CN110796114A - Intelligent video monitoring and early warning system based on biological vision - Google Patents
Intelligent video monitoring and early warning system based on biological vision Download PDFInfo
- Publication number
- CN110796114A CN110796114A CN201911083934.5A CN201911083934A CN110796114A CN 110796114 A CN110796114 A CN 110796114A CN 201911083934 A CN201911083934 A CN 201911083934A CN 110796114 A CN110796114 A CN 110796114A
- Authority
- CN
- China
- Prior art keywords
- monitoring
- abnormal
- visual
- early warning
- area
- 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.)
- Granted
Links
Images
Classifications
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Alarm Systems (AREA)
Abstract
The invention discloses an intelligent video monitoring and early warning system based on biological vision, which consists of three functional subsystems, namely abnormal monitoring of single or multiple moving targets in a local area based on fruit fly visual nerve, large-scale moving target group monitoring in a wide area based on locust visual nerve and automatic monitoring of entrance guard vehicles in parking lot scenes; the abnormal monitoring of single or multiple moving targets in the local area of the visual nerve of the fruit fly is realized by providing a visual nerve network of the fruit fly; the wide-area large-scale moving target group monitoring based on the locust visual nerve is realized by providing a locust visual nerve network; the automatic monitoring of the access control vehicle is to automatically identify the vehicle according to the vehicle video and automatically early warn abnormal vehicles. The intelligent video monitoring system can monitor the abnormal conditions of the targets or the target groups in the local and wide-area environments in a multi-channel manner in real time and automatically detect and monitor the abnormal conditions of the vehicles in the entrance guard environment, thereby realizing the unattended video monitoring.
Description
Technical Field
The invention relates to the fields of computer vision, biological vision, a visual neural network, image processing, video monitoring and the like, in particular to an intelligent video monitoring and early warning system based on biological vision.
Background
At present, monitoring of abnormal conditions in public places becomes an important topic concerned in the field of public safety, most of public place video monitoring methods implement monitoring in a mode of manually observing monitoring videos, and the mode has the defects of high cost, low efficiency, poor real-time performance and the like, and cannot meet the requirement of safety of modern public areas. With the popularization degree and the performance of the computer vision technology being higher and higher, the intelligent monitoring system taking the computer vision as the core conforms to the development trend of the monitoring industry of public areas.
The intelligent video monitoring system assists security personnel to more effectively control the safety condition of a public area by utilizing the technologies of computer vision, intelligent image processing and the like, and aims to achieve the aim of unattended intelligent video monitoring. Existing computer vision techniques still have significant limitations in terms of real-time performance of motion analysis, such as target detection, position estimation, target tracking, and information processing.
In the prior art, an invention patent (publication number: 105023278B) applied by the university of china mining discloses a moving object tracking method and system based on an optical flow method, wherein the method comprises the steps of providing a video image and preprocessing the image to generate a preprocessed image; carrying out edge detection on the preprocessed image, extracting target information from the preprocessed image by using an optical flow method, and fusing the edge detection information and the extracted target information to generate a complete moving target; estimating and analyzing the moving target by using an optical flow method and eliminating error matching points generated by illumination by using a forward and backward error algorithm based on a characteristic point track; and creating a template image and matching the template image to track the moving target.
The specific scheme comprises the following steps: providing a video image and preprocessing the image to generate a preprocessed image; the preprocessing adopts a method comprising low-illumination image enhancement based on Retinex and image denoising based on wavelet threshold; carrying out edge detection on the preprocessed image, extracting target information from the preprocessed image by using an optical flow method, and fusing the edge detection information and the extracted target information to generate a complete moving target; estimating and analyzing the moving target by using an optical flow method, and eliminating error matching points generated by illumination by using a forward and backward error algorithm based on a characteristic point track; creating a template image and matching the template image to track the moving target; when the target information is extracted from the preprocessed image by using an optical flow method, firstly, performing corner detection on the image by using an SUSAN corner detection algorithm and calculating optical flow under multi-scale; the step of generating the complete moving object by fusion further comprises the following steps: and carrying out binarization processing on the image after edge detection and the image generated by optical flow calculation.
However, such an optical flow method for analyzing abnormal behavior of a human population detects a target only by using a moving speed of a moving target, and is not suitable for an environment with strong illumination. However, the intelligent video monitoring system requires that the acquired video information can be fully utilized, and early warning can be timely and accurately sent out to abnormal conditions, so that the real-time requirement is high, and the current prior art cannot meet the requirement.
Disclosure of Invention
Aiming at the problems of high cost, low efficiency and the like of video monitoring in the traditional public place environment, the invention aims to provide a visual neural network anomaly detection technology and a monitoring system based on fruit flies and locusts, and realize unattended intelligent video monitoring.
The invention is realized by the following steps:
the intelligent video monitoring and early warning system is composed of three functional subsystems, namely abnormal monitoring of single or multiple moving targets (such as pedestrians) in a local area (local area) based on drosophila visual nerves, monitoring of large-scale moving target groups (such as crowds) in a wide area (square environment) based on locust visual nerves, and automatic monitoring of access control vehicles in parking lot scenes. The abnormal monitoring of the single or multiple moving targets in the local area based on the fruit fly vision is realized by providing a fruit fly vision neural network to automatically monitor and early warn the behavior and activity of the moving targets; the wide-area large-scale moving target group monitoring based on the locust visual nerve is realized by providing a locust visual nerve network, and automatically detecting and early warning the behavior of the moving target group in a wide-area scene; the automatic monitoring of the entrance guard vehicle is to automatically identify the vehicle according to the vehicle video and automatically early warn abnormal vehicles.
The intelligent video monitoring and early warning system can monitor the activity of the moving target or large-scale moving target group in a local area and wide area scene in real time and in multiple channels and automatically early warn abnormal vehicles, and can realize the unattended operation of video monitoring.
The following three aspects are explained in detail:
(1) the provided local dynamic detection and early warning technology based on the visual nerve of the fruit flies is a bionic artificial neural network designed according to the structure of the visual neural network of the fruit flies; the method can perform behavior detection on the moving target according to the speed, the size, the moving direction and the like of the moving target in the local area, and has the advantages of high detection speed, high accuracy and good real-time property; the technical scheme is described as follows:
1) the size of the gray scale image is M multiplied by M, the sampling time interval is T, the unit is second, and M is a multiple of 8;
2) inputting: the k frame gray scale image Pk;
3) According to PkAnd Pk-1Generating a differential map Ek(ii) a Will EkEqually divided into subfigure sequences Fij,i,j=1,2,...,M/8;
4) For each (i, j), a 3 × 3 template pair F is utilizedijPerforming convolution with FijConversion into 6 × 6 sub-gray-scale map Gij;
5) Sequentially converting the gray scale image GijSplicing (i, j ═ 1, 2.., M/8) to obtain a size M × M of ashDegree diagram Qk,m=0.75M;
6) According to the following formula
x'kl=-a×xkl+(b-xkl)×f(xkl),k,l=1,2,...,m;
Calculating the membrane potential x of the on-off channelklWhere a and b are non-negative constants and f (.) is a time-lag function; whereby m xklObtaining a gray scale map Rk;
7) Will gray scale image RkDivision into subblocks S by 3X 3ij,i,j=1,2,...,n,n=m/3;
8) For each subgraph SijCalculating the direction quantity of the node in the horizontal direction and the vertical direction;
9) by using
x'=-a×x+(b-x)×f(x)
Calculating the output quantity Y of the neural networklob;
10) According to YlobCalculating the early warning threshold value amount according to the current and historical data and the early warning scheme;
11) and (3) outputting: and (5) early warning signals.
(2) The provided large-scale moving target behavior activity detection and early warning technology based on locust visual nerve is a detection and monitoring technology designed based on a locust visual nerve network structure, and is characterized in that the behavior activity of a large-scale moving target group is detected based on the structural characteristics of a locust visual nerve system, the visual response mechanism of LGMD neurons in locust visual leaves and the retina processing mechanism of mammals. The method comprises the following specific steps:
1) inputting video image data, wherein the video image data mainly is a monitoring video image of a public place;
2) calculating the visual excitation of the cells in the P layer at the current moment;
3) calculating the amount of visual inhibition at the cells in layer I;
4) calculating the output quantity of each cell in the S layer;
5) calculating the variation of the space-time energy caused by the activities of the crowd and calculating the motor behavior quantity of the neuron;
6) calculating a membrane potential excitation value output by the neuron;
7) and sending a crowd escape early warning signal according to the membrane potential excitation value and the threshold scheme.
(3) The provided intelligent video monitoring system is a set of software and hardware monitoring system, and the functions of the system comprise automatic monitoring and early warning of vehicle access control abnormity, indoor and outdoor local area and wide area dynamic monitoring and early warning. The system consists of a detection and early warning technology, a computer, a network camera and a switch. The network camera is used for collecting video images in the monitoring area and transmitting the images to the computer through the switch; the switch is used for connecting the computer and each network camera and is responsible for image information transmission and IP address allocation of each network device; after receiving the video image, the computer preprocesses the image information; then, the automatic entrance guard detection is realized by utilizing a license plate recognition technology in parallel, the abnormal detection is carried out on the local environment by utilizing a drosophila visual neural network technology, and the behavior activity of a large-scale moving target group in a wide area is detected by utilizing a locust visual neural network technology; if the early warning triggering condition is met, an alarm signal is generated, abnormal information is displayed by the display, and the alarm signal is output through audio. The implementation steps of the system are as follows:
1) a user logs in the system by using a user name and a password;
2) the user selects corresponding functions according to the scene, and the system provides the selected functions at the moment, wherein the selected functions comprise local area environment monitoring, wide area environment monitoring and vehicle access control monitoring;
3) the system operates:
3.1 when the system selects local environment monitoring, the implementation steps are as follows:
3.1.1 the system collects image information through a camera;
3.1.2 detecting whether the moving target in the obtained image information has abnormal conditions by using a drosophila visual neural network model;
3.1.3 if the abnormal condition is determined after the system detects, the abnormal condition possibly exists in the target area shot by the camera, and the system sends alarm information;
3.1.4 if no abnormal condition is determined after the system detects, the abnormal condition does not exist in the target area shot by the camera, and the system does not send alarm information;
3.1.5 after completing the detection and alarm tasks, returning to the step 3.1.1;
3.2 when the system selects wide area environment monitoring, the implementation steps are as follows:
3.2.1 the system collects image information through a camera;
3.2.2 detecting whether the moving target in the obtained image information has abnormal conditions by using the locust visual neural network model;
3.2.3 if the abnormal condition is determined after the system detects, the abnormal condition possibly exists in the target area shot by the camera, and the system sends alarm information.
3.2.4 if no abnormal condition is determined after the system detects, the abnormal condition does not exist in the target area shot by the camera, and the system does not send alarm information;
3.2.5 after completing the task of detecting and alarming, returning to step 3.2.1;
3.3 when the system selects vehicle entrance guard monitoring, the implementation steps are as follows:
3.3.1 the system collects image information through a camera;
3.3.2, using a license plate recognition technology to carry out license plate recognition on the obtained image, detecting whether a vehicle exists in a camera shooting area, and carrying out license plate extraction on the condition that the vehicle appears;
3.3.3, if the system detects that no vehicle appears in the shooting area of the camera, returning to the step 3.3.1;
3.3.4 if the system detects that the camera shooting area contains the vehicle information, comparing the license plate information with the vehicle information in the database, and detecting whether abnormal vehicles exist in the target area;
3.3.5 if the system detects that the abnormal vehicle exists, the system sends alarm information;
3.3.6 if the system detects that the abnormal vehicle does not appear, the system does not send out early warning;
3.3.7 returning to step 3.3.1;
4) the system exits.
Compared with the prior art, the invention has the beneficial effects that:
1. the invention has low cost, can finish the detection and monitoring of abnormal behaviors in local and wide-area environments only by a network camera, a switch and a computer, realizes video intelligent monitoring and achieves the aim of unattended operation;
2. the anomaly detection technology is a visual neural network designed based on a biological visual neural mechanism, has high speed and low sensitivity, and can perform anomaly detection and early warning in time;
3. the invention has strong expandability and can add other peripheral equipment to achieve the corresponding purpose.
Drawings
FIG. 1 is a flow chart of the present invention for detecting abnormal activity behavior of moving objects of one or more moving objects within a local area;
FIG. 2 is a flow chart of the detection of abnormal activity behavior of a large-scale moving object in a wide area according to the present invention;
fig. 3 is a flow chart of an implementation of the video surveillance system.
Detailed Description
The following takes the monitoring and early warning of the crowd abnormity as an example, and the invention is further described in detail by combining the accompanying drawings and the embodiment.
Fig. 1 to 2 are flowcharts of a detection and early warning technique according to an embodiment of the present invention, and fig. 3 is a schematic block diagram of an early warning system according to an embodiment of the present invention. The anomaly detection technology of the invention consists of two parts, namely local anomaly behavior detection and monitoring based on the visual nerve of the drosophila and wide-area anomaly behavior detection and monitoring based on the visual nerve of the locust.
The invention obtains a visual neural network technology for identifying abnormal conditions of single or multiple moving targets (such as pedestrians) and large-scale moving target groups (such as crowds) from the visual neural mechanism of locust and fruit flies, and obtains an automatic entrance guard monitoring and early warning technology by means of a computer geometry and image processing method, thereby developing an unattended monitoring system with three functions of automatic entrance guard detection, dynamic monitoring in a local area and large-scale moving target group behavior and activity monitoring in a wide area
The fruit fly visual nerve-based local abnormal behavior monitoring and early warning is a bionic artificial neural network provided according to a fruit fly neural network structure, can be used for performing behavior detection on the moving target according to conditions such as speed, size and moving direction of the moving target, and is high in detection speed, high in accuracy and good in real-time performance.
The specific detection and early warning implementation scheme comprises the following steps:
1) the size of the gray scale image is M multiplied by M, the sampling time interval is T, the unit is second, and M is a multiple of 8;
2) inputting: the k frame gray scale image Pk;
3) According to PkAnd Pk-1Generating a differential map Ek(ii) a Will EkEqually divided into subfigure sequences Fij,i,j=1,2,...,M/8;
4) For each (i, j), a 3 × 3 template pair F is utilizedijPerforming convolution with FijConversion into 6 × 6 sub-gray-scale map Gij;
5) Sequentially converting the gray scale image GijPerforming splicing (i, j ═ 1, 2.., M/8) to obtain a grayscale map Q with the size of M × Mk,m=0.75M;
6) According to the following formula
x'kl=-a×xkl+(b-xkl)×f(xkl),k,l=1,2,...,m;
Calculating the membrane potential x of the on-off channelkl(ii) a Whereby m xklObtaining a gray scale map Rk;
7) Will gray scale image RkDivision into subblocks S by 3X 3ij,i,j=1,2,...,n,n=m/3;
8) For each subgraph SijCalculating the direction quantity of the node in the horizontal direction and the vertical direction;
9) by using
x'=-a×x+(b-x)×f(x)
Calculating the output quantity Y of the neural networklob;
10) According to YlobCalculating the early warning threshold value amount according to the current and historical data and the early warning scheme;
11) and (3) outputting: and (5) early warning signals.
The invention discloses a large-scale moving target group activity behavior detection technology based on locust visual nerves, and relates to a moving target group behavior detection neural network invented according to a locust visual nerve mechanism. The network can effectively sense dangerous motion behaviors of the moving target in the visual field and quickly respond to the dangerous motion behaviors. The technology is designed according to the characteristics of layered visual signal processing structures in locust visual nerves, and the locust visual nerves sequentially comprise three cell layers and large-scale moving target behavior activity detection neurons. Wherein, the first layer calculates the lumen brightness variation between two adjacent frames in the input video sequence and takes the lumen brightness variation as the perceived visual excitation; the second layer delays the output of the first layer by one frame to obtain the visual inhibition amount; the third layer performs fusion treatment on the visual excitation amount and the visual inhibition amount according to the treatment mode of the visual signals in the retina of the mammal; finally, the neuron receives the output signal of the third layer and detects the abnormal state by adopting a specific dynamic threshold early warning scheme. The specific monitoring and early warning implementation scheme comprises the following steps:
1) inputting video image data, wherein the video image data mainly is a monitoring video image of a public place;
2) calculating the visual excitation of the cells in the P layer at the current moment;
3) calculating the amount of visual inhibition at the cells in layer I;
4) calculating the output quantity of each cell in the S layer;
5) calculating the variation of the space-time energy caused by the activities of the crowd and calculating the motor behavior quantity of the neuron;
6) calculating a membrane potential excitation value output by the neuron;
7) and sending a crowd escape early warning signal according to the membrane potential excitation value and the threshold scheme.
The intelligent video monitoring system is based on the license plate recognition technology and the local area and wide area abnormal behavior monitoring and supervision intelligent video monitoring system; the system has three functions of automatic entrance guard detection, local area dynamic monitoring and abnormal monitoring and early warning in a wide area environment, and can detect and monitor multi-channel video data in real time and realize unattended video monitoring.
Of course, the above is only a specific application example of the present invention, and other embodiments of the present invention are also within the scope of the present invention.
Claims (4)
1. The utility model provides an intelligence video monitoring and early warning system based on biological vision which characterized in that: the system consists of three functional subsystems, namely abnormal monitoring of one or more moving targets in a local area based on drosophila visual nerves, monitoring of large-scale moving target groups in a wide area based on locust visual nerves and automatic monitoring of entrance guard vehicles in parking lot scenes; the abnormal monitoring of single or multiple moving targets in the local area of the visual nerve of the fruit fly is realized by providing a visual nerve network of the fruit fly, and the behavioral activity of the moving targets is automatically monitored and early warned; the wide-area large-scale moving target group monitoring based on the locust visual nerve is realized by providing a locust visual nerve network, and automatically detecting and early warning the behavior of the moving target group in a wide-area scene; the automatic monitoring of the entrance guard vehicle is to automatically identify the vehicle according to the vehicle video and automatically early warn abnormal vehicles.
2. The intelligent video monitoring and early warning system based on biological vision of claim 1, characterized in that: the abnormal monitoring of the single or multiple moving targets in the local area based on the visual nerve of the fruit flies is to provide a visual nerve network of the fruit flies with an input of a video sequence under a static scene according to a visual nerve information processing mechanism of the fruit flies, monitor the behavior activity of the moving targets in the local area, automatically monitor the abnormal behavior and send out an early warning; the abnormal monitoring of the single or a plurality of moving targets in the local area is carried out according to the following steps:
1) the size of the gray scale image is M multiplied by M, the sampling time interval is T, the unit is second, and M is a multiple of 8;
2) inputting: the k frame gray scale image Pk;
3) According to PkAnd Pk-1Generating a differential map Ek(ii) a Will EkEqually divided into subfigure sequences Fij,i,j=1,2,...,M/8;
4) For each (i, j), a 3 × 3 template pair F is utilizedijPerforming convolution with FijConversion into 6 × 6 sub-gray-scale map Gij;
5) Sequentially converting the gray scale image GijPerforming splicing (i, j ═ 1, 2.., M/8) to obtain a grayscale map Q with the size of M × Mk,m=0.75M;
6) According to the following formula
x'kl=-a×xkl+(b-xkl)×f(xkl),k,l=1,2,...,m;
Calculating the membrane potential x of the on-off channelkl(ii) a Whereby m xklObtaining a gray scale map Rk;
7) Will gray scale image RkDivision into subblocks S by 3X 3ij,i,j=1,2,...,n,n=m/3;
8) For each subgraph SijCalculating the direction quantity of the node in the horizontal direction and the vertical direction;
9) by using
x'=-a×x+(b-x)×f(x)
Calculating the output quantity Y of the neural networklobWhere a and b are non-negative constants and f (.) is a time-lag function;
10) according to YlobCalculating the early warning threshold value amount according to the current and historical data and the early warning scheme;
11) and (3) outputting: and (5) early warning signals.
3. The intelligent video monitoring and early warning system based on biological vision of claim 1, characterized in that: the wide-area large-scale moving target group monitoring based on the locust visual nerve is to monitor the activity of the large-scale moving target group based on the structural characteristics of the locust visual nerve system, the visual response mechanism of LGMD neurons in locust visual leaves and the retina processing mechanism of mammals, and automatically monitor and send out early warning for abnormal behaviors; the anomaly monitoring of the large-scale moving target group in the wide area is carried out according to the following steps:
inputting a video frame;
step 1, setting parameters: the resolution, the signal threshold, the signal transmission coefficient, the membrane potential excitation threshold, the gain coefficient, the neuron spike counting threshold and the time step of the video frame;
step 2, calculating the visual excitation of the cells in the P layer at the current moment;
step 3, calculating the visual inhibition amount of the cells in the layer I;
step 4, calculating the output quantity of each cell in the S layer;
step 5, calculating the variation of the space-time energy caused by the movement of the crowd and calculating the movement behavior quantity of the neuron;
step 6, calculating a membrane potential excitation value output by the neuron;
and 7, sending a crowd escape early warning signal according to the membrane potential excitation value and the threshold scheme.
4. The intelligent video monitoring and early warning system based on biological vision of claim 1, characterized in that: the method is characterized in that a single or multiple moving target abnormity detection and monitoring technology based on the visual nerve of drosophila, a large-scale moving target group abnormity detection and early warning technology based on the visual nerve of locust and a license plate recognition technology are integrated to design a set of unattended monitoring system with three functions of automatic entrance guard detection and early warning, local area dynamic monitoring and large-scale moving target group behavior activity monitoring, so that multi-channel video data are monitored in real time, and video monitoring is unattended; the implementation process comprises the following steps:
1) a user logs in the system by using a user name and a password;
2) the user selects corresponding functions according to the scene, and the system provides the selected functions at the moment, wherein the selected functions comprise local area environment monitoring, wide area environment monitoring and vehicle access control monitoring;
3) the system operates:
3.1 when the system selects local environment monitoring, the implementation steps are as follows:
3.1.1 the system collects image information through a camera;
3.1.2 detecting whether the moving target in the obtained image information has abnormal conditions by using a drosophila visual neural network model;
3.1.3 if the abnormal condition is determined after the system detects, the abnormal condition possibly exists in the target area shot by the camera, and the system sends alarm information;
3.1.4 if no abnormal condition is determined after the system detects, the abnormal condition does not exist in the target area shot by the camera, and the system does not send alarm information;
3.1.5 after completing the detection and alarm tasks, returning to the step 3.1.1;
3.2 when the system selects wide area environment monitoring, the implementation steps are as follows:
3.2.1 the system collects image information through a camera;
3.2.2 detecting whether the moving target in the obtained image information has abnormal conditions by using the locust visual neural network model;
3.2.3 if the abnormal condition is determined after the system detects, the abnormal condition possibly exists in the target area shot by the camera, and the system sends alarm information;
3.2.4 if no abnormal condition is determined after the system detects, the abnormal condition does not exist in the target area shot by the camera, and the system does not send alarm information;
3.2.5 after completing the task of detecting and alarming, returning to step 3.2.1;
3.3 when the system selects vehicle entrance guard monitoring, the implementation steps are as follows:
3.3.1 the system collects image information through a camera;
3.3.2, using a license plate recognition technology to carry out license plate recognition on the obtained image, detecting whether a vehicle exists in a camera shooting area, and carrying out license plate extraction on the condition that the vehicle appears;
3.3.3, if the system detects that no vehicle appears in the shooting area of the camera, returning to the step 3.3.1;
3.3.4 if the system detects that the camera shooting area contains the vehicle information, comparing the license plate information with the vehicle information in the database, and detecting whether abnormal vehicles exist in the target area;
3.3.5 if the system detects that the abnormal vehicle exists, the system sends alarm information;
3.3.6 if the system detects that the abnormal vehicle does not appear, the system does not send out early warning;
3.3.7 returning to step 3.3.1;
4) the system exits.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911083934.5A CN110796114B (en) | 2019-11-07 | 2019-11-07 | Intelligent video monitoring and early warning system based on biological vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911083934.5A CN110796114B (en) | 2019-11-07 | 2019-11-07 | Intelligent video monitoring and early warning system based on biological vision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110796114A true CN110796114A (en) | 2020-02-14 |
CN110796114B CN110796114B (en) | 2022-08-26 |
Family
ID=69443328
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911083934.5A Active CN110796114B (en) | 2019-11-07 | 2019-11-07 | Intelligent video monitoring and early warning system based on biological vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110796114B (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201315654Y (en) * | 2008-08-25 | 2009-09-23 | 云南正卓信息技术有限公司 | Special SkyEyes** intelligent monitoring system for prison |
CN103248867A (en) * | 2012-08-20 | 2013-08-14 | 苏州大学 | Monitoring method of intelligent video monitoring system based on multi-camera data fusion |
CN204069175U (en) * | 2014-03-14 | 2014-12-31 | 贵州大学 | A kind of video frequency following system based on machine vision technique |
CN104813993A (en) * | 2015-04-22 | 2015-08-05 | 中国矿业大学 | Automatic monitoring device and method of micro agricultural insects based on machine vision |
CN109033451A (en) * | 2018-08-21 | 2018-12-18 | 北京深瞐科技有限公司 | People's vehicle dynamic file analysis method and device |
CN109712110A (en) * | 2018-11-20 | 2019-05-03 | 北京农业智能装备技术研究中心 | A kind of real-time water consumption quantity monitoring method of crop based on machine vision and device |
CN209486212U (en) * | 2019-01-10 | 2019-10-11 | 东北石油大学 | A kind of intelligent appliance fault detection and control system device based on drosophila Optimal BP Algorithm |
-
2019
- 2019-11-07 CN CN201911083934.5A patent/CN110796114B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201315654Y (en) * | 2008-08-25 | 2009-09-23 | 云南正卓信息技术有限公司 | Special SkyEyes** intelligent monitoring system for prison |
CN103248867A (en) * | 2012-08-20 | 2013-08-14 | 苏州大学 | Monitoring method of intelligent video monitoring system based on multi-camera data fusion |
CN204069175U (en) * | 2014-03-14 | 2014-12-31 | 贵州大学 | A kind of video frequency following system based on machine vision technique |
CN104813993A (en) * | 2015-04-22 | 2015-08-05 | 中国矿业大学 | Automatic monitoring device and method of micro agricultural insects based on machine vision |
CN109033451A (en) * | 2018-08-21 | 2018-12-18 | 北京深瞐科技有限公司 | People's vehicle dynamic file analysis method and device |
CN109712110A (en) * | 2018-11-20 | 2019-05-03 | 北京农业智能装备技术研究中心 | A kind of real-time water consumption quantity monitoring method of crop based on machine vision and device |
CN209486212U (en) * | 2019-01-10 | 2019-10-11 | 东北石油大学 | A kind of intelligent appliance fault detection and control system device based on drosophila Optimal BP Algorithm |
Non-Patent Citations (1)
Title |
---|
胡常俊: "基于果蝇视觉机理的车辆预警算法及其应用", 《中国优秀硕士学位论文全文数据库 (工程科技Ⅱ辑)》 * |
Also Published As
Publication number | Publication date |
---|---|
CN110796114B (en) | 2022-08-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Andrade et al. | Hidden markov models for optical flow analysis in crowds | |
CN105426820B (en) | More people's anomaly detection methods based on safety monitoring video data | |
CN103517042B (en) | A kind of nursing house old man's hazardous act monitoring method | |
CN106006266B (en) | A kind of machine vision method for building up applied to elevator safety monitoring | |
CN106241534B (en) | More people's boarding abnormal movement intelligent control methods | |
CN105744232A (en) | Method for preventing power transmission line from being externally broken through video based on behaviour analysis technology | |
CN102254394A (en) | Antitheft monitoring method for poles and towers in power transmission line based on video difference analysis | |
KR20200071799A (en) | object recognition and counting method using deep learning artificial intelligence technology | |
CN100565557C (en) | System for tracking infrared human body target based on corpuscle dynamic sampling model | |
CN113963301A (en) | Space-time feature fused video fire and smoke detection method and system | |
CN105227907A (en) | Based on the nothing supervision anomalous event real-time detection method of video | |
Marcomini et al. | A comparison between background modelling methods for vehicle segmentation in highway traffic videos | |
CN112926522B (en) | Behavior recognition method based on skeleton gesture and space-time diagram convolution network | |
CN115546899A (en) | Examination room abnormal behavior analysis method, system and terminal based on deep learning | |
CN114359976A (en) | Intelligent security method and device based on person identification | |
CN114267082A (en) | Bridge side falling behavior identification method based on deep understanding | |
CN116776279A (en) | Multi-mode data collaborative power transmission line flash explosion early warning abnormal target detection method | |
Chauhan et al. | Study of moving object detection and tracking for video surveillance | |
CN101299276A (en) | Method and system for distributed multiple target tracking | |
CN110796114B (en) | Intelligent video monitoring and early warning system based on biological vision | |
CN115083229B (en) | Intelligent recognition and warning system of flight training equipment based on AI visual recognition | |
CN111160150A (en) | Video monitoring crowd behavior identification method based on depth residual error neural network convolution | |
CN115147921B (en) | Multi-domain information fusion-based key region target abnormal behavior detection and positioning method | |
CN115909144A (en) | Method and system for detecting abnormity of surveillance video based on counterstudy | |
Frejlichowski et al. | Extraction of the foreground regions by means of the adaptive background modelling based on various colour components for a visual surveillance system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
CB03 | Change of inventor or designer information |
Inventor after: Zhang Zhuhong Inventor after: Yang Changxi Inventor after: Hu Bin Inventor after: Hu Changjun Inventor before: Zhang Zhuhong Inventor before: Yang Changxi Inventor before: Hu Bin |
|
CB03 | Change of inventor or designer information | ||
GR01 | Patent grant | ||
GR01 | Patent grant |