CN117596363A - Big data real-time acquisition and analysis information system - Google Patents

Big data real-time acquisition and analysis information system Download PDF

Info

Publication number
CN117596363A
CN117596363A CN202311552457.9A CN202311552457A CN117596363A CN 117596363 A CN117596363 A CN 117596363A CN 202311552457 A CN202311552457 A CN 202311552457A CN 117596363 A CN117596363 A CN 117596363A
Authority
CN
China
Prior art keywords
monitoring
picture
alpha
distinguishing
data acquisition
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202311552457.9A
Other languages
Chinese (zh)
Inventor
高飞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suzhou Langriqing Media Technology Co ltd
Original Assignee
Suzhou Langriqing Media Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Suzhou Langriqing Media Technology Co ltd filed Critical Suzhou Langriqing Media Technology Co ltd
Priority to CN202311552457.9A priority Critical patent/CN117596363A/en
Publication of CN117596363A publication Critical patent/CN117596363A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/44—Event detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G06V20/54—Surveillance or monitoring of activities, e.g. for recognising suspicious objects of traffic, e.g. cars on the road, trains or boats
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/20—Movements or behaviour, e.g. gesture recognition
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/18—Status alarms
    • G08B21/24—Reminder alarms, e.g. anti-loss alarms
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B7/00—Signalling systems according to more than one of groups G08B3/00 - G08B6/00; Personal calling systems according to more than one of groups G08B3/00 - G08B6/00
    • G08B7/06—Signalling systems according to more than one of groups G08B3/00 - G08B6/00; Personal calling systems according to more than one of groups G08B3/00 - G08B6/00 using electric transmission, e.g. involving audible and visible signalling through the use of sound and light sources
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
    • H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/08—Detecting or categorising vehicles

Abstract

The invention discloses a big data real-time acquisition and analysis information system, which comprises: the monitoring system is configured in each monitoring node in the city, the monitoring system collects pictures and sounds covered by the monitoring nodes and transmits the pictures and the sounds, the monitoring system marks the monitoring pictures different from daily pictures, a deep learning model is configured in the monitoring system and is used for analyzing the monitoring pictures and identifying the marked distinguishing events, each monitoring in the monitoring system performs picture identification, different pictures and sounds in different times in the monitoring are processed, the event marking is performed on the distinguishing pictures different from the daily monitoring pictures, and the event marking is transmitted to the traffic management system through an additional transmission channel, so that the picture transmission in the monitoring and the transmission of monitoring around the monitoring are improved.

Description

Big data real-time acquisition and analysis information system
Technical Field
The invention relates to the technical field of big data processing, in particular to a big data real-time acquisition and analysis information system.
Background
The big data real-time acquisition analysis information system is an information system which can acquire data from various data sources in real time and buffer, calculate, store, display and analyze the data. The system can improve the value and timeliness of the data and support various data applications and decisions.
There are many application scenarios of the big data real-time collection and analysis information system, for example:
e-commerce platform: by collecting and analyzing the behaviors, preferences, demands and the like of the users in real time and combining the properties, evaluation, heat and the like of commodities or contents, a personalized recommendation list is generated, and the satisfaction degree and the conversion rate of the users are improved. Meanwhile, the operation management and customer service are optimized by collecting and analyzing information such as orders, payments, logistics and the like in real time.
Financial industry: by collecting and analyzing the credit, transaction, behavior and the like of the user in real time and combining the risk model and rules, the risk assessment and control are carried out on the user, so that fraud and loss are prevented. Meanwhile, intelligent investment advice and prediction are provided by collecting and analyzing market quotations such as stocks, funds, futures and the like in real time.
Smart city: by collecting and analyzing the running states, performance indexes, abnormal conditions and the like of various devices or systems in real time, the problems are found and processed in time, and the operation and maintenance efficiency and quality are improved. Meanwhile, intelligent management and service are provided by collecting and analyzing data in traffic, environment, safety and the like in real time.
The Internet of things: the intelligent sensing, monitoring, control and optimization of people, objects and environments are realized by collecting and analyzing mass data generated by the sensors, the cameras and other intelligent terminals in real time. Meanwhile, personalized products and services are provided by collecting and analyzing the demands, feedback, evaluation and the like of users in real time.
In city management, the influence of the emergency is often determined by time for the generation of the emergency, but in traffic management, the emergency is often informed by the on-site masses, and the monitoring system in the city is large in quantity and complex in picture and cannot be found out in time, so that for city management, the real-time acquisition and analysis information system can greatly improve the response time of the emergency and reduce the influence of the emergency.
Disclosure of Invention
The invention aims to provide a big data real-time acquisition and analysis information system, which can screen pictures in a city monitoring system to a certain extent, identify picture elements, and transmit the monitoring and surrounding monitoring coverage range in real time, so that the big data can be timely transmitted to an traffic management system and timely responded, and the processing efficiency of an emergency is improved.
In order to achieve the above purpose, the invention is realized by the following technical scheme: a big data real-time acquisition analysis information system, comprising:
monitoring system the monitoring system alpha is configured at each monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) In the monitoring system, the alpha acquisition monitoring node (alpha 1 ，α 2 ，α 3 ，…，α n ) The method comprises the steps that a covered picture and sound are transmitted, a monitoring system alpha marks a distinguishing event gamma on a monitoring picture different from a daily picture, a deep learning model tau is configured in the monitoring system alpha, and the deep learning model tau is used for analyzing the monitoring picture and identifying the marked distinguishing event gamma;
the data acquisition end is used for acquiring the data of each monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) Data acquisition is performed on images and sounds and the data acquisition is performed on each monitoring node (alpha) 1 ，α 2 ，α 3 ，...，α n ) Marking, wherein a data acquisition end is provided with two data transmission channels, one of which is a common transmission channel W 1 The other is a special transmission channel, a special transmission channel W 2 For transmitting a distinguishing event gamma tag;
main processing system R 1 Differential processing system R 2 The data acquisition end is positioned at the upper level of the data acquisition end, the data acquisition end synthesizes images and sounds related to the distinguishing event gamma through the distinguishing event gamma mark, and the main processing system R 1 The processed data acquisition end passes through a common transmission channel W 1 The transmitted monitoring picture and sound distinguish processing system R 2 Processing the picture and sound marked as distinguishing event gamma;
the display end is positioned in the main processing system R 1 And displays the main processing system R 1 The picture and sound displayed after processing.
In one or more embodiments of the present invention, the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) The collected pictures are connected with the traffic control signal lamp at the position, and the monitoring node (alpha 1 ，α 2 ，α 3 ，…，α n ) The detected picture range of the monitoring node (alpha) 1 ，α 2 ，α 3 ，…，α n ) The middle picture is divided into a waiting interval rho and a passing interval sigma according to the change of the signal lamp, the behavior of the pedestrians in the picture is marked, the waiting state, the passing state and the abnormal state are marked, and the scoring statistics is carried out on the pedestrian state.
In one or more embodiments of the present invention, the monitoring system α monitors the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) Recording the daily pedestrian state in the system, wherein when the pedestrian makes a behavior conforming to the waiting interval rho or the passing interval sigma, the behavior of the pedestrian is not scored, and when the pedestrian makes a behavior violating the waiting interval rho or the passing interval rho, scoring statistics is carried out according to the quantity of the pedestrian, namely:
wherein d is the pedestrian behavior result score, delta is the behavior in violation of the interval, s 1 To violate the number of pedestrians in the interval, epsilon is consistent with the behavior in the interval, s 2 To match the number of pedestrians in the interval, S is the number of pedestrians in the monitoring node (alpha 1 ，α 2 ，α 3 ，…，α n ) Waiting interval ρ or total number of pedestrians in passing interval σ, generating time scale of pedestrian behavior result score dMarked as discriminating event gamma.
In one or more embodiments of the present invention, the monitoring node image performs pedestrian range marking, marks the pedestrian distribution points, generates a marking point k, the marking point k is the head position of the pedestrian, generates a marking point k position diagram, the adjacent marking points k form areas with different shapes, performs marking point k identification according to the area generated by the marking point k, and determines the pedestrian state in the marking point k.
In one or more embodiments of the present invention, the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) The collected sound is counted correspondingly according to the time period, the sound decibel range in the time period is set, and the sound exceeding 10% of sound decibel is marked as a distinguishing event gamma;
according to the monitoring node (alpha) 1 ，α 2 ，α 3 ，…，α n ) The recorded range of pedestrians in the picture is marked, namely, the density area of pedestrians is divided in the picture, and when the pedestrians are in the non-pedestrian movement area in the corresponding time period, the pedestrians are marked as distinguishing events gamma.
In one or more embodiments of the present invention, the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) The method can also identify the vehicles appearing in the picture, and score according to the positions of the vehicles in the road, the vehicle running states of the vehicles and the vehicle lamplight states, namely: d= ((β×f) + (β×g))a;
wherein D is the score of the vehicle behavior result, beta is the vehicle which violates the road running behavior, F is the position of the lane where the outermost lane score is the lowest, the inward approach score is gradually increased, G is the vehicle light state, the light flashing state is recorded as 1, the rest states are recorded as 0, and A is the number of vehicles which violate the road running behavior.
In one or more embodiments of the present invention, the marked distinguishing event γ is identified by a deep learning model τ, that is, the distinguishing event γ picture and the sound content are identified according to the mark, and the behavior of the pedestrian in different emergencies is learned and recorded, and the behavior of the pedestrian in the deep learning model τ is taken to compare with the distinguishing event γ picture and the sound content, so as to determine the nature of the distinguishing event γ.
In one or more embodiments of the present invention, the data collection terminal collects different monitoring nodes (α) in the monitoring system α 1 ，α 2 ，α 3 ，…，α n ) And the picture and sound of different monitoring nodes (alpha 1 ，α 2 ，α 3 ，…，α n ) Transmitting the data to a display end of the jurisdiction, splitting the data of the distinguishing event gamma identified by the deep learning model tau through a data acquisition end, and transmitting the data through a special transmission channel W 2 A monitoring node (alpha) transmitting without distinguishing event gamma 1 ，α 2 ，α 3 ，…，α n ) The pictures pass through a common transmission channel W 1 Transmission, special transmission channel W 2 And the common transmission channel W 1 Can be converted from each other.
In one or more embodiments of the present invention, the pictures and sounds generated by the distinguishing event γ pass through the special transmission channel W 2 Delivered to a differential processing system R 2 To process and synthesize the picture and the sound, and the picture and the sound of the non-distinguishing event gamma are transmitted through the common transmission channel W 1 To the main processing system R 1 Processing, a main processing system R 1 Multiple different processing systems R are respectively connected with display ends in different jurisdictions 2 And connecting display ends of all jurisdictions.
In one or more embodiments of the present invention, the discrimination event γ is generated by the discrimination processing system R 2 Transmitting to display end and reminding, distinguishing the monitoring node (alpha) in the jurisdiction display end to which the event gamma belongs 1 ，α 2 ，α 3 ，...，α n ) The picture can be amplified and reminded, and the display end is provided with a warning module which can generate sound and light warning.
The invention provides a big data real-time acquisition and analysis information system. Compared with the prior art, the method has the following beneficial effects:
1. according to the monitoring system alpha, each monitoring is subjected to picture identification, different pictures and sounds are processed according to different times in the monitoring, event marking is carried out on different pictures different from daily monitoring pictures, the pictures are transmitted to an traffic management system through an additional transmission channel, picture transmission in the monitoring and monitoring transmission around the monitoring are improved, real-time transmission of the pictures is realized, and meanwhile, the monitored marks can be highlighted.
2. In the daily monitoring process, the monitoring system alpha identifies emergencies according to different pictures in the monitoring process, namely the positions, the stay states and other information of pedestrians in the pictures, and when the pictures are abnormal, the pictures can be marked as distinguishing events gamma, the distinguishing events gamma are identified through a deep learning model tau, and the accuracy of the picture identification emergencies is improved.
3. When the data of each monitoring is collected, each monitoring picture is synchronously transmitted, after a distinguishing event gamma occurs in the monitoring system alpha, the data collection can be realized by calling a transmission channel, the transmission efficiency of pictures and sounds is improved, the delay in picture transmission is reduced, the instantaneity of the pictures and the sounds is ensured, and the processing efficiency of emergencies can be improved.
4. When the picture is transmitted, different transmission can be carried out according to different monitoring pictures by changing the call of the transmission channel, real-time transmission can be carried out when the different events gamma marked by the monitoring pictures are detected, further, the data aging can be ensured, the display end can be subjected to amplified display processing, corresponding warning can be carried out, and the processing efficiency of the emergency is improved.
Drawings
FIG. 1 is a block diagram of a real-time acquisition and analysis information system of the present invention;
fig. 2 is a schematic diagram of the distribution of pedestrian marking points k of the present invention.
Detailed Description
Various embodiments of the invention are disclosed in the accompanying drawings, and for purposes of explanation, numerous practical details are set forth in the following description. However, it should be understood that these practical details are not to be taken as limiting the invention. That is, in some embodiments of the invention, these practical details are unnecessary. Furthermore, for the purpose of simplifying the drawings, some of the presently available structures and elements are shown in a simplified schematic form, and the same reference numerals will be used throughout the drawings to designate the same or similar elements. And features of different embodiments may be interactively applied, if implementation is possible.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have their ordinary meaning as understood by one of ordinary skill in the art. Furthermore, the definitions of the words and phrases used herein should be understood and interpreted to have a meaning consistent with the understanding of the relevant art and technology. These terms are not to be construed as idealized or overly formal meanings unless expressly so defined.
Referring to fig. 1-2, the present invention provides a real-time big data acquisition and analysis information system, which performs image marking by processing different events in a monitoring image and realizes real-time image transmission, thereby improving the processing efficiency of a traffic management system for urban emergencies. Comprising the following steps:
monitoring system alpha, configured at each monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) In the monitoring system, the alpha acquisition monitoring node (alpha 1 ，α 2 ，α 3 ，…，α n ) The method comprises the steps that a covered picture and sound are transmitted, a monitoring system alpha marks a distinguishing event gamma on a monitoring picture different from a daily picture, a deep learning model tau is configured in the monitoring system alpha, and the deep learning model tau is used for analyzing the monitoring picture and identifying the marked distinguishing event gamma;
the data acquisition end is used for acquiring the data of each monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) Data acquisition is performed on images and sounds and the data acquisition is performed on each monitoring node (alpha) 1 ，α 2 ，α 3 ，...，α n ) Marking, and configuring two types of data acquisition terminalsData transmission channel, one is common transmission channel W 1 The other is a special transmission channel, a special transmission channel W 2 For transmitting a distinguishing event gamma tag;
main processing system R 1 Differential processing system R 2 The data acquisition end is positioned at the upper level of the data acquisition end, the data acquisition end synthesizes images and sounds related to the distinguishing event gamma through the distinguishing event gamma mark, and the main processing system R 1 The processed data acquisition end passes through a common transmission channel W 1 The transmitted monitoring picture and sound distinguish processing system R 2 Processing the picture and sound marked as distinguishing event gamma;
the display end is positioned in the main processing system R 1 And displays the main processing system R 1 The picture and sound displayed after processing.
In this embodiment, the monitoring system α performs image acquisition to realize the acquisition of each monitoring node (α 1 ，α 2 ，α 3 ，...,α n ) Picture of internal monitoring, while monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) In the monitoring picture of (a), the influence of the emergency event will cause the monitoring node (alpha) 1 ，α 2 ，α 3 ，...，α n ) The special result different from the daily monitor screen appears in the monitor screen, namely, different results are generated in the monitor screen, and the identification of the event in the screen is performed based on the results.
Wherein there are a large number of and complex monitoring nodes (alpha) within the jurisdiction due to the different jurisdictions in which each traffic management system is administered 1 ，α 2 ，α 3 ，...，α n ) A large number of monitoring nodes (alpha 1 ，α 2 ，α 3 ，...，α n ) All that is required to be transmitted to the processing system in the jurisdiction, and as such, the monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) The transmitted picture has high delay and needs to ensure stable operation of each monitor picture, and thus, for the transmitted picture sound data, there is caused a problem that the data timeliness is poor.
In one embodiment, the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) The collected pictures are connected with the traffic control signal lamp at the position, and the monitoring node (alpha 1 ，α 2 ，α 3 ，…，α n ) The detected picture range of the monitoring node (alpha) 1 ，α 2 ，α 3 ，…，α n ) The middle picture is divided into a waiting interval rho and a passing interval sigma according to the change of the signal lamp, the behavior of the pedestrians in the picture is marked, the waiting state, the passing state and the abnormal state are marked, and the pedestrian states are counted in a scoring mode.
In this embodiment, different pedestrians are divided into different states, events reflected in a picture in different states are different, different influence results are generated, and whether the different events are different events gamma is judged according to different results in the picture, so that the effect of actively judging the different events gamma is formed, and further analysis can be performed according to the different states of the pedestrians.
Based on the characteristics of favoring and looking around the pedestrians, when the distinguishing event gamma occurs, a large number of pedestrians are often looked around the event occurrence area, and because the area where the event occurs is not a specific area, the positions where the pedestrians are looked at are different, the large number of pedestrians often mean that factors attracting pedestrians exist at the positions, and therefore analysis is accurate based on the states of the pedestrians.
In one embodiment, the monitoring system α monitors the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) Recording the daily pedestrian state in the system, wherein when the pedestrian makes a behavior conforming to the waiting interval rho or the passing interval sigma, the behavior of the pedestrian is not scored, and when the pedestrian makes a behavior violating the waiting interval rho or the passing interval sigma, the scoring statistics is carried out according to the quantity of the pedestrian, namely:
wherein d is the pedestrian behavior result score, delta is the violation of the intervalBehavior in s 1 To violate the number of pedestrians in the interval, epsilon is consistent with the behavior in the interval, s 2 To match the number of pedestrians in the interval, S is the number of pedestrians in the monitoring node (alpha 1 ，α 2 ，α 3 ，…，α n ) The total number of pedestrians in the waiting interval ρ or the passing interval σ is marked as a distinguishing event γ when the pedestrian behavior result score d is generated.
In the present embodiment, the monitoring node (α 1 ，α 2 ，α 3 ，...,α n ) The coverage range is often larger, and some smaller emergencies are not enough to generate larger influence, so different pedestrian behavior result scores can be obtained through the number of people who make a reaction against the convention in different interval time, namely, the difference of the scores obtained by d and d can reflect the size and influence of the emergencies.
Wherein the monitoring node (alpha 1 ，α 2 ，α 3 ，...,α n ) The number of pedestrians existing at different times is different, the generated impact scores are different, the ratio of the pedestrians in violation of the total number of pedestrians is calculated, the score is calculated according to the ratio result, and the monitoring node (alpha) 1 α 2 ，α 3 ，...,α n ) The higher the pedestrian behavior result score d, the more important the distinction of the event γ is.
In one embodiment, the monitoring node picture marks the pedestrian range, marks the pedestrian distribution points, generates a mark point k, wherein the mark point k is the head position of the pedestrian, generates a mark point k position diagram, forms areas with different shapes by adjacent mark points k, carries out mark point k identification according to the areas generated by the mark points k, and judges the pedestrian state in the mark points k.
In this embodiment, when an emergency occurs at a plurality of pedestrian positions, pedestrians are often stationed around the emergency, and distributed in a round package, and the pedestrians are distributed in a round package at a monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) Generating a mark point k position after identifying the pedestrian head position in the pictureAfter the drawing, the similar mark points k can be connected, and the identification can be performed according to the generated different shapes, so that the situation of aggregation among pedestrians is not generated in a normal state, and the distribution is uniform, and therefore, the judgment of the distinguishing event gamma can be performed according to the surrounding positions of the pedestrians.
In one embodiment, the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) The collected sound is counted correspondingly according to the time period, the sound decibel range in the time period is set, and the sound exceeding 10% of sound decibel is marked as a distinguishing event gamma;
according to the monitoring node (alpha) 1 ，α 2 ，α 3 ，…，α n ) The recorded range of pedestrians in the picture is marked, namely, the density area of pedestrians is divided in the picture, and when the pedestrians are in the non-pedestrian movement area in the corresponding time period, the pedestrians are marked as distinguishing events gamma.
In the present embodiment, since the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) The recorded picture is a 24-hour generated picture, so that the density of pedestrians in the picture is different in different time periods, and the density area is changed according to the time periods, and the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) In the monitored crossroads, the dense areas of pedestrians at the nodes of the on-off duty can generate waiting areas at two sides of the road, and after supper, the dense areas of pedestrians can be generated on squares beside the road, and pedestrians in different time periods can generate changes of the dense areas.
Wherein part of the monitoring nodes (alpha 1 ，α 2 ，α 3 ，…，α n ) An area to be set up beside a road includes both a part of the road and another part of the area (a cell fence, a school fence or an enclosure of an unpublished place), and no pedestrian remains in the area near the other part, so that pedestrian behavior determination can be performed by pedestrians located in different areas, and nodes when a current person appears in an area where a pedestrian is unlikely to appear are markedThe monitoring can be conveniently called.
While the monitoring node (alpha 1 ，α 2 ，α 3 ，…，α n ) The noise disturbance problem generated in midnight and late night can be marked according to sound decibel change at night, namely, the behavior of driving at high speed at night and generating noise can be marked, and the behavior can be processed by distinguishing event gamma.
In one embodiment, the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) The method can also identify the vehicles appearing in the picture, and score according to the positions of the vehicles in the road, the vehicle running states of the vehicles and the vehicle lamplight states, namely:
D＝((β*F)+(β*G))*A；
wherein D is the score of the vehicle behavior result, beta is the vehicle which violates the road running behavior, F is the position of the lane where the outermost lane score is the lowest, the inward approach score is gradually increased, G is the vehicle light state, the light flashing state is recorded as 1, the rest states are recorded as 0, and A is the number of vehicles which violate the road running behavior.
In this embodiment, two states are present in the running process of the road, and these two states are generated in the waiting section ρ and the passing section σ, respectively, the running of the vehicle in the passing section σ is normal, the stopping is marked as a violation of the running behavior of the road, the running in the waiting section ρ is marked as a violation of the running behavior of the road, and the vehicle is warned by means of flickering of light after the running of the vehicle is failed or has a collision or the like.
Wherein, since the monitoring node (alpha 1 ，α 2 ，α 3 ，…，α n ) The number of vehicles displayed on the screen is different, and the greater the number of vehicles, that is, the greater the variable a, the higher the vehicle behavior result score D, which indicates the monitoring node (α 1 ，α 2 ，α 3 ，…，α n ) The greater the impact of the incident occurring in the system.
In one embodiment, the marked distinguishing event γ is identified by a deep learning model τ, that is, the distinguishing event γ picture and the sound content are identified according to the marked distinguishing event γ picture and the sound content, and the behavior actions of pedestrians in different emergencies are learned and recorded, and the behavior actions of pedestrians in the emergency are called out from the deep learning model τ to be compared with the distinguishing event γ picture and the sound content, so as to judge the nature of the distinguishing event γ.
In this embodiment, analysis of the behavior of the pedestrian is performed through the deep learning model τ, so as to form analysis of the distinguishing event γ, that is, to realize determination of whether the distinguishing event γ is an emergency event under the behaviors of different pedestrians, and the behavior of the pedestrians in the distinguishing event γ can be learned by modifying the distinguishing event γ determined through the deep learning model τ.
The deep learning model tau is set to learn different behaviors of pedestrians, the different behaviors correspond to different events, the distinguishing event gamma judged by the monitoring system alpha is further screened, and transmission of the error distinguishing event gamma is reduced.
In one embodiment, the data acquisition end acquires different monitoring nodes (alpha 1 ，α 2 ，α 3 ，…，α n ) And the picture and sound of different monitoring nodes (alpha 1 ，α 2 ，α 3 ，…，α n ) Transmitting the data to a display end of the jurisdiction, splitting the data of the distinguishing event gamma identified by the deep learning model tau through a data acquisition end, and transmitting the data through a special transmission channel W 2 A monitoring node (alpha) transmitting without distinguishing event gamma 1 ，α 2 ，α 3 ，…，α n ) The pictures pass through a common transmission channel W 1 Transmission, special transmission channel W 2 And the common transmission channel W 1 Can be converted from each other.
In the present embodiment, a special transmission channel W 2 And the common transmission channel W 1 The mutual conversion can transmit the multi-channel distinguishing event gamma, namely, when a plurality of distinguishing events gamma exist, a special transmission channel W 2 Side common transmission channel W 1 Conversion into special transmission channels W 2 And transmitting the distinguishing event gamma picture.
In order to ensure the transmission efficiency, when the transmission rate is low, the real-time transmission is performed by adopting a split transmission mode, namely splitting the picture and the sound in the distinguishing event gamma, and adopting different special transmission channels W 2 And transmission is carried out, so that the transmission efficiency is quickened.
In one embodiment, the pictures and sounds generated by the distinguishing event gamma pass through a special transmission channel W 2 Delivered to a differential processing system R 2 To process and synthesize the picture and the sound, and the picture and the sound of the non-distinguishing event gamma are transmitted through the common transmission channel W 1 To the main processing system R 1 Processing, a main processing system R 1 Multiple different processing systems R are respectively connected with display ends in different jurisdictions 2 And connecting display ends of all jurisdictions.
In the present embodiment, the monitoring node (α) that generates the distinguishing event γ 1 ，α 2 ，α 3 ，...，α n ) The number is very small, and therefore, the special transmission channel W is passed 2 The transmission of pictures and sounds contained in the distinguishing event gamma can be performed more quickly, while the rest of the monitoring nodes (alpha) 1 ，α 2 ，α 3 ，…，α n ) All need to be transmitted at the same time, so the main processing system R 1 The processed data is more, and the data is processed by the distinguishing processing system R 2 The picture and sound of the distinguishing event gamma can be more quickly transmitted.
In one embodiment, the discrimination event γ is defined by the discrimination processing system R 2 Transmitting to display end and reminding, distinguishing the monitoring node (alpha) in the jurisdiction display end to which the event gamma belongs 1 ，α 2 ，α 3 ，…，α n ) The picture can be amplified and reminded, and the display end is provided with a warning module which can generate sound and light warning.
In this embodiment, the corresponding speed of the emergency event can rapidly control the emergency event, and the control time of the emergency event is often equal to the emergency eventThe influence of the incident is related, so that the condition of the included distinguishing incident gamma can be observed more quickly in a warning mode, and the related monitoring node (alpha can be carried out more quickly when an emergency alarm occurs 1 ，α 2 ，α 3 ，…，α n ) Is used for positioning.
In summary, the technical solution disclosed in the above embodiment of the present invention has at least the following advantages:
1. according to the monitoring system alpha, each monitoring is subjected to picture identification, different pictures and sounds are processed according to different times in the monitoring, event marking is carried out on different pictures different from daily monitoring pictures, the pictures are transmitted to an traffic management system through an additional transmission channel, picture transmission in the monitoring and monitoring transmission around the monitoring are improved, real-time transmission of the pictures is realized, and meanwhile, the monitored marks can be highlighted.
2. In the daily monitoring process, the monitoring system alpha identifies emergencies according to different pictures in the monitoring process, namely the positions, the stay states and other information of pedestrians in the pictures, and when the pictures are abnormal, the pictures can be marked as distinguishing events gamma, the distinguishing events gamma are identified through a deep learning model tau, and the accuracy of the picture identification emergencies is improved.
3. When the data of each monitoring is collected, each monitoring picture is synchronously transmitted, after a distinguishing event gamma occurs in the monitoring system alpha, the data collection can be realized by calling a transmission channel, the transmission efficiency of pictures and sounds is improved, the delay in picture transmission is reduced, the instantaneity of the pictures and the sounds is ensured, and the processing efficiency of emergencies can be improved.
4. When the picture is transmitted, different transmission can be carried out according to different monitoring pictures by changing the call of the transmission channel, real-time transmission can be carried out when the different events gamma marked by the monitoring pictures are detected, further, the data aging can be ensured, the display end can be subjected to amplified display processing, corresponding warning can be carried out, and the processing efficiency of the emergency is improved.
Although the present invention has been described in connection with the above embodiments, it should be understood that the invention is not limited thereto, but may be variously modified and modified by those skilled in the art without departing from the spirit and scope of the invention, and the scope of the invention is accordingly defined by the appended claims.

Claims (9)

1. A big data real-time acquisition and analysis information system, comprising:
monitoring system alpha, configured at each monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) In the monitoring system, the alpha acquisition monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) The method comprises the steps that a covered picture and sound are transmitted, a monitoring system alpha marks a distinguishing event gamma on a monitoring picture different from a daily picture, a deep learning model tau is configured in the monitoring system alpha, and the deep learning model tau is used for analyzing the monitoring picture and identifying the marked distinguishing event gamma;
the data acquisition end is used for acquiring the data of each monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) Data acquisition is performed on images and sounds and the data acquisition is performed on each monitoring node (alpha) 1 ，α 2 ，α 3 ，...，α n ) Marking, wherein a data acquisition end is provided with two data transmission channels, one of which is a common transmission channel W 1 The other is a special transmission channel, a special transmission channel W 2 For transmitting a distinguishing event gamma tag;
main processing system R 1 Differential processing system R 2 The data acquisition end is positioned at the upper level of the data acquisition end, the data acquisition end synthesizes images and sounds related to the distinguishing event gamma through the distinguishing event gamma mark, and the main processing system R 1 The processed data acquisition end passes through a common transmission channel W 1 The transmitted monitoring picture and sound distinguish processing system R 2 Processing the picture and sound marked as distinguishing event gamma;
the display end is positioned in the main processing system R 1 And displays the main processing system R 1 The picture and sound displayed after processing; the monitoring system α monitors the monitoring node (α 1 ，α 2 ，α 3 ，...，α n ) Recording the daily pedestrian state in the system, wherein when the pedestrian makes a behavior conforming to the waiting interval rho or the passing interval sigma, the behavior of the pedestrian is not scored, and when the pedestrian makes a behavior violating the waiting interval rho or the passing interval sigma, the scoring statistics is carried out according to the quantity of the pedestrian, namely:
wherein d is the pedestrian behavior result score, delta is the behavior in violation of the interval, s 1 To violate the number of pedestrians in the interval, epsilon is consistent with the behavior in the interval, s 2 To match the number of pedestrians in the interval, S is the number of pedestrians in the monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) The total number of pedestrians in the waiting interval ρ or the passing interval σ is marked as a distinguishing event γ when the pedestrian behavior result score d is generated.
2. The real-time big data acquisition and analysis information system according to claim 1, wherein the monitoring node picture carries out pedestrian range marking and marks pedestrian distribution points to generate a mark point k, the mark point k is the head position of a pedestrian, a mark point k position diagram is generated, adjacent mark points k form areas with different shapes, mark point k identification is carried out according to the areas generated by the mark points k, and the pedestrian state in the mark point k is judged.
3. A real-time big data acquisition and analysis information system according to claim 1, characterized in that the monitoring node (α 1 ，α 2 ，α 3 ，...，α n ) The collected pictures are connected with the traffic control signal lamp at the position, and the monitoring node (alpha 1 ，α 2 ，α 3 ，...，α n ) Marking the detected picture range of the pictureMonitoring node (alpha) 1 ，α 2 ，α 3 ，...，α n ) The middle picture is divided into a waiting interval rho and a passing interval sigma according to the change of the signal lamp, the behavior of the pedestrians in the picture is marked, the waiting state, the passing state and the abnormal state are marked, and the scoring statistics is carried out on the pedestrian state.
4. A real-time big data acquisition and analysis information system according to claim 1, characterized in that the monitoring node (α 1 ，α 2 ，α 3 ，...，α n ) The collected sound is counted correspondingly according to the time period, the sound decibel range in the time period is set, and the sound exceeding 10% of sound decibel is marked as a distinguishing event gamma;
according to the monitoring node (alpha) 1 ，α 2 ，α 3 ，...，α n ) The recorded range of pedestrians in the picture is marked, namely, the density area of pedestrians is divided in the picture, and when the pedestrians are in the non-pedestrian movement area in the corresponding time period, the pedestrians are marked as distinguishing events gamma.
5. A real-time big data acquisition and analysis information system according to claim 1, characterized in that the monitoring node (α 1 ，α 2 ，α 3 ，...，α n ) The method can also identify the vehicles appearing in the picture, and score according to the positions of the vehicles in the road, the vehicle running states of the vehicles and the vehicle lamplight states, namely:
D＝((β*F)+(β*G))*A；
wherein D is the score of the vehicle behavior result, beta is the vehicle which violates the road running behavior, F is the position of the lane where the outermost lane score is the lowest, the inward approach score is gradually increased, G is the vehicle light state, the light flashing state is recorded as 1, the rest states are recorded as 0, and A is the number of vehicles which violate the road running behavior.
6. The real-time big data acquisition and analysis information system according to claim 4, wherein the marked distinguishing event gamma is identified by a deep learning model tau, namely, the distinguishing event gamma picture and the sound content are identified according to the marked distinguishing event gamma picture and the sound content, the behavior actions of pedestrians in different emergencies are learned and recorded, the behavior actions of the pedestrians in the emergencies in the deep learning model tau are called to be compared with the distinguishing event gamma picture and the sound content, and the nature of the distinguishing event gamma is judged.
7. The real-time big data acquisition and analysis information system according to claim 1, wherein the data acquisition end acquires the information of different monitoring nodes (α 1 ，α 2 ，α 3 ，...，α n ) And the picture and sound of different monitoring nodes (alpha 1 ，α 2 ，α 3 ，...，α n ) Transmitting the data to a display end of the jurisdiction, splitting the data of the distinguishing event gamma identified by the deep learning model tau through a data acquisition end, and transmitting the data through a special transmission channel W 2 A monitoring node (alpha) transmitting without distinguishing event gamma 1 ，α 2 ，α 3 ，...，α n ) The pictures pass through a common transmission channel W 1 Transmission, special transmission channel W 2 And the common transmission channel W 1 Can be converted from each other.
8. The real-time big data acquisition and analysis information system according to claim 7, wherein the frames and sounds generated by the distinguishing event γ pass through the special transmission channel W 2 Delivered to a differential processing system R 2 To process and synthesize the picture and the sound, and the picture and the sound of the non-distinguishing event gamma are transmitted through the common transmission channel W 1 To the main processing system R 1 Processing, a main processing system R 1 Multiple different processing systems R are respectively connected with display ends in different jurisdictions 2 And connecting display ends of all jurisdictions.
9. The real-time big data acquisition and analysis information system according to claim 8, wherein the distinguishing event gamma is generated by the distinguishing processing systemR 2 Transmitting to display end and reminding, distinguishing the monitoring node (alpha) in the jurisdiction display end to which the event gamma belongs 1 ，α 2 ，α 3 ，...，α n ) The picture can be amplified and reminded, and the display end is provided with a warning module which can generate sound and light warning.
CN202311552457.9A 2023-11-21 2023-11-21 Big data real-time acquisition and analysis information system Pending CN117596363A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311552457.9A CN117596363A (en) 2023-11-21 2023-11-21 Big data real-time acquisition and analysis information system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311552457.9A CN117596363A (en) 2023-11-21 2023-11-21 Big data real-time acquisition and analysis information system

Publications (1)

Publication Number Publication Date
CN117596363A true CN117596363A (en) 2024-02-23

Family

ID=89912738

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311552457.9A Pending CN117596363A (en) 2023-11-21 2023-11-21 Big data real-time acquisition and analysis information system

Country Status (1)

Country Link
CN (1) CN117596363A (en)

Similar Documents

Publication Publication Date Title
Svensson A method for analysing the traffic process in a safety perspective
Silva et al. Ethical implications of social internet of vehicles systems
Ali et al. An Extreme Value Theory approach to estimate crash risk during mandatory lane-changing in a connected environment
Meir et al. Understanding complex traffic road scenes: The case of child-pedestrians’ hazard perception
Barić et al. Introducing experiment in pedestrian behaviour and risk perception study at urban level crossing
CN115861983A (en) Intelligent management system and method for mechanical equipment
CN114241589B (en) Bus driver violation behavior determination method and device based on vehicle-mounted video
CN108053153A (en) A kind of ecotourism environment bearing capacity early warning system
Desai et al. Accident detection using ml and ai techniques
Sadler et al. Factors Affecting Retention Behavior: A Model To Predict At-Risk Students. AIR 1997 Annual Forum Paper.
CN117596363A (en) Big data real-time acquisition and analysis information system
CN109903150A (en) Vehicle borrows risk evaluating system and method
Mirmohammadi et al. Investigation of road accidents and casualties factors with MCDM methods in Iran
CN117392852A (en) Cloud computing resource scheduling optimization system oriented to big data analysis
US20190362432A1 (en) Compliance Aware Crime Risk Avoidance System
Hawas A microscopic simulation model for incident modeling in urban networks
CN116777685A (en) Scenic spot guiding system, method and device and management server
CN110705854A (en) Driving level evaluation method and system
Wu et al. Information Volume Threshold for Graphical Variable Message Signs Based on Drivers’ Visual Cognition Behavior
Shaheed et al. Motorcycle conspicuity: what factors have the greatest impact
Bäumler et al. Report on validation of the stochastic traffic simulation (Part B)
KR20210097555A (en) Service methods for providing information on traffic accidents
Asif et al. CNN-RNN Algorithm-based Traffic Congestion Prediction System using Tri-Stage Attention
CN111985386A (en) Method for identifying pedestrian illegal-passing behavior based on planned behavior theory
Shubenkova et al. Machine vision in autonomous vehicles: designing and testing the decision making algorithm based on entity attribute value model

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication