vehicle license plate and the background color matching behind the characters generally comprise several types of blue bottom white characters, yellow bottom black characters, white bottom red characters, green bottom white characters, black bottom white characters and the like, the region and the background can be obviously distinguished by utilizing different color channels (RGB channels, namely comprising an R channel, a G channel and a B channel), for example: for the license plate with the blue bottom and the white characters, when a blue channel (a channel B) is separated from a vehicle snapshot image, the license plate area is a bright rectangle, and characters in the license plate are not presented in the rectangular area, because blue (255, 0, 0) and white (255, 255, 255) in a color channel are not distinguished in the blue channel, similarly, a red channel (an R channel) can be used for the license plate with the white bottom and the black characters, the position of the license plate area can be obviously presented by using a green channel (a channel G), then edge extraction is carried out on the area position, the boundary part with obvious peripheral local brightness change at the area position is detected, the edge of the license plate can be detected, and the area of the license plate is positioned by enclosing the edge of the license plate; secondly, processing an image corresponding to the positioned license plate region by an image enhancement technology, such as a noise removal method and a contrast enhancement method, so that the text outline in the license plate region can be enhanced, and performing character segmentation on the image corresponding to the license plate region after the image enhancement technology to segment an independent character image; and finally, performing character normalization processing and character recognition on each individual character image, and converging the recognized characters into a license plate number.

Wherein, the character normalization process is to perform 0 to 1 normalization process on the character image labeled with the serial number, the character recognition can be a recognition process of firstly performing binarization on the character image and scaling the size of the character image into the size of a template in a character database, then matching with all templates, and finally selecting the best matching as a result, or a recognition process of performing feature extraction on the character image by using a character recognition network, then classifying by using an obtained feature to perform trained classifier, obtaining a recognition process of a result with the highest probability after classification, the training process of the character recognition network is to input the trained character image containing a character label, and perform character feature extraction on the character image by using the network containing parameters, the parameters of the network can be obtained by transfer learning, and classify according to the extracted character features, so as to classify the character recognition result of the corresponding character, calculating a loss value between the character recognition result and the character label associated with the character image, iteratively updating parameters of the character recognition network according to the loss value until the loss value reaches a convergence condition, wherein the convergence condition can be that the loss value is not changed any more or the iteration times reach preset times, stopping training, and recording the character recognition network after convergence as the character recognition network after training.

And carrying out vehicle type attribute identification on the vehicle snapshot image to obtain an attribute result of the vehicle snapshot image.

Understandably, the vehicle type attribute recognition is a recognition process of the vehicle type of the vehicle in the vehicle snapshot image, and the vehicle type attribute recognition process is as follows: extracting vehicle type characteristics from an input vehicle snapshot image through a trained vehicle type detection model, classifying according to the extracted vehicle type characteristics to classify probability values of various vehicle type categories, acquiring a vehicle type category corresponding to a maximum probability value as an attribute result, wherein the attribute result reflects the vehicle type category of the vehicle in the vehicle snapshot image, and the vehicle type characteristics are characteristics related to the vehicle type category, wherein the vehicle type category comprises categories such as taxis, buses, private cars, trucks, engineering vehicles and the like, a training process of a vehicle type detection network is to input a trained vehicle type image containing a vehicle type label, the vehicle type characteristics are extracted from the vehicle type image through the network containing parameters, the parameters of the network can be obtained through transfer learning, and are classified according to the extracted vehicle type characteristics to classify vehicle type recognition results of corresponding vehicle types, calculating a loss value between the vehicle type recognition result and a vehicle type label associated with the vehicle type image, iteratively updating parameters of the vehicle type detection network according to the loss value until the loss value reaches a convergence condition, wherein the convergence condition can be that the loss value is not changed any more or the iteration times reach preset times, stopping training, and recording the vehicle type detection network after convergence as a trained vehicle type detection network.

And recording the license plate number and the attribute result of the vehicle snapshot image as a second identification result.

Understandably, the license plate number and the attribute result of the vehicle snapshot image are determined as a second identification result, and the second identification result comprises the owner area image, the license plate number and the attribute result.

The license plate number of the vehicle snapshot image is obtained by carrying out license plate recognition on the vehicle snapshot image; carrying out vehicle type attribute identification on the vehicle snapshot image to obtain an attribute result of the vehicle snapshot image; the license plate number and the attribute result of the vehicle snap-shot image are recorded as a second recognition result, so that the license plate number and the attribute result containing the vehicle type can be automatically recognized by applying license plate recognition and vehicle type attribute recognition, the second recognition result recognized by the vehicle sub-attribute is finally determined, manual recognition is not needed, the labor cost is reduced, and the efficiency and the accuracy of vehicle sub-attribute recognition are improved.

S105, recording the first recognition result and the second recognition result corresponding to the raw data as image data.

Understandably, the image data corresponding to one original data includes the first recognition result and the second recognition result, so that after all the original data are traversed, the image data corresponding to each original data is obtained.

The invention realizes the purpose of acquiring a plurality of original data; for each original data, carrying out vehicle identification on the snapshot image in the original data to obtain a first identification result corresponding to the original data; based on the first identification results corresponding to the original data, vehicle screening is carried out on all the original data, and the original data containing the vehicles are screened out; extracting the snapshot image in the screened original data based on the vehicle area in the first identification result to obtain at least one vehicle snapshot image, and performing vehicle sub-attribute identification on the vehicle snapshot image to obtain a second identification result corresponding to the original data; the first identification result and the second identification result corresponding to the original data are recorded as the image data, so that whether the original data contain the vehicle or not can be automatically identified through vehicle identification, the original data containing the vehicle can be automatically screened out, the vehicle sub-attribute identification is carried out on the screened out original data, the second identification result of the vehicle attention attribute can be automatically identified, and the image data can be determined.

And S20, determining the vehicle number of the target vehicles of the specific vehicle types in the area to be monitored in the preset time period and the owner information of the target vehicles according to the plurality of image data.

Understandably, the specific vehicle type is a vehicle type or a vehicle attribute category for which an abnormal attribute rule is set in advance, such as: the method comprises the steps that a preset abnormal attribute rule is a license plate number of a certain area, such as the front license plate number of Yue A or Liao C and the like, or the preset abnormal attribute rule is a certain vehicle type, such as a taxi, a bus and the like, the process of determining a target vehicle of a specific vehicle type in a to-be-monitored area within a preset time period is to screen image data which accord with the specific vehicle type from all image data, determine the vehicle in the screened image data as the target vehicle, count the number of the target vehicle in the screened image data, and acquire vehicle owner information related to the license plate number of the target vehicle through the license plate number of the target vehicle.

In one embodiment, the step S20 of determining the number of vehicles of the target vehicle of the specific vehicle type present in the area to be monitored within the preset time period and the owner information of the target vehicle according to the plurality of image data includes:

acquiring a preset number of continuous historical time periods adjacent to the preset time period; each historical time period is the same time period as the preset time period.

Understandably, the preset number may be set according to the requirement, for example, the preset number may be 5 or 7, the historical time period may be the same time period as the preset time period, or may be one time period in a complete day of the historical record, for example: obtaining a preset number of consecutive historical time periods adjacent to the preset time period may be: the preset time period is a time period (from 14:00 to 14:59) from the current time, the preset number is 7, and the adjacent preset number of continuous historical time periods are 7 days (14:00 to 14:59) before the current day.

And carrying out average processing on the historical snapshot values associated with all the historical time periods to obtain a historical average value.

Understandably, each historical time period is associated with a historical snapshot value, the historical snapshot values are the number of vehicles snapshot in the historical time period, the historical snapshot values associated with the historical time periods are obtained for a historical segment, the averaging processing is the processing of averaging the obtained historical snapshot values, and the historical average value corresponding to the preset time period is obtained after the averaging processing.

The statistics of the historical snapshot values are statistics of the total number of the vehicle regions of the first recognition result in the image data corresponding to the monitoring time falling into the same historical time period.

And carrying out snapshot statistics on all image data to obtain snapshot values to be compared.

Understandably, the snapshot system is used for counting the total number of the vehicle areas of the first identification result in the image data corresponding to the monitoring time falling into the same preset time period, and recording the counting result as a snapshot value to be compared.

And judging whether the snapshot value to be compared is larger than the historical average value.

And when the snapshot value to be compared is larger than the historical mean value, screening all the image data according to the specific vehicle types to screen out the target image data corresponding to the target vehicle.

Understandably, if the snapshot value to be compared is larger than the historical mean value, the screening of the specific vehicle type is a screening process for screening out the image data corresponding to the target vehicle conforming to the specific vehicle type, the target vehicle conforming to the specific vehicle type can be determined through the screened image data, and the image data corresponding to the target vehicle is recorded as the target image data.

In one embodiment, the screening of all image data for a specific vehicle type to screen out target image data corresponding to a target vehicle includes:

and determining the superscalar value in the preset time period according to the snapshot value to be compared and the historical mean value.

Understandably, the sizes of the corresponding historical snapshot value and the snapshot value to be compared in the preset time period are judged, if the historical snapshot value is larger than or equal to the snapshot value to be compared, the superscript value in the preset time period is determined to be zero, if the historical snapshot value is smaller than the snapshot value to be compared, the difference value between the snapshot value to be compared and the historical snapshot value is calculated, and the ratio of the difference value and the historical snapshot value is determined to be the superscript value in the preset time period.

And if the exceeding value corresponding to the preset time period is greater than or equal to the preset threshold value, screening the image data corresponding to the exceeding value according to the specific vehicle type, and screening out the target image data corresponding to the target vehicle.

Understandably, if it is detected that the superscript value is greater than or equal to the preset threshold, the preset threshold may be set according to the requirement, for example, the preset threshold is set to 20%, which indicates that the snapshot value to be compared exceeds the floating standard of the historical snapshot value, and there is a gathering condition.

If the corresponding exceeding value in the preset time period is smaller than the preset threshold value, the image data corresponding to the exceeding value is not screened for the specific vehicle type, understandably, if the exceeding value is detected to be smaller than the preset threshold value, the condition of no aggregation exists.

Therefore, the image data with abnormal aggregation can be automatically output through the determination of the super-standard value and the judgment between the super-standard value and the preset threshold value, the specific abnormal vehicle type is screened, the target image data of the target vehicle is automatically screened, and the automatic screening of the image data with the target vehicle is realized without manual judgment.

In an embodiment, after determining whether the snapshot value to be compared is greater than the historical mean value, the method includes:

when the snapshot value to be compared is less than or equal to the historical mean value, determining an abnormal aggregation detection result of the area to be monitored without abnormal aggregation within a preset time period, understandably, if the snapshot value to be compared is less than or equal to the historical mean value, indicating that the area to be monitored does not have aggregation.

And counting the quantity of all the target image data to obtain the quantity of the vehicles, and acquiring the owner information of the target vehicles by acquiring the license plate number in each target image data.

Understandably, the quantity statistics is a process of counting the quantity of the target vehicles in all the target image data, in the counting process, the same target vehicle is subjected to duplicate removal, so that the obtained quantity of the vehicles is the total quantity of different target vehicles, the duplicate removal of the same target vehicle can be the duplicate removal of the same license plate number, the license plate number of the target vehicle in all the target image data is subjected to duplicate removal, a license plate number to be checked is obtained, owner information related to the license plate number to be checked is searched from an owner information base, and the searched owner information is related to the corresponding target vehicle, so that the owner information of each target vehicle is obtained.

The taxi owner information base stores the association relation between all license plate numbers and the information of the car owners, one license plate number can be associated with a plurality of pieces of car owner information, and the taxi can be matched with the car owners in two shifts in the day and night within a preset time period and then the car owners associated with the inquiry are determined.

The invention realizes that the historical time periods with continuous preset number adjacent to the preset time period are obtained; carrying out average processing on historical snapshot values associated with all historical time periods to obtain a historical average value; carrying out snapshot statistics on all image data to obtain snapshot values to be compared; judging whether the snapshot value to be compared is larger than the historical average value; when the snapshot value to be compared is larger than the historical mean value, screening all the image data of a specific vehicle type to screen out target image data corresponding to a target vehicle; the quantity statistics is carried out on all target image data to obtain the quantity of the vehicles, and the owner information of the target vehicles is obtained by obtaining the license plate number in each target image data.

And S30, determining the owner ratio value of the abnormal aggregation record based on the number of the vehicles, the owner information of the target vehicle and a preset abnormal aggregation blacklist.

Understandably, the abnormal aggregation blacklist is a list of people with history records of abnormal aggregation predecessors, abnormal aggregation potential risks or abnormal aggregation participation,

in one embodiment, the step S30 of determining the owner ratio value with the abnormal aggregation record based on the number of vehicles, the owner information of the target vehicle, and the preset abnormal aggregation blacklist includes:

and matching the owner information of the target vehicle with the abnormal aggregation blacklist to obtain the target vehicle with the abnormal aggregation record.

Understandably, matching the owner information of the target vehicle with each name in the abnormal aggregation blacklist by using a text matching algorithm, recording the target vehicle corresponding to the owner information matched with any name in the abnormal aggregation blacklist as the target vehicle with the abnormal aggregation record, calculating the complete matching degree between the owner information and each name in the abnormal aggregation blacklist by using the text matching algorithm, and determining the completely matched owner information and name as the owner information matched with the abnormal aggregation blacklist.

The text matching algorithm may be set according to requirements, for example, the text matching algorithm is BF, RK, KMP, BM, Sunday, or the like, and preferably, the text matching algorithm is KMP, which is also referred to as a fast string matching algorithm.

In one embodiment, the owner information includes an owner zone image; matching the owner information of the target vehicle with the abnormal aggregation blacklist to obtain the target vehicle with the abnormal aggregation record, wherein the method comprises the following steps:

and inputting the vehicle owner region image in the vehicle owner information into a blacklist face detection model, and extracting the face characteristics of the vehicle owner region image through the blacklist face detection model.

Understandably, the blacklist face detection model is a neural network model which is trained and finished through face images listed in an abnormal aggregation blacklist, the blacklist face detection model can identify whether the input image containing the face is the model of the face listed in the abnormal aggregation blacklist, each vehicle owner area image is input into the blacklist face detection model, the blacklist face detection model extracts the face characteristics of the vehicle owner area image to obtain a face characteristic vector to be compared, the extraction process is to perform gray processing on the two-dimensional three-channel vehicle owner area image, convert the vehicle owner area image into an image of one channel, perform convolution on the image of the channel, perform convolution through different convolution kernels to extract the face characteristics, and finally perform dimension reduction processing on the extracted characteristic vector to obtain a one-dimensional vector, for example: the end-to-end connection of the face features is converted into a column vector, and the vector is 400 dimensions assuming that the size of the image is 20 x 20.

The abnormal aggregation blacklist is a list of people with history records of abnormal aggregation predecessors, abnormal aggregation potential risks or abnormal aggregation, for example, the abnormal aggregation predecessors are fighting aggregation and the like, and the face features are features of the face of the abnormal aggregation blacklist.

And comparing the extracted face features with the face features of the persons contained in the preset abnormal aggregation blacklist to determine a comparison result.

Understandably, similarity comparison is carried out according to the extracted face features (face feature maps to be compared) and the face features (face feature vectors of people) of people included in the abnormal aggregation blacklist, the similarity comparison is difference comparison of the face features between two face feature vectors, a cosine similarity algorithm can be used for calculating the similarity between the two face feature vectors, corresponding probability values are mapped according to the similarity, so that the probability value corresponding to each person can be obtained, the maximum probability value in the probability values corresponding to all the people is obtained, and the maximum probability value is determined as a comparison result.

The cosine similarity algorithm is an algorithm for calculating a cosine value between two vector values as a similarity.

And obtaining the target vehicle with the abnormal aggregation record based on the comparison result.

Understandably, if the comparison result is greater than or equal to the preset probability threshold, it is determined that the target vehicle corresponding to the owner zone corresponding to the comparison result is the target vehicle with the abnormal aggregation record, and if the comparison result is less than the preset probability threshold, it is determined that the target vehicle corresponding to the owner zone corresponding to the comparison result is not the target vehicle with the abnormal aggregation record.

The invention realizes that the face characteristics of the vehicle owner region image are extracted through the blacklist face detection model by inputting the vehicle owner region image in the vehicle owner information into the blacklist face detection model; comparing the extracted face features with face features of persons contained in a preset abnormal aggregation blacklist to determine a comparison result; and obtaining the target vehicle with the abnormal aggregation record based on the comparison result, thus extracting the face characteristics of the vehicle owner in the vehicle owner region image through the blacklist face detection model, automatically comparing the extracted face characteristics with the face characteristics of the abnormal aggregation blacklist personnel to obtain the comparison result, automatically determining the target vehicle with the abnormal aggregation record, and improving the identification efficiency and accuracy of the target vehicle with the abnormal aggregation record.

And obtaining the vehicle owner occupation ratio based on the ratio of the total number of the target vehicles with the abnormal aggregation records to the number of the vehicles.

Understandably, the ratio of the total number of the target vehicles with the abnormal aggregation records to the number of the vehicles is calculated and is recorded as the vehicle owner occupation ratio.

The method and the device realize that the target vehicle with the abnormal aggregation record is obtained by matching the owner information of the target vehicle with the abnormal aggregation blacklist; the vehicle owner ratio is obtained based on the ratio of the total number of the target vehicles with the abnormal aggregation records to the number of the vehicles, so that the target vehicles with the abnormal aggregation records can be automatically matched through the abnormal aggregation blacklist, the vehicle owner ratio is obtained, the ratio of the target vehicles with the abnormal aggregation records to the number of the vehicles is automatically output, and a data basis is provided for subsequent abnormal aggregation detection.

And S40, determining the abnormal aggregation detection result of the area to be monitored in the preset time period according to the number of the vehicles and the vehicle owner ratio.

Understandably, the process of determining the abnormal cluster detection result may be: recording the difference value between the vehicle quantity and the preset aggregation quantity as an aggregation quantity index, acquiring an abnormal aggregation level value corresponding to the aggregation quantity index from an aggregation quantity index-abnormal aggregation level value comparison table, mapping an abnormal aggregation level value corresponding to the owner ratio from an owner ratio-abnormal aggregation level value mapping table according to the owner ratio, recording the maximum abnormal aggregation level value or the average value/sum of the maximum abnormal aggregation level value and the abnormal aggregation level value corresponding to the owner ratio as an abnormal aggregation detection result, wherein the abnormal aggregation detection result represents the level value of the abnormal aggregation degree represented by an area to be monitored in a preset time period, the higher the level value is, the higher the abnormal aggregation degree is, and the level range falls into according to different level values, and determining corresponding different color early warnings of high, medium and low so as to perform further related transactions, such as alarming or patrolling and the like.

The process of determining the abnormal aggregation detection result may further include: multiplying the number of the vehicles by the ratio of the number of the vehicle owners to obtain the number of the blacklists, subtracting the number of the blacklists from the preset number to obtain an early warning value if the number of the blacklists is larger than the preset number, determining a grade value of the abnormal aggregation detection result based on the early warning value, namely mapping the corresponding grade value of the abnormal aggregation degree according to the range in which the early warning value falls, and determining that the abnormal aggregation detection result is risk-free if the number of the blacklists is smaller than or equal to the preset number and the obtained early warning value is zero.

Wherein the process of determining the anomalous aggregation detection result further comprises: the vehicle number and the vehicle owner ratio value are weighted and summed to obtain the number of the prediction blacklists, the weights of the vehicle number and the vehicle owner ratio value can be set according to requirements, and the sum of the weights of the vehicle number and the vehicle owner ratio value is one, such as: the weight of the number of the vehicles is 0.4, the weight of the vehicle owner ratio is 0.6, if the number of the predicted blacklists is larger than the preset number, the number of the predicted blacklists is subtracted from the preset number to obtain an early warning value, the grade value of the abnormal aggregation detection result is determined based on the early warning value, namely the corresponding grade value of the abnormal aggregation degree is mapped according to the range in which the early warning value falls, if the number of the predicted blacklists is smaller than or equal to the preset number, the obtained early warning value is zero, and the abnormal aggregation detection result is determined to be risk-free.

The method and the device realize that a plurality of image data obtained by shooting in the area to be monitored in the preset time period are obtained; determining the number of target vehicles of a specific vehicle type in a region to be monitored within a preset time period and owner information of the target vehicles according to the plurality of image data; determining the owner ratio of the abnormal aggregation records based on the number of vehicles, the owner information of the target vehicle and a preset abnormal aggregation blacklist; according to the number of the vehicles and the ratio of the vehicle owners, the abnormal aggregation detection result of the area to be monitored in the preset time period is determined, therefore, the invention automatically determines the vehicle number and the vehicle owner information of the target vehicle with the specific vehicle type from all the image data through the image data shot in the area to be monitored in the preset time period, automatically determines the vehicle owner ratio based on the abnormal aggregation blacklist, integrates the vehicle number and the vehicle owner ratio, and automatically outputs the abnormal aggregation detection result of the area to be monitored, thereby obtaining the image data from the massive shot data, determining the vehicle number and the vehicle owner information of the target vehicle with the specific vehicle type, objectively, accurately and automatically identifying whether the abnormal aggregation detection result of the area to be monitored has the abnormal aggregation condition based on the abnormal aggregation blacklist, and not needing to manually identify whether the image data has the aggregation condition, the labor cost and the energy are greatly saved, and the accuracy and the correctness of monitoring the abnormal gathering of the vehicles are improved.

In an embodiment, in step S40, the determining the abnormal aggregation detection result of the area to be monitored in the preset time period according to the number of vehicles and the vehicle owner ratio includes:

and determining an aggregation result according to the number of the vehicles and the vehicle owner ratio.

Understandably, the determination process of the aggregated result may be: carry out the weighted summation with vehicle quantity and car owner ratio value, obtain the gathering result, the weight of vehicle quantity and car owner ratio value can be set for according to the demand, the weight of vehicle quantity and car owner ratio value is all not greater than 1, the weight sum of the weight of vehicle quantity and the weight sum of car owner ratio value is not more than 0.9, for example: the weight of the number of vehicles is 0.4, the weight of the owner ratio is 0.3, and the determination process of the aggregation result may also be: and multiplying the number of the vehicles by the ratio of the number of the vehicle owners to obtain an aggregation result.

And acquiring a risk level associated with the area to be monitored, and determining an abnormal aggregation detection result according to the aggregation result and the risk level.

Understandably, a risk level is associated with one to-be-monitored area, the risk level represents the risk degree of aggregation of the monitored area, the risk level is continuously updated along with the time, for example, the number of times of abnormal aggregation detection results with abnormal aggregation is accumulated to determine the risk level, that is, the risk level of the monitored area is analyzed along with the relevant historical data of the monitored area, the risk level associated with the to-be-monitored area is obtained, and according to the aggregation result and the risk level, the process of determining the abnormal aggregation detection results may be: and performing weighted summation on the aggregation result and the risk level to obtain an abnormal aggregation detection result, wherein the weight of the aggregation result is one, and the sum of the weight of the risk level, the weight of the number of the vehicles and the weight of the proportion of the owners is one.

The invention realizes that the aggregation result is determined according to the number of the vehicles and the vehicle owner ratio; the risk grade associated with the area to be monitored is obtained, and the abnormal aggregation detection result is determined according to the aggregation result and the risk grade, so that the risk grade of the area to be monitored can be introduced to comprehensively output the abnormal aggregation detection result, the abnormal aggregation result whether the area to be monitored exists in the current preset time period can be reflected more accurately and objectively, and the accuracy of monitoring the abnormal aggregation of the vehicle is improved.

For each aggregate data, multiplying and summing the risk level value, the flow value and the blacklist ratio value corresponding to the aggregate data by the respective preset weight, wherein the sum of all the preset weights is 1, so as to obtain an abnormal aggregate value of the aggregate data, for example, the weight corresponding to the risk level value is 0.5, the weight corresponding to the flow value is 0.2, the weight corresponding to the blacklist ratio value is 0.3, the abnormal early warning is to map out a corresponding grade according to the abnormal aggregate value, for example, the value in the range of 0.1 to 0.5 corresponds to the low grade, the value in the range of 0.5 to 0.9 corresponds to the middle grade, and the like, classifying and summarizing all the abnormally aggregated grades to obtain a vehicle abnormal aggregate monitoring result, and rapidly displaying the corresponding monitoring area identifier on a visual interface through the vehicle abnormal aggregate monitoring result, or displaying different colors on the visual interface through different grades, for further monitoring or for executing related transactions.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

In one embodiment, a vehicle abnormality aggregation monitoring device is provided, which corresponds one-to-one to the vehicle abnormality aggregation monitoring methods in the above-described embodiments. As shown in fig. 4, the vehicle abnormality gathering monitoring apparatus includes an acquisition module 11, a statistic module 12, a determination module 13, and an output module 14. The functional modules are explained in detail as follows:

the acquisition module 11 is configured to acquire a plurality of image data captured in a preset time period in an area to be monitored;

the statistical module 12 is configured to determine, according to the plurality of image data, the number of vehicles of a target vehicle of a specific vehicle type appearing in an area to be monitored within a preset time period and owner information of the target vehicle;

the determining module 13 is configured to determine an owner ratio value with an abnormal aggregation record based on the number of vehicles, owner information of the target vehicle, and a preset abnormal aggregation blacklist;

and the output module 14 is configured to determine an abnormal aggregation detection result for the area to be monitored within a preset time period according to the number of the vehicles and the vehicle owner ratio.

For specific limitations of the vehicle abnormal aggregation monitoring apparatus, reference may be made to the above limitations of the vehicle abnormal aggregation monitoring method, which are not described herein again. The respective modules in the vehicle abnormality gathering monitoring apparatus described above may be implemented in whole or in part by software, hardware, and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a client or a server, and its internal structure diagram may be as shown in fig. 5. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a readable storage medium and an internal memory. The readable storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the readable storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a vehicle anomaly aggregation monitoring method.

In one embodiment, a computer device is provided, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the vehicle abnormality gathering monitoring method in the above embodiments is implemented.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the vehicle abnormality aggregation monitoring method in the above-described embodiments.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, databases, or other media used in embodiments provided herein may include non-volatile and/or volatile memory. 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 a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.

Claims

1. A vehicle abnormal aggregation monitoring method, characterized by comprising:

determining the number of target vehicles of a specific vehicle type appearing in an area to be monitored within a preset time period and owner information of the target vehicles according to the plurality of image data;

2. The vehicle abnormal aggregation monitoring method according to claim 1, wherein the determining an abnormal aggregation detection result for the area to be monitored within a preset time period according to the number of vehicles and the vehicle owner ratio includes:

determining an aggregation result according to the number of the vehicles and the vehicle owner ratio;

and acquiring a risk grade associated with the area to be monitored, and determining the abnormal aggregation detection result according to the aggregation result and the risk grade.

3. The vehicle abnormal collection monitoring method according to claim 1, wherein the acquiring of the plurality of image data captured in the area to be monitored within the preset time period includes:

acquiring a plurality of original data;

for each original data, carrying out vehicle identification on the snapshot image in the original data to obtain a first identification result corresponding to the original data; the first recognition result shows whether the snapshot image contains a vehicle or not;

based on the first identification result corresponding to each original data, vehicle screening is carried out on all the original data, and the original data containing vehicles are screened out;

extracting the snapshot image in the screened original data based on the vehicle area in the first identification result to obtain at least one vehicle snapshot image, and performing vehicle sub-attribute identification on the vehicle snapshot image to obtain a second identification result corresponding to the original data;

recording the first recognition result and the second recognition result corresponding to the original data as the image data.

4. The vehicle abnormal gathering monitoring method as claimed in claim 3, wherein the vehicle sub-attribute recognition of the vehicle snap-shot image to obtain a second recognition result corresponding to the original data comprises:

carrying out license plate recognition on the vehicle snapshot image to obtain a license plate number of the vehicle snapshot image;

carrying out vehicle type attribute identification on the vehicle snapshot image to obtain an attribute result of the vehicle snapshot image;

and recording the license plate number and the attribute result of the vehicle snapshot image as the second identification result.

5. The vehicle abnormal collection monitoring method according to claim 4, wherein the determining of the number of the target vehicles of the specific vehicle type appearing in the area to be monitored within the preset time period and the owner information of the target vehicles according to the plurality of image data comprises:

acquiring a preset number of continuous historical time periods adjacent to the preset time period; each historical time period and the preset time period are the same time period;

carrying out average processing on the historical snapshot values associated with all the historical time periods to obtain a historical average value;

carrying out snapshot statistics on all the image data to obtain snapshot values to be compared;

judging whether the snapshot value to be compared is larger than the historical mean value or not;

when the snapshot value to be compared is larger than the historical mean value, screening all the image data of a specific vehicle type to screen out target image data corresponding to a target vehicle;

counting the quantity of all the target image data to obtain the quantity of the vehicles, and acquiring the owner information of the license plate in each target image data to obtain the owner information of the target vehicle.

6. The vehicle abnormal aggregation monitoring method according to claim 5, wherein the determining of the vehicle owner ratio value in which the abnormal aggregation record exists based on the number of the vehicles, the vehicle owner information of the target vehicle, and a preset abnormal aggregation blacklist includes:

matching the owner information of the target vehicle with the abnormal aggregation blacklist to obtain the target vehicle with the abnormal aggregation record;

7. The vehicle abnormality gathering monitoring method according to claim 6, characterized in that the vehicle owner information includes a vehicle owner zone image; the matching the owner information of the target vehicle with the abnormal aggregation blacklist to obtain the target vehicle with the abnormal aggregation record includes:

inputting the vehicle owner region image in the vehicle owner information into a blacklist face detection model, and extracting the face characteristics of the vehicle owner region image through the blacklist face detection model;

comparing the extracted face features with face features of persons contained in a preset abnormal aggregation blacklist to determine a comparison result;

8. An abnormal aggregation monitoring apparatus for a vehicle, comprising:

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the vehicle anomaly aggregation monitoring method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the vehicle abnormality aggregation monitoring method according to any one of claims 1 to 7.