An intelligent compression method, system and device for AI law enforcement vehicle recognition events

Through the intelligent compression method of AI law enforcement vehicle identification events, GPS and azimuth data are used to make intelligent judgments, which solves the problem of inefficient image target recognition in urban management, and realizes efficient processing of urban management pictures and event recognition.

CN118762240BActive Publication Date: 2024-11-22BEIJING HUIMENG XINAN SOFTWARE TECH CO LTD +1
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
CN202411244145.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-06
Publication Date
2024-11-22
Estimated Expiration
2044-09-06

AI Technical Summary

Technical Problem

The prior art is difficult to quickly identify the differences between the two urban management pictures in the front and rear, resulting in inefficient image target recognition in urban management and unable to meet the needs of efficiently processing urban image data.

Method used

Using the intelligent compression method of AI law enforcement vehicles to identify events, the GPS latitude and longitude information and azimuth data of urban management images are obtained in real time, and the neural network model is input to output event information, and intelligent judgment is made using latitude and longitude distance, azimuth difference and IOU overlap. , realize intelligent compression processing of the same events.

Benefits of technology

Effectively identifying various target subjects and their corresponding events in multiple pictures improves the image processing efficiency and management level in urban management, and solves the problem of inefficient processing of target events in multiple pictures.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118762240B_ABST
    Figure CN118762240B_ABST
Patent Text Reader

Abstract

The present application relates to the field of urban management image recognition and classification, and in particular to an intelligent compression method, system and device for AI law enforcement vehicles to identify events. The technical solution provided by the present application uses an AI law enforcement vehicle to collect pictures, analyze the GPS, azimuth, IOU and other data of the pictures, make intelligent judgments on the identified events, and intelligently compress the same events. The method is based on the analysis of urban governance picture data, effectively identifies various target subjects in multiple pictures and their corresponding events, and designs an event intelligent compression and automatic elimination method. The method proposed in the invention is based on an AI law enforcement vehicle and an on-board computing box, which can effectively solve the problem of inefficiency in the target event processing method of multiple pictures in the process of urban governance, and effectively improve the level and efficiency of urban management.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of urban management image recognition and classification, and in particular to an intelligent compression method, system and device for AI law enforcement vehicle recognition events. Background Art

[0002] Image recognition technology in urban management is an important means to improve the efficiency of urban case discovery and handling and save urban human and material resources. However, in practical application, this technology faces some difficulties and challenges, including the following aspects.

[0003] 1. Complexity of urban scale and structure: With the increasing expansion of urban scale and the increasing complexity of urban structure, urban governance is becoming increasingly difficult. This means that when extracting images related to urban governance, it is necessary to consider multi-dimensional information, including urban infrastructure, residents' lives, environmental conditions, etc., which increases the complexity of image extraction.

[0004] 2. Multi-dimensional requirements of urban governance: Urban governance involves not only the efficiency of urban operation at the macro level, but also the happiness of residents at the micro level. This requires that image extraction can reflect the "temperature", "precision" and intelligence of urban governance, which increases the diversity and depth of image content.

[0005] 3. Challenges brought by rapid urbanization: Problems such as "big city disease" that arise during rapid urbanization have put forward higher requirements for urban governance. This requires that image extraction should not only focus on the current state of the city, but also be able to reflect the problems and challenges that arise during urban development.

[0006] 4. Algorithm island problem: Traditional urban governance AI applications drive algorithms based on specific events, forming algorithm islands. As a result, the same batch of data needs to be repeatedly calculated using different algorithms, which increases the complexity and technical cost of image processing.

[0007] 5. Backward technical means: In some urban governance practices, the level of informatization is low, manual identification is difficult, and the traditional method of relying on manual inspections is prone to misjudgment and missed judgments, and lacks efficient early warning means and rapid response capabilities.

[0008] 6. Challenges of data collection and analysis: Although modern technology provides a way to collect a large amount of image data, how to effectively use this data, develop a standardized method framework to extract and understand visual information, and how to establish a methodology to quantify the material environment of urban space are still issues to be resolved.

[0009] 7. Complexity of refined governance: Refined urban governance requires more accurate reading of the status and details of social facts and the design of more refined management and service items, which places higher demands on the accuracy and intelligence level of image processing. In order to understand the development status of existing technologies, we searched, compared and analyzed existing papers and patents, and screened out the following technical information that is highly relevant to this application:

[0010] Technical Solution 1: The patent number CN202010816260.1, "A Smart City Security Monitoring System Based on Image Recognition", discloses a smart city security monitoring system based on image recognition, including a follow-up monitoring and analysis device and a self-starting device; the self-starting device includes a wall and a clamping assembly arranged at a preset height of the wall, and the clamping assembly clamps a bullet-receiving cloth; the clamping assembly includes a base plate fixed to the wall through a bolt hole, the bolt hole is opened on the wall, and a corresponding through hole is opened on the base plate, and a nut is passed through the through hole to connect to the bolt hole; the present application monitors the content of the high-rise building through a follow-up monitoring and analysis device, and collects images with the help of a camera. After analyzing the image, when an object falls, the corresponding threat point is automatically obtained according to the position of the threat point, and the self-starting device is automatically combined with the drive unit 1 and the drive unit 2 to perform protection and self-start, and the drive unit 1 is automatically controlled to drive the rotation, and the rotating column drives the front support rod to rotate; the front support rod opens the bullet-receiving cloth to play a protective role.

[0011] Technical Solution 2: Patent No. CN 201310222110, "Image Acquisition Method for Spatial Precision Positioning for Landscape Planning", discloses an image acquisition method for spatial precision positioning for landscape planning. This method is based on the existing open scenic spots that are difficult to accurately locate, and proposes a method based on GPRS fixed point. First, the location of the observation point is determined by GPRS positioning, and then the grid image is acquired for the entire open area. Finally, the image and coordinates of each point are input into the computer for linking, so that tourists can independently determine the scope and route of the tour. This application adopts quantitative survey, statistics, and calculation methods, and has objective and rigorous characteristics. The image acquisition technology based on GPS spatial precision positioning described in this application promotes the comprehensive display of high-quality tourist landscapes, provides tourists with the convenience of completely independently choosing routes, has good social, economic and ecological benefits, and provides a rational and scientific basis for urban cruise line planning, urban tourism planning and related construction, and more efficiently and reasonably utilizes precious urban tourism resources.

[0012] The method provided by Technical Solution 1 recognizes and processes smart city intelligent security pictures, and identifies events in the pictures as targets. The method provided by Technical Solution 2 processes urban grid pictures through GPRS positioning. The above two methods do not consider the rapid identification of the difference between the previous and the next pictures, and are not suitable for the requirements of high efficiency of image target recognition in the scenario of urban management. There are technical defects in processing urban image data. Summary of the invention

[0013] In view of this, the purpose of this application is to provide an intelligent compression method, system and equipment for AI law enforcement vehicle recognition events, so as to solve the problem that the prior art does not consider the rapid identification of the difference between the previous and next pictures, is not suitable for the requirements of high-efficiency image target recognition in urban management scenarios, and has technical defects in processing urban image data.

[0014] According to a first aspect of an embodiment of the present application, there is provided an intelligent compression method for AI law enforcement vehicle recognition events, comprising:

[0015] Acquire the city management images to be processed in real time according to a preset frequency;

[0016] Obtaining GPS longitude and latitude information in the adjacent city management image to be processed and azimuth data information of the adjacent city management image to be processed;

[0017] Inputting the GPS longitude and latitude information in the adjacent city management image to be processed, the azimuth data information of the adjacent city management image to be processed, and the adjacent city management image to be processed into a preset neural network model, and outputting all event information in the adjacent city management image to be processed;

[0018] If there are events of the same category in all event information in the adjacent city management images to be processed, the preset latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into the preset event intelligent judgment model to obtain a first processing result;

[0019] If the first processing result meets the preset standard value, it indicates that the same event information exists in all the event information in the adjacent urban management images to be processed, and then the same event information is intelligently compressed to obtain the processing result of intelligent recognition and compression of the urban management image.

[0020] Further, the obtaining of GPS longitude and latitude information in the adjacent city management image to be processed and azimuth data information of the adjacent city management image to be processed includes:

[0021] Using a preset Beidou chip to obtain GPS longitude and latitude information in the adjacent city management images to be processed, marking the GPS longitude and latitude information in one of the city management images to be processed as P1, and marking the GPS longitude and latitude information in the adjacent city management image to be processed as P2;

[0022] Using P1 and P2, a preset first formula is used to obtain the y-axis coordinate and the x-axis coordinate in the two-dimensional coordinate system;

[0023] The first formula is as follows:

[0024] (1)

[0025] Among them, lat represents the latitude in GPS longitude and latitude information, and long represents the longitude in GPS longitude and latitude information;

[0026] Input the obtained y-axis coordinate and x-axis coordinate into a preset second formula to obtain the azimuth data information of the adjacent city management image to be processed;

[0027] The second formula is as follows:

[0028] Bearing(P1) = arctan(y,x)(2)

[0029] Wherein, the Bearing (P1) represents the azimuth data information of the adjacent city management image to be processed.

[0030] Furthermore, if there are events of the same category in all event information in the adjacent city management images to be processed, the preset latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into the preset event intelligent judgment model to obtain a first processing result, including:

[0031] If there are events of the same category in all event information in the adjacent city management images to be processed, then obtaining the latitude and longitude distance information, azimuth angle difference information, and IOU overlap in the adjacent city management images to be processed;

[0032] The acquired latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into a preset event intelligent judgment model to obtain a first processing result.

[0033] Further, obtaining the latitude and longitude distance information in the adjacent city management image to be processed includes:

[0034] Using the latitude and longitude information in the adjacent city management image to be processed, and using a preset GPS latitude and longitude distance formula, the latitude and longitude distance information in the adjacent city management image to be processed is obtained;

[0035] The GPS latitude and longitude distance formula includes:

[0036] (3)

[0037] Among them, lat represents the latitude in the GPS longitude and latitude information, and long represents the longitude in the GPS longitude and latitude information; lat1 and long1 represent the latitude and longitude information in P1; lat2 and long2 represent the latitude and longitude information in P2.

[0038] Further, obtaining the azimuth difference information in the adjacent city management images to be processed includes:

[0039] Obtaining the azimuth angles of the city management images to be processed that are adjacent to each other, marking the azimuth angle of one of the city management images to be processed as Bearing (P1), and marking the azimuth angle of the city management image to be processed that is adjacent to P1 as Bearing (P2);

[0040] The absolute value of the difference between the azimuth angles Bearing (P1) and Bearing (P2) of the adjacent city management images to be processed is taken to obtain the longitude and latitude distance information of the adjacent city management images to be processed.

[0041] Further, obtaining the IOU overlap in the adjacent city management images to be processed includes:

[0042] Based on events of the same category, the location information of events in adjacent urban management images to be processed is respectively marked;

[0043] The position information is used to identify the same position and calculate the IOU overlap of adjacent urban management images to be processed.

[0044] Furthermore, if the first processing result meets the preset standard value, it indicates that the same event information exists in all event information in the adjacent city management images to be processed, including:

[0045] If the latitude and longitude distance information, azimuth difference information, and IOU overlap of the adjacent city management images to be processed in the first processing result are all less than their respective preset thresholds, it indicates that the same event information exists in all event information of the adjacent city management images to be processed.

[0046] According to a second aspect of an embodiment of the present application, there is provided an intelligent compression system for AI law enforcement vehicle recognition events, comprising:

[0047] An acquisition module, used for acquiring the city management images to be processed in real time according to a preset frequency;

[0048] The first processing module is used to obtain GPS longitude and latitude information in the adjacent city management image to be processed and azimuth data information of the adjacent city management image to be processed;

[0049] The second processing module is used to input the GPS longitude and latitude information in the adjacent city management image to be processed, the azimuth data information of the adjacent city management image to be processed, and the adjacent city management image to be processed into a preset neural network model, and output all event information in the adjacent city management image to be processed;

[0050] The third processing module is used for inputting the preset latitude and longitude distance information, azimuth difference information, and IOU overlap of the adjacent city management images to be processed into a preset event intelligent judgment model to obtain a first processing result if there are events of the same category in all event information in the adjacent city management images to be processed;

[0051] The fourth processing module is used to perform intelligent compression processing on the same event information if the first processing result meets the preset standard value, indicating that the same event information exists in all the event information in the adjacent urban management images to be processed, so as to obtain the processing result of intelligent recognition and compression of the urban management image.

[0052] According to a third aspect of an embodiment of the present application, there is provided an intelligent compression device for AI law enforcement vehicle recognition events, including:

[0053] a memory having an executable program stored therein;

[0054] A processor is used to execute the executable program in the memory to implement the steps of any one of the above methods.

[0055] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, including:

[0056] The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the steps of any one of the above methods.

[0057] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0058] It can be understood that the technical solution provided by the present application is to acquire the city management image to be processed in real time according to a preset frequency; obtain the GPS longitude and latitude information and the azimuth data information of the adjacent city management image to be processed; then, input the GPS longitude and latitude information, the azimuth data information of the adjacent city management image to be processed and the adjacent city management image to be processed into a preset neural network model, and output all event information of the adjacent city management image to be processed; if the same category of events exists in all event information of the adjacent city management images to be processed, then input the preset longitude and latitude distance information, azimuth difference information and IOU overlap of the adjacent city management images to be processed into a preset event intelligent judgment model to obtain a first processing result; finally, if the first processing result meets the preset standard value, it indicates that the same event information exists in all event information of the adjacent city management images to be processed, then the same event information is intelligently compressed to obtain the processing result of intelligent recognition and compression of the city management image. It is understandable that the technical solution provided by this application uses an AI law enforcement vehicle to collect pictures, analyzes the GPS, azimuth, IOU and other data of the pictures, makes intelligent judgments on the identified events, and intelligently compresses the same events. This method is based on the analysis of urban governance picture data, effectively identifies various target subjects and their corresponding events in multiple pictures, and designs an event intelligent compression and automatic elimination method. The method proposed in this invention is based on an AI law enforcement vehicle and an on-board computing box, which can effectively solve the problem of inefficiency in the method of processing target events in multiple pictures during urban governance, and effectively improve the level and efficiency of urban management.

[0059] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0061] Figure 1 It is a schematic diagram of the steps of an intelligent compression method for an AI law enforcement vehicle to identify an event according to an exemplary embodiment;

[0062] Figure 2 is a schematic diagram showing the composition of GPS latitude, longitude and azimuth of an image according to an exemplary embodiment;

[0063] Figure 3It is a schematic diagram of the structure of an intelligent compression system for AI law enforcement vehicle recognition events according to an exemplary embodiment;

[0064] Figure 4 It is a schematic diagram of the structural composition of an intelligent compression device for AI law enforcement vehicle recognition events according to an exemplary embodiment. DETAILED DESCRIPTION

[0065] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0066] Embodiment 1

[0067] See also Figure 1 , Figure 1 The figure is a schematic diagram of steps of an intelligent compression method for an AI law enforcement vehicle to identify an event according to an exemplary embodiment, the method comprising:

[0068] S1. Obtain the city management image to be processed in real time at a preset frequency;

[0069] S2. Obtaining GPS latitude and longitude information of the adjacent city management image to be processed, and azimuth data information of the adjacent city management image to be processed;

[0070] S3. The GPS longitude and latitude information of the adjacent city management image to be processed, the azimuth data information of the adjacent city management image to be processed, and the adjacent city management image to be processed are input into a preset neural network model, and all event information of the adjacent city management image to be processed is output;

[0071] S4. If there are events of the same category in all event information in the adjacent city management images to be processed, the preset latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into the preset event intelligent judgment model to obtain a first processing result;

[0072] S5. If the first processing result meets the preset standard value, it indicates that the same event information exists in all the event information in the adjacent urban management images to be processed, and then the same event information is intelligently compressed to obtain the processing result of intelligent recognition and compression of the urban management image.

[0073] As described in steps S1-S2, in the specific implementation, the AI ​​law enforcement vehicle is equipped with a camera and an on-board edge computing box, which has computing power and is connected to the city governance cloud platform. The AI ​​law enforcement vehicle travels at a speed of 30km / h. The system extracts pictures from the camera mounted on the law enforcement vehicle every 0.1 seconds, and simultaneously obtains the GPS longitude and latitude of the picture through the Beidou chip, and calculates the GPS azimuth Bearing (P) of the front and back pictures in real time through the longitude and latitude, and transmits the picture, GPS longitude and latitude, and azimuth data to the analysis module of the on-board edge computing box.

[0074] Further, obtaining GPS latitude and longitude information in the adjacent city management image to be processed and azimuth data information of the adjacent city management image to be processed includes:

[0075] Using a preset Beidou chip to obtain GPS longitude and latitude information in the adjacent city management images to be processed, marking the GPS longitude and latitude information in one of the city management images to be processed as P1, and marking the GPS longitude and latitude information in the adjacent city management image to be processed as P2;

[0076] Using P1 and P2, a preset first formula is used to obtain the y-axis coordinate and the x-axis coordinate in the two-dimensional coordinate system;

[0077] The first formula is as follows:

[0078] (1)

[0079] Among them, lat represents the latitude in GPS longitude and latitude information, and long represents the longitude in GPS longitude and latitude information;

[0080] Input the obtained y-axis coordinate and x-axis coordinate into a preset second formula to obtain the azimuth data information of the adjacent city management image to be processed;

[0081] The second formula is as follows:

[0082] Bearing(P1) = arctan(y,x)(2)

[0083] Wherein, the Bearing (P1) represents the azimuth data information of the adjacent city management image to be processed.

[0084] For specific implementation, please refer to Figure 2 , Figure 2 The coordinates (latitude and longitude) of three GPS points and the azimuth between them are shown in Figure 1. The details are as follows:

[0085] 1.GPS Point

[0086] P1: The first GPS point, with coordinates (latitude, longitude).

[0087] P2: The second GPS point, with coordinates (latitude, longitude).

[0088] P3: The third GPS point, with coordinates (latitude, longitude).

[0089] 2. Bearing

[0090] Azimuth is the direction from one GPS point to the next.

[0091] The azimuth from P1 to P2 is marked as Bearing(P1).

[0092] The azimuth from P2 to P3 is marked as Bearing(P2).

[0093] 3. Calculation of the azimuth angle Bearing (P1)

[0094] First calculate y and x, refer to formula (1);

[0095] To calculate the azimuth angle from y and x, refer to formula (2).

[0096] In one embodiment, as described in step S3, the GPS longitude and latitude information in the adjacent city management image to be processed, the azimuth data information of the adjacent city management image to be processed, and the adjacent city management image to be processed are input into a preset neural network model, and all event information in the adjacent city management image to be processed is output, including:

[0097] The GPS longitude and latitude information in the adjacent city management image to be processed, the azimuth data information of the adjacent city management image to be processed, and the adjacent city management image to be processed are input into the preset neural network model, and the vehicle-mounted edge computing box analysis module is used to identify the events in the picture through the neural network. For example: 0001 indicates road cracks, 0002 indicates road potholes, etc.

[0098] In one embodiment, as described in step S4, if there are events of the same category in all event information in the adjacent city management images to be processed, the preset latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into the preset event intelligent judgment model to obtain a first processing result, including:

[0099] If there are events of the same category in all event information in the adjacent city management images to be processed, then obtaining the latitude and longitude distance information, azimuth angle difference information, and IOU overlap in the adjacent city management images to be processed;

[0100] The acquired latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into a preset event intelligent judgment model to obtain a first processing result.

[0101] Further, obtaining the latitude and longitude distance information in the adjacent city management image to be processed includes:

[0102] Using the latitude and longitude information in the adjacent city management image to be processed, and using a preset GPS latitude and longitude distance formula, the latitude and longitude distance information in the adjacent city management image to be processed is obtained;

[0103] The GPS latitude and longitude distance formula includes:

[0104] (3)

[0105] Among them, lat represents the latitude in the GPS longitude and latitude information, and long represents the longitude in the GPS longitude and latitude information; lat1 and long1 represent the latitude and longitude information in P1; lat2 and long2 represent the latitude and longitude information in P2.

[0106] Further, obtaining the azimuth difference information in the adjacent city management images to be processed includes:

[0107] Obtaining the azimuth angles of the city management images to be processed that are adjacent to each other, marking the azimuth angle of one of the city management images to be processed as Bearing (P1), and marking the azimuth angle of the city management image to be processed that is adjacent to P1 as Bearing (P2);

[0108] Taking the absolute value of the difference between the azimuth angles Bearing (P1) and Bearing (P2) of the adjacent city management images to be processed, and obtaining the longitude and latitude distance information of the adjacent city management images to be processed;

[0109] Specifically, the formula for taking the absolute value of the difference between the azimuth angles Bearing (P1) and Bearing (P2) of the adjacent city management images to be processed before and after the solution is as follows:

[0110] BR(P1) = |Bearing(P2) - Bearing(P1)|(4)

[0111] Further, obtaining the IOU overlap in the adjacent city management images to be processed includes:

[0112] Based on events of the same category, the location information of events in adjacent urban management images to be processed is respectively marked;

[0113] The position information is used to identify the same position and calculate the IOU overlap of adjacent urban management images to be processed.

[0114] In the specific implementation, the calculation method of the IOU overlap rate of the same type of event in the two pictures before and after is

[0115] 1. Event type: for example, 0001 means illegal umbrella holding;

[0116] 2. The position of the event in the picture: For example, if the picture size is 1920*1080, the coordinates of the upper left corner of the picture are x1=0, y1=0, and the coordinates of the lower right corner of the picture are x2=1920, y2=1080. The position of the illegal umbrella-holding event in the picture is x1, y1, x2, y2. These positions are all based on the coordinates of the upper left corner of the picture (0,0).

[0117] 3. Same location recognition method: For example, the coordinates of the illegal outdoor advertisement in the first picture are x11, y11, x12, y12, and the coordinates of the illegal outdoor advertisement in the second picture are x21, y21, x22, y22. Calculate the IOU overlap of these two rectangular boxes.

[0118] In the specific implementation, as described in step S5, if the first processing result meets the preset standard value, it indicates that the same event information exists in all the event information in the adjacent city management images to be processed, then the same event information is intelligently compressed to obtain the processing result of the intelligent recognition compression of the city management image, including:

[0119] If the latitude and longitude distance information, azimuth difference information, and IOU overlap of the adjacent city management images to be processed in the first processing result are all less than their respective preset thresholds, it indicates that the same event information exists in all event information of the adjacent city management images to be processed;

[0120] The same event information is subjected to intelligent compression processing to obtain the processing result of intelligent recognition compression of the urban management image.

[0121] In specific implementation, if the latitude and longitude distance < preset threshold D, the azimuth angle difference BR(P1) < preset threshold B, and the IOU overlap < preset threshold I, it means that the events in the previous and next two pictures are the same event, and event compression processing is performed.

[0122] In one embodiment, the technical solution provided by the present application obtains the city management image to be processed in real time according to a preset frequency; obtains the GPS longitude and latitude information and the azimuth data information of the adjacent city management image to be processed; then, inputs the GPS longitude and latitude information, the azimuth data information of the adjacent city management image to be processed and the adjacent city management image to be processed into a preset neural network model, and outputs all event information of the adjacent city management image to be processed; if there are events of the same category in all event information of the adjacent city management images to be processed, then inputs the preset longitude and latitude distance information, azimuth difference information and IOU overlap of the adjacent city management images to be processed into a preset event intelligent judgment model to obtain a first processing result; finally, if the first processing result meets the preset standard value, it indicates that there is the same event information in all event information of the adjacent city management images to be processed, then the same event information is intelligently compressed to obtain the processing result of intelligent recognition compression of the city management image. It is understandable that the technical solution provided by this application uses an AI law enforcement vehicle to collect pictures, analyzes the GPS, azimuth, IOU and other data of the pictures, makes intelligent judgments on the identified events, and intelligently compresses the same events. This method is based on the analysis of urban governance picture data, effectively identifies various target subjects and their corresponding events in multiple pictures, and designs an event intelligent compression and automatic elimination method. The method proposed in this invention is based on an AI law enforcement vehicle and an on-board computing box, which can effectively solve the problem of inefficiency in the method of processing target events in multiple pictures during urban governance, and effectively improve the level and efficiency of urban management.

[0123] More specifically, in order to solve the problem of quickly identifying image targets in urban management, this application obtains urban management images from fixed cameras in the city, and then identifies illegal outdoor advertising, illegal umbrella holding, road damage and other events in the images. The image processing method of multiple image reasoning proposed in this article can automatically compress or clear the events identified in the image, thereby effectively improving the target recognition rate in urban management images and realizing the rapid processing of urban management events.

[0124] See also Figure 3 , Figure 3 The figure is a schematic diagram showing the structure of an intelligent compression system for AI law enforcement vehicle recognition events according to an exemplary embodiment, wherein the system comprises:

[0125] An acquisition module 10 is used to acquire the city management image to be processed in real time according to a preset frequency;

[0126] The first processing module 20 is used to obtain GPS longitude and latitude information in the adjacent city management image to be processed and azimuth data information of the adjacent city management image to be processed;

[0127] The second processing module 30 is used to input the GPS longitude and latitude information in the adjacent city management image to be processed, the azimuth data information of the adjacent city management image to be processed, and the adjacent city management image to be processed into a preset neural network model, and output all event information in the adjacent city management image to be processed;

[0128] The third processing module 40 is used to input the preset latitude and longitude distance information, azimuth difference information, and IOU overlap of the adjacent city management images to be processed into a preset event intelligent judgment model to obtain a first processing result if there are events of the same category in all event information in the adjacent city management images to be processed;

[0129] The fourth processing module 50 is used to perform intelligent compression processing on the same event information if the first processing result meets the preset standard value, indicating that the same event information exists in all the event information in the adjacent urban management images to be processed, so as to obtain the processing result of intelligent recognition compression of the urban management image.

[0130] See also Figure 4 , Figure 4 The figure is a schematic diagram showing the structure of an intelligent compression device for AI law enforcement vehicle recognition events according to an exemplary embodiment, wherein the device comprises:

[0131] A memory 41 on which an executable program is stored;

[0132] The processor 42 is used to execute the executable program in the memory 41 to implement the steps of any one of the above methods.

[0133] It should be noted that, in the description of this application, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, the meaning of "plurality" refers to at least two.

[0134] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0135] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0136] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0137] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0138] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0139] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0140] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in the field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. An intelligent compression method for AI law enforcement vehicle recognition events, characterized in that: The method comprises: Acquire the city management images to be processed in real time according to a preset frequency; Obtaining GPS longitude and latitude information in the adjacent city management image to be processed and azimuth data information of the adjacent city management image to be processed; Inputting the GPS longitude and latitude information in the adjacent city management image to be processed, the azimuth data information of the adjacent city management image to be processed, and the adjacent city management image to be processed into a preset neural network model, and outputting all event information in the adjacent city management image to be processed; If there are events of the same category in all event information in the adjacent city management images to be processed, the preset latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into the preset event intelligent judgment model to obtain a first processing result; If the first processing result meets the preset standard value, it indicates that the same event information exists in all the event information in the adjacent city management images to be processed, then the same event information is intelligently compressed to obtain the processing result of the intelligent recognition compression of the city management image; The obtaining of GPS longitude and latitude information in the adjacent city management image to be processed and azimuth data information of the adjacent city management image to be processed includes: Using a preset Beidou chip to obtain GPS longitude and latitude information in the adjacent city management images to be processed, marking the GPS longitude and latitude information in one of the city management images to be processed as P1, and marking the GPS longitude and latitude information in the adjacent city management image to be processed as P2; Using P1 and P2, a preset first formula is used to obtain the y-axis coordinate and the x-axis coordinate in the two-dimensional coordinate system; The first formula is as follows: (1) Among them, lat represents the latitude in GPS longitude and latitude information, and long represents the longitude in GPS longitude and latitude information; Input the obtained y-axis coordinate and x-axis coordinate into a preset second formula to obtain the azimuth data information of the adjacent city management image to be processed; The second formula is as follows: Bearing(P1) = arctan(y,x)(2) Wherein, the Bearing (P1) represents the azimuth data information of the adjacent city management image to be processed.

2. The method according to claim 1, characterized in that: If there are events of the same category in all event information in the adjacent city management images to be processed, the preset latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into the preset event intelligent judgment model to obtain a first processing result, including: If there are events of the same category in all event information in the adjacent city management images to be processed, then obtaining the latitude and longitude distance information, azimuth angle difference information, and IOU overlap in the adjacent city management images to be processed; The acquired latitude and longitude distance information, azimuth difference information, and IOU overlap in the adjacent city management images to be processed are input into a preset event intelligent judgment model to obtain a first processing result.

3. The method according to claim 2, characterized in that Acquiring the latitude and longitude distance information of the adjacent city management image to be processed, including: Using the latitude and longitude information in the adjacent city management image to be processed, and using a preset GPS latitude and longitude distance formula, the latitude and longitude distance information in the adjacent city management image to be processed is obtained; The GPS latitude and longitude distance formula includes: (3) Among them, lat represents the latitude in the GPS longitude and latitude information, and long represents the longitude in the GPS longitude and latitude information; lat1 and long1 represent the latitude and longitude information in P1; lat2 and long2 represent the latitude and longitude information in P2.

4. The method according to claim 2, characterized in that Acquiring azimuth difference information in adjacent city management images to be processed, including: Obtaining the azimuth angles of the city management images to be processed that are adjacent to each other, marking the azimuth angle of one of the city management images to be processed as Bearing (P1), and marking the azimuth angle of the city management image to be processed that is adjacent to P1 as Bearing (P2); The absolute value of the difference between the azimuth angles Bearing (P1) and Bearing (P2) of the adjacent city management images to be processed is taken to obtain the longitude and latitude distance information of the adjacent city management images to be processed.

5. The method according to claim 2, characterized in that: Obtaining the IOU overlap in the adjacent city management images to be processed, including: Based on events of the same category, the location information of events in adjacent urban management images to be processed is respectively marked; The position information is used to identify the same position and calculate the IOU overlap of adjacent urban management images to be processed.

6. The method according to claim 1, characterized in that If the first processing result meets the preset standard value, it indicates that the same event information exists in all the event information in the adjacent city management images to be processed, including: If the latitude and longitude distance information, azimuth difference information, and IOU overlap of the adjacent city management images to be processed in the first processing result are all less than their respective preset thresholds, it indicates that the same event information exists in all event information of the adjacent city management images to be processed.

7. An intelligent compression system for AI law enforcement vehicle recognition events, applied to an intelligent compression method for AI law enforcement vehicle recognition events as described in any one of claims 1-6, characterized in that: The method comprises: An acquisition module, used for acquiring the city management images to be processed in real time according to a preset frequency; The first processing module is used to obtain GPS longitude and latitude information in the adjacent city management image to be processed and azimuth data information of the adjacent city management image to be processed; The second processing module is used to input the GPS longitude and latitude information in the adjacent city management image to be processed, the azimuth data information of the adjacent city management image to be processed, and the adjacent city management image to be processed into a preset neural network model, and output all event information in the adjacent city management image to be processed; The third processing module is used for inputting the preset latitude and longitude distance information, azimuth difference information, and IOU overlap of the adjacent city management images to be processed into a preset event intelligent judgment model to obtain a first processing result if there are events of the same category in all event information in the adjacent city management images to be processed; The fourth processing module is used to perform intelligent compression processing on the same event information if the first processing result meets the preset standard value, indicating that the same event information exists in all the event information in the adjacent urban management images to be processed, so as to obtain the processing result of intelligent recognition and compression of the urban management image.

8. An intelligent compression device for AI law enforcement vehicle identification events, characterized in that: The device comprises: A memory having an executable program stored therein; A processor, configured to execute the executable program in the memory to implement the steps of the method according to any one of claims 1 to 6.

Citation Information

Patent Citations

  • Landscape-planning-oriented image acquisition method for accurate space positioning

    CN103335645A

  • Smart city security monitoring system based on image recognition

    CN111914791A

  • Intelligent identification classification method and system for city management image, and integration device

    CN108133236A

  • Pixhawk-based method for building three-dimensional modeling using UAV

    CN109708622A

  • Object popularity analysis system, method and device

    CN113298597A