KR101075550B1 - Image sensing agent and security system of USN complex type - Google Patents

Abstract

According to the present invention, when analyzing a video object through an image object recognition algorithm in determining whether an event occurs by installing a surveillance camera and a sensor node in an event occurrence monitoring area, the reliability and accuracy according to the event are improved by combining USN sensor technology. It relates to an image sensing agent and a USN hybrid security system capable of obtaining the.
The present invention provides a surveillance camera which is installed in an event occurrence area and outputs a captured video signal to a data transmission means; It consists of a number of sensors installed in the event generating area and applies a network for communication between each other to perform a computing function to detect the sensing signal that detects a certain type of temperature, humidity, and pressure value that can occur in the event generating area. A sensor node outputting to the transmission means; Data transmission means for encoding the analog video signal transmitted from the surveillance camera and the analog detection signal transmitted from the sensor node, converting the analog video signal into a digital signal and transmitting the digital signal to a video recognition server system unit; After receiving the image data and the sensor data transmitted from the data transmission means, the image data is streamed to the control center for real-time monitoring, driving the image recognition processor to analyze the image of the object through the event detection algorithm An image recognition server system unit configured to extract data, drive sensor data processing unit to extract sensor analysis data through a predetermined threshold of a sensor, and determine whether an event occurs by combining the image analysis data and sensor analysis data; Image sensing agent and USN complex security system.

Description

Image sensing agent and security system of USN complex type}

The present invention relates to an image sensing agent and a USN complex security system. More particularly, the present invention relates to an image sensing agent and a USN complex security system. The present invention relates to an image sensing agent and a USN complex security system that combine USN sensor technology to obtain a result of improved reliability and accuracy according to an event occurrence.

In general, image recognition technology using CCTV (Close Circuit Television) has been used as a traditional method for recognizing and locating objects. The image recognition technology has been newly constructed after storing and storing image database for each object. In comparison with the input image, a method of recognizing the image object has been used.

The CCTV is a device that transmits video using a cable television in a specific building or a specific facility, and its use is widely used for industrial, educational, medical, disaster prevention, and in-house image information transmission. Used as a means of visual expansion, it is widely used in various fields for the purpose of observing remote places, observing places where the human body is impossible to approach, observing multiple people simultaneously or intensively monitoring.

The image recognition system using the CCTV to monitor the abnormality in the surveillance area using the CCTV, it is installed in the CCTV monitoring place, the video signal taken by the CCTV is displayed by transmitting to the monitor installed in the central management room, The user can keep an eye on the screen displayed on the monitor device to monitor the occurrence of abnormal conditions such as theft or fire.

In the conventional CCTV image recognition system, the administrator can be placed and monitored in the central management room and simply display the image captured by the CCTV through a monitor. Even if an abnormal situation such as a fire alarm (hereinafter referred to as an event) occurs, if the administrator does not grasp in real time, there is a problem in that it is not possible to appropriately deal with the event situation.

Recently, with the advent of ubiquitous technology, many changes have been made to CCTV camera and DVR (Digital Video Recorder) technology, and conventional DVR products are gradually being converted into NVR (Network Video Recorder) products with enhanced network. Conventional CCTV video system is an intelligent video system that integrates these technologies due to the limitations of direct monitoring and analysis by humans due to the rapid spread of CCTV cameras, increasing demand for real-time monitoring system rather than post processing, and rapid development of IT technology. Call for new evolution.

The intelligent video security system includes functions such as moving object detection, classification, tracking, and behavior recognition based on automatic image analysis and analysis technology. The main fields of the intelligent video security include image-based moving object detection, tracking, and video. Based on personal identification and identification, large-scale video security system construction technology.

The intelligent video security system is a change detection, region localization, region tracking modules that are commonly required regardless of a specific application area, and an object classification requiring a function change according to an application. It is divided into variable modules such as classification, object identification and tracking.

The motion detection is a process of finding a region different from the previous image in the input image, and the feature can be used to detect and track the region by using a feature, and the object classification is performed by using an object model previously classified and defined. This classifier can classify motions by classifying them, recognize them, and track them, and use these predefined behavioral models.

On the other hand, due to the recent rapid development of the information and communication environment, ubiquitous, a communication environment that can be connected to the network regardless of time and place, is in the spotlight.In the information and communication field, service competition aims at a communication environment that transcends time and place. In addition, various mobile communication devices are released with products that emphasize portability and convenience, and can turn on and off the boilers in the home by telephone or the Internet, and control the machines in overseas plants from their seats. It refers to an information technology environment that makes it easy to communicate with users by inserting computers into various devices or objects.

In addition, Ubiquitous Sensor Network (USN), which is a network configured to collect information detected by various sensors wirelessly, is also spotlighted as an infrastructure for achieving ubiquitous, and the USN is a vulnerable district inaccessible to humans. Hundreds of sensor network nodes are installed in the network to enable human monitoring. Sensor network technology is very active with the development of wireless personal area network (WPAN) technology and micro network device technology. In the USN, the most important technology is reliability to quickly transmit accurate information, and power consumption is also very important. Currently, USN technology, which can be used for one year with a single battery, has been developed in Korea.

Therefore, each image analysis data is combined with image object recognition technology that detects and analyzes the object through image transmitted from the video surveillance apparatus installed in various places and facilities, and sensor technology that detects the object using USN sensor nodes. The need for a more accurate and reliable security system by combining data with sensor data is emerging.

The present invention has been made to solve the above-mentioned conventional problems, the object recognition according to the occurrence of the event through the image analysis data by analyzing the image captured by the surveillance camera installed in the event location and the detection signal detected through the sensor node In order to prevent accidents by increasing the reliability and accuracy of object recognition according to event occurrence using sensor analysis data.

Surveillance camera installed in the event generation area to achieve the above object and outputs the video signal to the data transmission means; It consists of a number of sensors installed in the event generating area and applies a network for communication between each other to perform a computing function to detect the sensing signal that detects a certain type of temperature, humidity, and pressure value that can occur in the event generating area. A sensor node outputting to the transmission means; Data transmission means for encoding the analog video signal transmitted from the surveillance camera and the analog detection signal transmitted from the sensor node, converting the analog video signal into a digital signal and transmitting the digital signal to a video recognition server system unit; After receiving the image data and the sensor data transmitted from the data transmission means, the image data is streamed to the control center for real-time monitoring, driving the image recognition processor to analyze the image of the object through the event detection algorithm An image recognition server system unit configured to extract data, drive sensor data processing unit to extract sensor analysis data through a predetermined threshold of a sensor, and determine whether an event occurs by combining the image analysis data and sensor analysis data; It is intended to implement an image sensing agent and a USN complex security system.

The image recognition server system unit may be configured to transmit and receive image data, sensor data, and command signals from a data transmission means, and to transmit the image data to a control center so that real-time monitoring may be performed; Streaming the image information transmitted from the data transmission means to the control center to control the real-time monitoring, driving the sensor data processing unit and the image recognition processing unit by combining the processed image analysis data and sensor analysis data through the event situation detection unit A system operating unit supporting a control program and data to obtain a result value of whether an event occurs; A monitoring provider configured to receive and execute a control signal of the system operating unit to stream image data transmitted through a data transmission unit to a control center so that real-time monitoring can be performed; A sensor data processor that compares sensor data transmitted from a sensor node installed in an event occurrence region with threshold information of a corresponding sensor stored in a database unit, determines whether an event occurs, and transmits a result value to an event situation detector; An image recognition processor for comparing an image data transmitted from a surveillance camera installed in an event occurrence area with an original image of a corresponding place stored in a database unit to determine whether an event has occurred, and to transmit a result value to an event situation detection unit; When a true value, which is a warning value for an event occurrence, is output through the sensor data processing unit and the image recognition processing unit, it is determined that an event situation has occurred due to a true value generated through an AND logic operation. An event situation detector for outputting a predetermined alarm sound; A database manager executed to extract or store data stored in the database storage; The video information database unit classifies and stores the original video information on the place where the surveillance camera is installed as an event area according to the classification system and the threshold value for each sensor for the place where the sensor node is installed. And a database storage unit formed of a sensor information database unit which is classified and stored according to a category according to a predetermined classification system.

When the detection signal transmitted from the sensor node is received, the sensor data processor extracts threshold information of the corresponding sensor stored in the sensor information database to determine whether the received sensor information value passes the preset threshold of the corresponding sensor. After the determination, the threshold value elapses, and a true value, which is a warning value for an event occurrence, is output.

When the image data captured by the surveillance camera receives the image data, the image recognition processor extracts an original image of a corresponding place stored in the image information database unit and calculates a predetermined threshold value by using RGB color information with the received image data. After determining whether the elapsed time is passed, and if the threshold value passes, it is configured to output a true value, which is a warning value for an event occurrence, through a filtering process.

The present invention combines video object recognition technology and sensor network technology to recognize intelligent video objects for event situations such as intruder monitoring or fire occurrence, but uses the sensor data to provide reliable and accurate results of event occurrence. By raising the value, the user can be alerted by the occurrence of an event to the outside, thereby improving security.

1 is an overall configuration diagram of an image sensing agent and a USN hybrid security system applied to the present invention.
2 is a flowchart illustrating an image recognition process of an image sensing agent and a USN hybrid security system applied to the present invention.

Hereinafter, an image sensing agent and a USN hybrid security system applied to the present invention will be described in detail with reference to the accompanying drawings.

1 is an overall configuration diagram of an image sensing agent and a USN hybrid security system applied to the present invention, and FIG. 2 is a flowchart illustrating an image recognition process of the image sensing agent and the USN hybrid security system applied to the present invention.

As shown, the image signal captured by the surveillance camera 10 installed in the event generation area and the detection signal detected through the sensor node 20 is transferred to the image recognition server system unit 40 through the data transmission means 30. The image recognition server system unit 40 transmits an object recognition algorithm that determines an event situation such as a fire or smoke detection, and combines each image analysis data and sensor data to provide a reliable result in real time for the event situation. Configure to output the value.

The surveillance camera 10 transmits a video signal photographed by installing in a predetermined event occurrence area to a data transmission means 30 using a closed circuit television (CCTV) camera, and pans and tilts according to a control signal. Tilt) and Zoom function can be used to capture the image of the installed area using a PTZ program.

The sensor node 20 is installed in each event generating area using a USN sensor, and transmits the sensed signals to the data transmission means 30. Each sensor node 20 applies a network for wireless communication with each other. It is designed to detect and efficiently compute with each other, and to detect various types of temperature, humidity, and pressure values that can occur in each event area.

The data transmission means 30 encodes an analog video signal transmitted from the surveillance camera 10, converts the analog video signal into a digital video, and transmits the converted video signal to the image recognition server system 40, and the analog detection signal transmitted from the sensor node 20. Converts the signal into a digital signal and transmits the converted data to the image recognition server system 40.

The image recognition server system unit 40 may include a connection unit 41, a system operating unit 42, a monitoring provider 43, a sensor data processor 44, an image recognition processor 45, an event situation detector 46, It consists of a database management unit 47 and a database storage unit 48, so that when a large capacity is required to use a separate application server.

The connection unit 41 may transmit and receive the image data, the sensed sensor data, and the command signal photographed from the data transmission means 30 in which a session is formed over the Internet, and transmit the transmitted image data to the control center 50. To be monitored in real time.

The system operator 42 may monitor the image information transmitted from the data transmission means 30 in real time in the control center 50, and the image data and sensor data transmitted through the data transmission means 30. Upon reception, the sensor data processor 44 and the image recognition processor 45 are driven to combine the image analysis data through the corresponding event detection algorithm with the sensor analysis data using a predetermined threshold value of the sensor to detect the event situation 46. Supports the control program and data to determine whether an event has occurred.

The monitoring provider 43 may be executed by receiving a control signal from the system operator 42 to stream the image data transmitted through the data transmission means 30 to the control center 50 to perform real time monitoring. To help.

The sensor data processor 44 is implemented in software to determine whether an event occurs through sensor data transmitted from the sensor node 20 installed in the event generating region, and is detected through the sensor node 20. When the sensor data is received, the threshold information of the corresponding sensor stored in the sensor information database unit 48b is extracted, and after determining whether the received sensor information value passes the preset threshold of the corresponding sensor, the elapsed time passes. The result value of whether or not an event is generated, that is, a Treu value, which is a warning value of event occurrence, is transmitted to the event situation detection unit 47.

The image recognition processor 45 implements an event detection algorithm in software to determine whether an event occurs through the image data transmitted from the surveillance camera 10 installed in the event generation region. When the captured image data is received, the original image of the corresponding place stored in the image information database unit 48a is extracted, and whether RGB color information with the received image data passes a predetermined threshold through a calculation process. After the determination, the elapsed value of the event, that is, a True value, which is a warning value for the occurrence of the event, is transmitted to the event situation detector 47.

In this case, the operation module may perform object recognition on whether an event occurs through the image photographed by the surveillance camera 10 by using the database information on the event situation, which is captured by the surveillance camera 10. Object recognition can be achieved by extracting feature vectors that are robust to view, size, and lighting from images and comparing them with feature vectors of original images stored in a database.

The event situation detector 46 generates a warning value (True value) for event occurrence through the sensor data processor 44 and the image recognition processor 45, that is, generates a True value through an AND logic operation. In this case, it is determined that an event situation occurs so that a predetermined alarm sound such as a siren can be output.

The database manager 47 controls to extract and store data stored in the database storage unit 48 by receiving a control signal transmitted from the system operating unit 42, but includes a Common Gateway Interface (CGI) and an ASP (Active). Server Pages), PHP (Hypertext Preprocessor) scripts such as access to the database storage unit 48 to control the function to extract the data and to save the input data, or to delete and process the data.

The database storage unit 48 is composed of an image information database unit 48a and a sensor information database unit 48b. The image information database unit 48a is used for a place where the surveillance camera 10 is installed. Original image information is stored, and data necessary for recognizing an object through image processing is stored in the original image information, and specifically, other data such as image data, image feature amount or size, shape, etc. registered in advance for the object are stored. The information is classified and stored according to a category according to a predetermined classification system, and the sensor information database unit 48b is a threshold for a temperature value, a humidity value, a pressure value, etc. for a place where the sensor node 20 which is an event occurrence area is installed. The value is preset and classified and stored according to a category according to a predetermined classification system.

The operation of the image recognition security system using the USN-based sensor information configured according to an embodiment of the present invention as described above will be described.

The image signal captured by the surveillance camera 10 installed in the event generation area and the detection signal detected through the sensor node 20 are transmitted to the data transmission means 30 (S100), and the data transmission means 30 After encoding the analog video signal transmitted from the surveillance camera 10 to a digital video signal, and the analog detection signal transmitted from the sensor node 20 is encoded and converted into a digital signal, the image recognition server system unit 40 In this case, it may be transmitted in the form of video data compressed using the codec H.264 to improve security.

In this case, when an abnormal condition such as a specific event situation, that is, a fire alarm due to the occurrence of smoke or fire, access control of a door, detection of dynamic movement, or the like occurs among video information captured by the surveillance camera 10, According to the control signal transmitted from the recognition server system unit 40, the Pan, Tilt, and Zoom functions may be operated through a PTZ program embedded in the surveillance camera 10.

The image recognition server system unit 40 receives image data and sensor data transmitted from the data transmission means 30 through the connection unit 41, but the system operation unit 42 transmits the image data to the control center 50 in real time. In order to perform real-time monitoring by streaming streaming (S120), the sensor data processor 44 and the image recognition processor 45 are driven to analyze the sensor data and the image data to determine the occurrence of the event can be made. To control.

When the sensor data is received from the sensor node 20 installed at the event occurrence location as described above, the sensor data processing unit 44 is driven (S130) to extract the threshold information of the sensor stored in the sensor information database unit 48b ( S131), a sensor signal and a sensor signal such as a temperature value, a humidity value, a pressure value, and a concentration value of a detection signal detected at an event occurrence place transmitted from the sensor node 20, that is, the place where the sensor node 20 is installed. It is determined whether there is a sensor data value that passes the threshold value by comparing with the data about the threshold value for each sensor node 20 extracted from the information database unit 48b (S132), and the sensor data that passes the threshold value. If a value exists, a true value, which is a warning value for notifying occurrence of an event, is output (S133), and then transmitted to the event situation detection unit 46.

In addition, when image data is received from the surveillance camera 10 installed at the event occurrence place, the image recognition processing unit 45 is driven (S140) to extract the original image information stored in the image information database unit 48a (S141), It is determined whether the threshold value has elapsed by comparing the image information photographed at the event occurrence place transmitted from the surveillance camera 10 with the original image information extracted from the image information database unit 48a (S142), and the threshold value has elapsed. In the case of performing a filtering process (S143), a true value, which is a warning value indicating an event occurrence, is output (S144), and then transmitted to the event situation detection unit 46.

That is, in case of monitoring an event such as a fire occurrence at the event occurrence place, the RGB image information is transmitted in the case of the image data transmitted through the surveillance camera 10 and the original image of the corresponding place stored in the image information database unit 48a. Smoke can be recognized through two filtering processes. The first filtering method uses the difference image technique to determine the image data movement of the received smoke, and then uses the threshold value for the smoke. First, the smoke is judged, and when it is determined that the smoke is smoked, the motion of the smoke is determined by mathematical calculation through the second filtering, and when it is moved in various directions, it is determined that the smoke is smoked. The event situation detection unit 46 outputs the result.

The event situation detection unit 46 receives a true value, which is a warning value for the occurrence of an event output through the sensor data processing unit 44 and the image recognition processing unit 45, and then performs a true operation through an AND logic operation. When a value of True occurs, it is determined that an event situation such as a fire has occurred (S150), and an alarm sound such as a siren can be output to alert the user according to the event occurrence (S160).

As described above, the present invention combines an image object recognition technique for analyzing an object through an image captured by the surveillance camera 10 and a sensor technique for detecting an object using the sensor nodes 20 to generate an event. In order to output a result of improved reliability and accuracy with respect to the present invention, the present invention described above is not limited to those skilled in the art without departing from the spirit of the present invention. Since it is possible to substitute, modify, and change the branch, it is not limited to the above-described embodiment and the accompanying drawings.

10. Surveillance Camera 20. Sensor Node
30. Data transmission means 40. Image recognition server system unit
41. Connection 42. System Operation
43. Monitoring provider 44. Sensor data processor
45. Image recognition processor 46. Event situation detector
47. Database Manager 48. Database Store
48a. Image Information Database 48b. Sensor Information Database
50. Control Center

Claims (4)

A surveillance camera 10 installed at an event occurrence area and outputting the captured image signal to the data transmission means 30;
A sensor node 20 installed at an event occurrence area and outputting a detected signal to the data transmission means 30;
Data transmission means (30) for converting an image signal transmitted from the surveillance camera (10) and a detection signal transmitted from the sensor node (20) into a digital signal and transmitting the digital signal to an image recognition server system unit (40);
In the image sensing agent and USN complex security system consisting of; image recognition server system unit 40 for analyzing the image data and sensor data transmitted from the data transmission means 30 to determine whether an event occurs,
The image recognition server system unit 40 includes a connection unit 41 for receiving the transmitted image data and sensor data;
A system operator 42 controlling the video data to be streamed to the control center 50 so that real-time monitoring can be performed;
A monitoring provider 43 capable of monitoring image data in real time in the control center 50 under the control of the system operator 42;
An event situation detector 46 outputting a predetermined alarm sound when it is determined that an event situation occurs;
A database storage unit 48 for storing original image data and sensor data of an event occurrence area;
Database management unit 47 for extracting or storing the data stored in the database storage unit 48;
A sensor for comparing the sensor data transmitted from the sensor node 20 with the threshold information of the corresponding sensor stored in the database storage unit 48 to determine whether an event has occurred and transmitting the result value to the event situation detection unit 46. A data processor 44;
Image recognition processing unit for comparing the image data taken from the surveillance camera 10 with the original image of the place stored in the database storage unit 48 to determine whether the event occurs, and transmits the result value to the event situation detector 46 (45); Image sensing agent and USN complex security system, characterized in that consisting of. delete delete delete

Images