KR102300640B1 - System for analyzing cctv image information for occupancy realtime monitoring, and method for the same - Google Patents

Abstract

In the event of a disaster such as a fire in a high-rise building, in order to monitor the location and number of occupants (or visitors) moving to the exits such as emergency exits and elevators in real time, CCTV images can be collected and visualized to automatically and quickly and accurately analyze, According to the movement of occupants based on camera motion detection, even if several people move through the entrance at once, each occupant can be easily distinguished and recognized. As it is implemented as a dome-type IP camera with a built-in dynamic visual sensor (DVS) or CMOS image sensor (CIS) and thermal image sensor (TIS), it is possible to secure a stable detection rate without being affected under various lighting conditions, and also A CCTV image information analysis system and method for real-time monitoring of occupants, which can designate a recognition area according to a monitoring program, and can easily determine whether to enter or exit based on this designated recognition area, are provided.

Description

CCTV image information analysis system and method for real-time monitoring of occupants {SYSTEM FOR ANALYZING CCTV IMAGE INFORMATION FOR OCCUPANCY REALTIME MONITORING, AND METHOD FOR THE SAME}

The present invention relates to real-time monitoring of occupants, and more specifically, when a disaster such as a fire occurs in a high-rise building, CCTV images are collected in order to monitor the location and number of occupants (or occupants) moving to the exits such as emergency exits and elevators in real time. It relates to a CCTV image information analysis system and method for visualizing and automatically analyzing.

Recently, due to the concentration of population due to urbanization, large-scale, high-rise, and complex of various buildings are rapidly progressing. Accordingly, as the number of buildings that are relatively vulnerable to fire safety is increasing, there is an urgent need to establish effective fire countermeasures to secure human safety.

In particular, high-rise buildings, such as high-rise complex buildings, have a longer evacuation route than general buildings in case of fire because occupants are restricted in their accessibility to the outside, so the evacuation time is delayed. judged to be higher. That is, when a disaster occurs in a high-rise or large-scale building, it takes a long time to evacuate due to the characteristics of the building, and congestion increases. In addition, it is difficult to find an evacuation route because it is not possible to check information on the presence or location of occupants in the building despite the presence of occupants in the building. Accordingly, it is necessary to suggest a quick response method in the event of a disaster and to minimize human casualties.

According to the prior art, RFID (Radio Frequency InDentification) sensor, CCTV (Closed Circuit TeleVision), human body detection sensor (pyroelectric infrared sensor: PIR sensor), etc. are installed in the entrance and office area of a building, and the data detected through each sensor A technique for determining the presence or absence of an occupant based on the

Specifically, in the method using the RFID sensor, when a group of two or more people enters/exits, it is difficult to accurately measure the access data because the data is distorted due to misuse such as loss of ID card or not recognizing ID card when leaving.

In addition, the method using CCTV has the advantage of accurately counting the number of people entering and leaving the installed area by continuously detecting and analyzing the faces of people passing through a specific area, but it is difficult to install in the office due to privacy problems. There is a problem. Therefore, the conventional method using CCTV has limitations in measuring data for determining whether occupancy is present.

In addition, in the method using the human body sensor, the presence or absence of an occupancy can be accurately determined by installing the human body detecting sensor on the ceiling of a building and identifying the presence of a human body in a certain area. However, it is not possible to count the number of occupants, and if the human body sensor is affected by smoke and flames in case of a fire, there is a problem in that it may affect the determination of whether or not occupancy is present.

As a prior art for solving the above-described problems, the Republic of Korea Patent Registration No. 10-1871991 discloses an invention entitled “System and Method for Determining Occupancy Occupancy”, which will be described in detail with reference to FIGS. 1 and 2 . .

1 is a block diagram of a system for determining presence of occupants according to the prior art.

Referring to FIG. 1 , a occupant determination system according to the prior art includes a plurality of measurement sensors 10 installed in a building entrance, each floor, and a predetermined area, a data receiving unit 20, an entrance/exit analysis unit 30, and the presence or absence of an occupant. It includes a determination unit 40 , a correction unit 50 , and a database 60 .

The plurality of measurement sensors 10 are different types of measurement sensors, and are installed at an entrance, each floor, or a predetermined area in a building to measure whether or not a occupant enters or occupies the room. Here, the plurality of measurement sensors 10 are installed at the entrance or a certain area within the building, the RFID sensor 11 for detecting whether a occupant enters; a human body sensor 13 installed in a certain area of the building to detect the presence or absence of occupants; and a CCTV (12) installed at the entrance of the building, the emergency exit of each floor, or the elevator, and photographing the occupants entering and exiting the corresponding entrance or between the floors.

The data receiving unit 20 receives and collects measurement data sensed from a plurality of measurement sensors.

The entry/exit analysis unit 30 receives the measurement data sensed through the plurality of measurement sensors 10 from the data receiving unit 20, and the received measurement data is stored in the database 60, sensor installation information-here, the sensor installation The information includes the location information of the entrance, each floor, or a certain area of the building where a plurality of measurement sensors are installed - and based on the building access information, analyzes the entry and exit of the occupants in the building and the entry and exit location, Count the number of people entering each floor or in a certain area. Specifically, the entry/exit analysis unit 30 identifies the occupant ID and entry/exit location from the measurement data sensed by the RFID sensor 11 to analyze the entry and exit, and uses the measurement data captured from the CCTV 12 to determine the entrance or exit of the building. Analyze whether each floor has entered or not, and the number of people moving between each floor is analyzed.

The occupant presence determination unit 40 receives the measured data sensed through the plurality of measurement sensors 10 from the data receiving unit 20 , determines the presence or absence of occupants in the building, and counts through the entry/exit analysis unit 30 . Based on the number of people entering and leaving, the number of occupants present in each floor and in a certain area is calculated. Specifically, when calculating the number of occupants on each floor, the occupant presence determination unit 40 weights the measured data sensed from the RFID sensor 11 and the measured data sensed from the CCTV 12 based on Equation 1 However, when each floor is divided into two or more zones, the final number of occupants is calculated by reflecting the weight of each zone of the corresponding floor in the measurement data sensed from the CCTV 12 based on Equation 2 based on Equation 2.

[Equation 1]

(RFID 센서에 의한 재실자 수) + b

(CCTV에 의한 재실자 수)(단, a, b는 가중치이고 a < b 임)number of occupants = a

(number of occupants by RFID sensor) + b

(Number of occupants by CCTV) (where a and b are weights and a < b)

[Equation 2]

(RFID 센서에 의한 재실자 수) + {b

(CCTV에 의한 재실자 수)

(각 구역별 비중) }number of occupants = a

(Number of occupants by RFID sensor) + {b

(Number of occupants by CCTV)

(Proportion of each area) }

Proportion of each zone = RFID measurement value of the zone / {(RFID measurement value of zone A) + (RFID measurement value of zone B) + … } (where a and b are weights and a < b)

In this case, the specific gravity of each zone may be calculated by taking the total number of occupants of the RFID sensor 11 for all zones of the corresponding floor as the denominator, and using the number of occupants of the RFID sensor 11 in the corresponding zone as the numerator.

The correction unit 50 checks the number of people moving between floors based on the measured data sensed by the plurality of measurement sensors 10 , and corrects the number of people moving by subtracting the number of occupants from the number of occupants calculated by the occupancy determination unit 40 . . Specifically, the correction unit 50 determines whether the final number of occupants calculated by the occupant presence determination unit 40 or the number of RFID occupants calculated from the RFID sensor 11 exceeds the number of CCTV occupants calculated from the CCTV 12 . If it is determined and exceeds, the final number of occupants or the number of RFID occupants is corrected by the number of CCTV occupants calculated from the CCTV 12, and the number of CCTV occupants calculated from the CCTV 12 is the number of RFID occupants calculated from the RFID sensor 11 is abnormally exceeded, the weight for the number of RFID occupants is corrected upward, and the weight for the number of CCTV occupants is corrected downward.

Meanwhile, FIG. 2 is an operation flowchart for explaining a method of determining the presence or absence of an occupant using the occupancy determination system according to the related art.

1 and 2, in the method of determining the presence or absence of an occupant using the occupancy determination system according to the prior art, first, the occupant presence determination system collects data detected from a plurality of measurement sensors and Access information is modeled (S10). At this time, the modeling method is as follows: a random occupant enters the building from what time, and arbitrarily moves resident personnel in the following ways: toilet, smoking, access to other offices, elevators, stairs, outside work, use of the lobby on the 1st floor, group entry and exit, etc. can be modeled taking into account. Through this modeling, it is possible to statistically grasp the status of entering/exiting, moving, and occupancy of a building.

Next, the occupant presence determination system receives and collects the measured data sensed from the plurality of measurement sensors 10 by the data receiving unit 20 (S20).

Next, the entrance/exit analysis unit 30 of the occupant presence determination system receives the measurement data detected from the plurality of measurement sensors 10 from the data receiving unit 20 and based on the building entrance information modeling the received measurement data. Analyze whether or not to enter (S30). At this time, the analysis process of whether to enter or not can be divided according to whether it is the RFID sensor 11 or the CCTV 12 . Specifically, when the detected measurement data from the RFID sensor 11 is received. The entry/exit analysis unit 30 identifies the occupant ID and the entry/exit location from the measurement data (tag information) based on the sensor installation information and the building access information stored in the database 60, and corresponds to the identified entry/exit location by floor , and judges whether or not to enter each zone. In addition, when the detected measurement data from the CCTV 12 is received, the entry/exit analysis unit 30 may determine the entry/exit status based on the measurement data and whether the entry/exit is by floor, and may detect the inter-floor movement status.

Next, the entry/exit analysis unit 30 of the occupant presence determination system identifies an entrance, each floor, and a predetermined area corresponding to the corresponding entry/exit location, and counts the number of persons entering and exiting the corresponding entrance, each floor, and the predetermined area (S40).

Next, the occupancy determination unit 40 of the occupant presence determination system receives the measured data sensed through the plurality of measurement sensors 10 from the data receiving unit 20, and performs the measurement data and entry/exit analysis unit 30 Based on the number of entrances and exits counted through the system, the presence or absence of occupants is calculated according to the floor and area. Accordingly, the presence or absence of occupancy or the number of occupants may be predicted (S50).

Next, the correction unit 50 of the occupancy determination system checks and corrects calculation errors and errors with respect to the number of occupants calculated by the occupant presence determination unit 40 ( S60 ).

According to the occupancy determination system and method according to the prior art, by collecting and comprehensively analyzing the measurement data of a plurality of measurement sensors, the limitations of each measurement sensor can be supplemented, and the presence or number of occupants can be determined more accurately. have. In addition, it can be used to control the automatic opening and closing of a door or smoke window in case of a fire or disaster in a building, or it can be used as a crime prevention management for security. In addition, by providing an analysis algorithm using a plurality of measurement sensors, it is possible to utilize a plurality of measurement sensors that are built in an existing building or used in general. In addition, the number of occupants is calculated more accurately than the occupant prediction method using each individual sensor by subtracting the number of people moving between floors from the measured data taken through CCTV, correcting the error on the measured data calculated from each sensor and reflecting it in real time. can be calculated.

However, in the case of the occupancy determination system and the method according to the prior art, the configuration is complicated because a plurality of measurement sensors must be installed, and when CCTV is used, there is a limit in measuring data for determining whether occupancy is present. . For example, when the aforementioned CCTV is an IP camera, there are the following problems.

First, there is a problem in that there is no scalability because it is a method of reproducing video by inputting only the IP fixed in the IP camera in general. For example, it works only when the IP camera is connected to a fixed IP, and if the IP is not known, access is not possible. Also, since a fixed IP is assigned to the IP camera, individual tuning is required when adding equipment.

In addition, remote control of the IP camera is not possible, that is, operation method adjustment such as rebooting, resolution, and FPS designation is possible only at the IP camera installation site.

In addition, since the firmware of the camera equipment is used as it is without modification or supplementation, and in particular, it is not applicable when the camera model is changed because only a fixed specific folder is used. Since the settings of "motion detecting" and "detection area" are applied, only specific IP camera equipment must be used.

In addition, there is a problem that basic functions provided by the IP camera cannot be used properly. For example, a motion detection command can be performed by designating an area. In this case, although the area can be designated, there is a problem in that it is impossible to accurately detect a person due to shaking. For example, even when a motion is detected, it is not possible to distinguish whether the corresponding motion is a human or an animal (or object), and there is a problem in that a single counting is performed even when a plurality of people move together. Accordingly, there is a problem in that a useful detection result cannot be obtained as an actual execution result.

After all, when CCTV is used to determine the presence or absence of occupants, there is a need for a method for counting occupants more accurately.

Republic of Korea Patent No. 10-1871991 (application date: June 25, 2015), title of invention: "System and method for determining presence of occupants" Republic of Korea Patent No. 10-1492799 (application date: November 21, 2014), title of invention: "Access control integrated image storage system and method having access control function through tracking people entering and leaving the monitoring area" Republic of Korea Patent No. 10-1714479 (application date: February 5, 2015), title of invention: "Spatial information recognition method using optical sensor and mutual multi-object recognition type disaster response system using the same" Republic of Korea Patent No. 10-2010320 (application date: February 19, 2017), title of invention: "Occupancy check service device" Republic of Korea Patent No. 10-1972655 (Application Date: July 18, 2017), Title of Invention: "Occupant Detection Security System Using Ultrasound Recognition Technology" Republic of Korea Patent No. 10-1442658 (application date: January 15, 2013), title of invention: "Disaster evacuation system and method providing evacuation simulation" Republic of Korea Patent No. 10-1828005 (application date: November 22, 2017), title of invention: "Evacuation strategy support system and method using an evacuation elevator"

The technical task of the present invention to solve the above problems is to collect and visualize CCTV images in order to monitor in real time the location and number of occupants moving to the exits such as emergency exits and elevators when disasters such as fires occur in high-rise buildings. An object of the present invention is to provide a CCTV image information analysis system and method for real-time monitoring of occupants, which can be analyzed quickly and accurately automatically.

Another technical problem to be achieved by the present invention is real-time monitoring of the number of occupants that can easily distinguish and recognize each occupant even when several people move through the entrance at once according to the occupant's passage based on camera motion detection To provide a CCTV image information analysis system and method for

Another technical task to be achieved by the present invention is that each of a plurality of CCTVs is a dome-type IP camera with a CMOS image sensor (CIS) and a dynamic visual sensor (DVS) or a CMOS image sensor (CIS) and a thermal image sensor (TIS). It is to provide a CCTV image information analysis system and method for real-time monitoring of occupants, which can secure a stable detection rate without being affected under various illuminance environments as it is implemented.

As a means for achieving the above-described technical problem, the CCTV image information analysis system for real-time monitoring of occupants according to the present invention is installed at the entrance of each floor and each area of a building to generate a video stream. CCTV; a repeater that collects and transmits the video streams generated from each of the plurality of CCTVs; According to the CCTV monitoring program and automatic doorway occupant counting program, a video stream is collected from the repeater, and the collected video stream is processed and analyzed to count the occupants entering or exiting the doorway. CCTV image information processing device for generating and remotely controlling the plurality of CCTVs; and a control monitoring server for controlling and monitoring the entrance/passenger information processed by the CCTV image information processing device, wherein each of the plurality of CCTVs includes a CMOS image sensor (CIS) and a dynamic vision sensor: DVS) or a dome type IP (Internet Protocol) camera with a built-in CMOS image sensor (CIS) and a thermal imaging sensor (TIS); The CCTV image information processing device is implemented with a user interface for processing a CCTV monitoring program and an automatic doorway occupant counting program, and automatically detects the number of entrances and exits according to the occupant's movement line based on camera motion detection for real-time monitoring of the number of occupants; The CCTV image information processing apparatus includes: a video stream collecting unit for collecting video streams captured from the plurality of CCTVs via the repeater in real time; an entrance/exit detection unit that processes and analyzes the video stream collected from the video stream collection unit, and recognizes the entrance or exit of the occupant according to the movement of the occupant based on camera motion detection; an occupant counting unit automatically counting the occupants recognized by the entrance/exit detection unit to generate information on the number of people passing through the entrance; a data transmission unit for transmitting information on the number of people passing through the entrance, counted by the occupant counting unit, to the monitoring server; and a remote control unit for remotely controlling an operation method including rebooting, resolution, and number of frames per second of each of the plurality of CCTVs.

delete

Here, in each of the plurality of CCTVs, when sufficient illuminance is secured, the CMOS image sensor (CIS) operates to collect images, and when the illuminance is insufficient, the dynamic visual sensor (DVS) or the thermal image sensor (TIS) It can work instead to create a videostream.

Here, the CCTV image information processing apparatus may designate a recognition area in the CCTV monitoring program, and determine whether a occupant passing through an entrance/exit or not with the designated recognition area as a center.

Here, when the "motion detection area" is set through the CCTV monitoring program, a remote control signal is transmitted to the IP camera for CCTV to detect the movement of occupants in the corresponding area.

Here, in the case of “movement” through the CCTV monitoring program, the IP camera for CCTV stores the detection information in a specific folder of a memory implemented in the IP camera, and periodically polls the specific folder of the IP camera to file related information characterized by bringing

Here, the CCTV monitoring program processes entry counting (People-In Counting) when movement into a predetermined "motion detection area" is confirmed, and when movement out of the "motion detection area" near the boundary is confirmed, exit counting By performing (People-Out Counting) and storing the result, the number of occupants is determined by subtracting the outgoing counting (People-Out Counting) from the entry counting (People-In Counting).

delete

Here, the entrance/exit detection unit may distinguish and recognize occupants based on camera motion detection even when several people move at once.

Here, the entrance/exit detection unit includes: an image sequence extractor for automatically extracting an image sequence by extracting a video frame from each CCTV video stream collected in real time; an image analyzer for analyzing the image extracted by the image sequence extractor; a movement direction detection unit for automatically detecting the movement direction of the occupants according to the movement of the occupants based on the camera motion detection in response to the image analysis of the image analysis unit; and an entry/exit recognition unit that distinguishes and recognizes the entrance or exit of the occupant.

Here, the movement direction detecting unit may store a state in which there is no traffic as an initial state and automatically detect the movement direction when a occupant moves.

On the other hand, as another means for achieving the above-mentioned technical problem, the CCTV image information analysis method for real-time monitoring of occupants according to the present invention is a) Each CCTV installed at the entrance and exit of each floor and area of the building shoots the entrance. to generate a video stream; b) collecting, by the CCTV image information processing device, the video stream generated by each CCTV in real time via a repeater; c) the step of recognizing, by the CCTV image information processing device, the entry/exit of a occupant passing through the entrance/exit separately based on camera motion detection; d) generating, by the CCTV image information processing device, automatically counting the occupants entering or exiting the doorway to generate information on the number of people passing through the doorway; e) transmitting, by the CCTV image information processing device, information on the number of people passing through the entrance to a control monitoring server; and f) the CCTV image information processing device remotely controlling the operation of each CCTV, wherein each of the plurality of CCTVs includes a CMOS image sensor (CIS) and a dynamic vision sensor (DVS). ) or a dome type IP (Internet Protocol) camera with a built-in CMOS image sensor (CIS) and a thermal imaging sensor (TIS); The CCTV image information processing device is implemented with a user interface for processing a CCTV monitoring program and an automatic doorway occupant counting program, and automatically detects the number of entrances and exits according to the occupant's movement line based on camera motion detection for real-time monitoring of the number of occupants; Step c) includes: c-1) extracting each video frame from a video stream collected in real time from a plurality of CCTVs 100; c-2) if the video frame is the first frame, storing a first extracted image assuming no traffic as a basic image; c-3) converting the extracted frame image into a grayscale format to facilitate analysis; c-4) extracting an optical flow feature from the grayscale image; c-5) sequentially performing primary filtering for removing shake of the grayscale image, secondary filtering for removing flicker, and tertiary filtering for removing noise; c-6) performing an RGB comparison operation that is a pixel value of a comparison area on the stored basic image and the grayscale image on which the filtering is sequentially performed; c-7) extracting a current feature list with respect to the image on which the comparison operation of comparing pixel values (RGB) has been performed, and comparing the current feature list with a previous feature list to extract a movement direction; and c-8) recognizing whether each occupant who passes through the entrance or exits enters or exits.

Here, the video stream collected in step c) is processed and analyzed, and the entrance or exit of the occupant is recognized according to the occupant's passage based on camera motion detection; And it is characterized in that the data automatically counting the occupants entering or exiting the entrance obtained as a result of the recognition in step d) is generated.

delete

According to the present invention, when a disaster such as a fire occurs in a high-rise building, CCTV images are collected and visualized to automatically and quickly and accurately analyze the location and number of occupants (or occupants) moving to the exits such as emergency exits and elevators in real time. can

According to the present invention, it is possible to easily distinguish and recognize each occupant even when several people move through the entrance at once according to the occupant's passage based on camera motion detection.

According to the present invention, as each of a plurality of CCTVs is implemented as a dome-type IP camera having a CMOS image sensor (CIS) and a dynamic visual sensor (DVS) or a CMOS image sensor (CIS) and a thermal image sensor (TIS), various illumination environments It is possible to secure a stable detection rate without being affected even under low conditions.

According to the present invention, it is possible to designate a recognition zone according to a CCTV monitoring program, and it is possible to easily determine whether to enter/exit based on the designated recognition zone.

1 is a block diagram of a system for determining presence of occupants according to the prior art.
2 is an operation flowchart for explaining a method of determining the presence or absence of an occupant using the occupant presence determination system according to the related art.
3 is a schematic configuration diagram of a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention.
4 is a diagram for explaining the use of a dome-type CCTV as a CCTV applied to a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention.
5 is a diagram illustrating that CCTV applied to a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention is implemented as CIS-DVS or CIS-TIS.
6 is a diagram schematically illustrating that the CCTV image information processing apparatus uses the CCTV monitoring program and the automatic doorway occupant counting program in the CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention.
7 is a detailed configuration diagram of a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention.
8 is a detailed configuration diagram of the entrance/exit detection unit shown in FIG. 7 .
9 is a schematic operation flowchart of a CCTV image information analysis method for real-time monitoring of occupants according to an embodiment of the present invention.
10 is an operation flowchart specifically illustrating a process of recognizing an occupant's entry/exit in the CCTV image information analysis method for real-time monitoring of occupants according to an embodiment of the present invention.
11 is a diagram illustrating extraction of the number and direction of occupants of an input image in the CCTV image information analysis method for real-time monitoring of occupants according to an embodiment of the present invention.
12 is a diagram illustrating a list of corner features for an input image when analyzing CCTV image information for real-time monitoring of occupants according to an embodiment of the present invention.
13 is a view illustrating counting the number of people entering or leaving by analyzing the moving direction of the occupants in the CCTV image information method for real-time monitoring of the occupants according to an embodiment of the present invention.
14 is a diagram illustrating a program for detection and analysis in which a CCTV image information method for real-time monitoring of occupants according to an embodiment of the present invention is implemented.
FIG. 15 is a diagram specifically illustrating the program for detection and analysis shown in FIG. 14 .

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art can easily implement them. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part "includes" a certain element, it means that other elements may be further included, rather than excluding other elements, unless otherwise stated. In addition, terms such as “…unit” described in the specification mean a unit that processes at least one function or operation, which may be implemented as hardware or software or a combination of hardware and software.

Hereinafter, a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention will be described with reference to FIGS. 3 to 8 , and occupants according to an embodiment of the present invention with reference to FIGS. 9 to 15 . A CCTV image information analysis method for real-time monitoring will be described.

[CCTV video information analysis system for real-time monitoring of occupants]

3 is a schematic configuration diagram of a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention.

Referring to FIG. 3, the CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention is largely, a plurality of CCTVs 100, repeaters 200, CCTV image information processing apparatus 300 and control. and a monitoring server 400 .

A plurality of CCTVs 100 are implemented as IP cameras, and a video stream is generated by photographing an entrance, and a repeater (Hub: 200) transmits the video stream generated by the plurality of CCTVs 100 . Here, the entrance includes an emergency exit and an elevator of a high-rise building.

The CCTV image information processing device 300 collects the video stream passing through the repeater 200, analyzes the entrance video stream, and counts the number of occupants, and the control monitoring server 400 is the plurality of CCTVs ( 100) to perform control monitoring. At this time, the CCTV monitoring program may be implemented as software to control the CCTV of the entire entrance and collect meta data.

On the other hand, FIG. 4 is a view for explaining the use of a dome-type CCTV as a CCTV applied to a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention, and FIG. 5 is a diagram according to an embodiment of the present invention As a diagram illustrating that CCTV applied to CCTV image information analysis system for real-time monitoring of occupants is implemented as CIS-DVS or CIS-TIS, that is, CCTV is a CMOS image sensor (CIS) and a dynamic visual sensor (DVS) or It can be implemented with a CMOS image sensor (CIS) and a thermal image sensor (TIS).

In the case of the CCTV 100 applied to the CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention, as shown in FIG. 4 , CIS and DVS or CIS and TIS are arranged in an array form. Since it is a form of a bundled camera, it is necessary to use a dome-type IP camera with a wide front part.

In addition, in the case of a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention, as shown in FIG. 5 , each of a plurality of CCTVs 100 is a CMOS Image Sensor (CIS)-DVS (Dynamic Vision) Sensor) or CIS-TIS (Thermal Imaging Sensor) is preferably a dome-type IP camera. For example, each of the plurality of CCTVs 100 collects images by operating the CMOS image sensor (CIS) when sufficient illumination is secured, and when the illumination is insufficient, the dynamic visual sensor (DVS) or thermal imager A sensor (TIS) may act instead to create a Video Stream.

Meanwhile, FIG. 6 is a diagram schematically illustrating that the CCTV image information processing apparatus uses a CCTV monitoring program and an automatic door/exit occupant counting program in the CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention.

In the case of a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention, as shown in FIG. 6 , the CCTV image information processing apparatus 300 operates according to a CCTV monitoring program and an automatic entrance and exit occupant counting program. do.

In this case, a CCTV network suitable for a fire situation in a high-rise building can be configured by the CCTV monitoring program, and a control interface for remotely controlling and monitoring an IP camera for CCTV can be built. Accordingly, the CCTV image information processing apparatus 300 may generate a remote control signal, and may collect a video stream from each of the plurality of CCTVs 100 .

In addition, in the case of the CCTV monitoring program, it is designed as a CCTV integrated user interface screen, and a shooting method such as a CCTV resolution can be specified remotely. In addition, it is possible to remotely designate the entrance/exit area of the entrance/exit personnel, for example, if the recognition area is designated in the CCTV monitoring program, it is possible to determine whether to enter/exit based on the designated recognition area. In addition, a database (DB) that can monitor the situation of the person at each entrance by the CCTV monitoring program can be built.

In addition, the automatic occupant counting program at the entrance is a software (Automatic People Counting Program: APCP) that automatically detects the number of people entering each entrance (emergency exit and elevator). It is possible to provide counting data by identifying the etc., and in particular, in the event of a fire, it is possible to process the counting data in consideration of the movement of the occupants gathered at once.

Meanwhile, FIG. 7 is a detailed configuration diagram of a CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention, and FIG. 8 is a detailed configuration diagram of the entrance/exit detection unit shown in FIG. 7 .

7 and 8, the CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention includes a plurality of CCTVs 100, repeaters 200, CCTV image information processing apparatus 300 and Including a control monitoring server 400, the CCTV image information processing device 300 includes a video stream collecting unit 310, an entrance/exit detection unit 320, an occupant counting unit 330, and a data transmission unit 340. and a remote control 350 .

A plurality of CCTVs (100, 100a ~ 100n), as shown in Fig. 5 described above, is installed at the entrance 500 for each floor and each area of the building, and generates a video stream in real time by photographing the entrance 500 do. Here, each of the plurality of CCTVs 100 is preferably a dome-type IP camera having a DVS, which is a CMOS image sensor, built-in, and the entrance 500 includes an emergency exit and an elevator.

The repeater 200 collects and transmits the video streams generated from each of the plurality of CCTVs 100 .

The CCTV image information processing device 300 collects a video stream from the repeater 200 according to a CCTV monitoring program and an automatic doorway occupant counting program, and processes and analyzes the collected video stream to enter or exit the doorway. by counting to generate information on the number of people passing through the entrance according to the counting data, and also remotely control the plurality of CCTVs 100 . In this case, the CCTV image information processing apparatus 300 may designate a recognition area according to the above-described CCTV monitoring program, and determine whether a occupant passing through the entrance/exit is entered with the designated recognition area as a center. For example, if a "motion detection area" is set through the CCTV monitoring program, a remote control signal is transmitted to the IP camera for CCTV 100 to detect the movement of occupants in the area, and also, through the CCTV monitoring program, " When there is movement", the IP camera for CCTV 100 stores the detection information in a specific folder of the memory implemented in the IP camera, periodically polls the specific folder of the IP camera to get the related information file, The CCTV monitoring program processes entry counting (People-In Counting) when movement into a predetermined "motion detection area" is confirmed, and when movement out of the "motion detection area" is confirmed near the boundary, exit counting (People-In Counting) -Out Counting) and storing the result, the number of occupants may be determined as deducting People-Out Counting from People-In Counting.

In addition, the CCTV image information processing apparatus 300 implements a user interface for processing a CCTV monitoring program and an automatic door/exit occupant counting program, and is based on camera motion detection for real-time monitoring of occupants. The number of entrances and exits is automatically detected according to the occupant's flow.

Specifically, the video stream collecting unit 310 of the CCTV image information processing device 300 collects the video streams captured from the plurality of CCTVs 100 via the repeater 200 in real time.

The entrance/exit detection unit 320 of the CCTV image information processing device 300 processes and analyzes the video stream collected from the video stream collection unit 310, and the passage of occupants based on camera motion detection. The occupant's entrance or exit is recognized by dividing it according to the movement line. Also, the entrance/exit detection unit 320 distinguishes and recognizes occupants based on camera motion detection even when several people move at once. More specifically, as shown in FIG. 8 , the entrance/exit detection unit 320 includes an image sequence extraction unit 321 , an image analysis unit 322 , a movement direction detection unit 323 , and an entry/exit recognition unit. (324).

The image sequence extracting unit 321 automatically extracts an image sequence by extracting video frames from each CCTV video stream collected in real time, and the image analysis unit 322 is The extracted image is analyzed in real time according to the CCTV monitoring program.

The movement direction detection unit 323 automatically detects the movement direction of the occupants according to the occupant's movement line based on camera motion detection in response to the image analysis of the image analysis unit 322, and also recognizes entry/exit. The unit 324 distinguishes and recognizes the entrance or exit of the occupant. In this case, the movement direction detecting unit 323 may store the non-traffic state as an initial state and automatically detect the movement direction when the occupant moves.

Referring back to FIG. 7 , the occupant counting unit 330 of the CCTV image information processing device 300 automatically counts the occupants recognized by the entrance/exit detection unit 320 to generate information on the number of people passing through the entrance.

The data transmission unit 340 of the CCTV image information processing device 300 transmits the information on the number of people passing through the entrance and exit counted by the occupant counting unit 330 to the monitoring server 400 .

The remote control unit 350 of the CCTV image information processing device 300 remotely controls the operation method including rebooting, resolution, and frames per second (FPS) designation of each of the plurality of CCTVs 100 .

The control monitoring server 400 controls and monitors the information on the number of people passing through the entrance processed by the CCTV image information processing device 300 . Here, the monitoring server 400 includes: a doorway pass person information database (DB) for storing pass person information processed by the CCTV image information processing device 300; and a control monitor for controlling and monitoring each of the plurality of CCTVs 100 according to the number of passengers information stored in the passenger information database DB.

After all, according to the CCTV image information analysis system for real-time monitoring of occupants according to an embodiment of the present invention, even when several people move through the doorway at once according to the occupant's passage based on camera motion detection, each occupant is It can be easily distinguished and recognized. In addition, as each of a plurality of CCTVs is implemented as a dome-type IP camera with CIS-DVS or CIS-TIS, it is possible to secure a stable detection rate without being affected under various illuminance environments, and also recognize according to the CCTV monitoring program. It is possible to designate a zone, and it is possible to easily determine whether to enter/exit based on this designated recognition zone.

[CCTV video information analysis method for real-time monitoring of occupants]

9 is a schematic operation flowchart of a method for analyzing CCTV image information for real-time monitoring of occupants according to an embodiment of the present invention, and FIG. 10 is a method for analyzing CCTV image information for real-time monitoring of occupants according to an embodiment of the present invention. It is an operation flow chart specifically showing the process of recognizing the entry/exit of a occupant.

9 and 10, the CCTV image information analysis method for real-time monitoring of occupants according to an embodiment of the present invention is, first, each CCTV 100 installed at the entrance 500 for each floor and each area of the building. Creates a video stream (Video Stream) by photographing the entrance (500) (S100).

Next, the CCTV image information processing device 300 collects the video stream generated by each CCTV 100 in real time via a repeater (Hub) (S200).

Next, the CCTV image information processing apparatus 300 distinguishes and recognizes the entry/exit of the occupants passing through the entrance/exit 500 based on the camera motion detection (S300). More specifically, referring to FIG. 10 , each video frame is extracted from a video stream collected in real time from a plurality of CCTVs 100 ( S301 ), and then, it is checked whether it is the first frame ( S302 ). .

Next, when the video frame is the first frame, the first extracted image assuming no passage is stored as a basic image (S303), and then, the extracted frame image is grayscaled to facilitate analysis. (Grayscale) format is converted (S304), and then, the optical flow (Optical Flow) feature is extracted (S305).

Next, primary filtering is performed for removing the shake of the grayscale image (S306), and then secondary filtering is performed for removing flicker of the grayscale image (S307), Next, third-order filtering is performed to remove noise from the grayscale image (S308).

Next, an RGB comparison operation, which is a pixel value of a comparison area, is performed on the stored base image and the grayscale image on which the filtering is sequentially performed (S309).

Next, a current feature list is extracted with respect to the image on which the comparison operation of the comparison region pixel value (RGB) has been performed (S310), and then, the current feature list is compared with the previous feature list to determine the movement direction (S311). ).

Next, it is recognized whether each of the occupants passing through the doorway 500 enters or exits (S312). In this case, the above-described steps S304 to S312 are repeatedly performed for all video frames of the video stream.

Accordingly, the CCTV image information processing device 300 is implemented with a user interface for processing the CCTV monitoring program and the automatic doorway occupant counting program, so that the occupant's real-time monitoring is performed according to the occupant's movement line based on camera motion detection. Personnel can be automatically recognized and detected.

Referring back to FIG. 9 , next, the CCTV image information processing apparatus 300 automatically counts the occupants entering or exiting the entrance 500 to generate information on the number of people passing through the entrance (S400).

Next, the CCTV image information processing device 300 transmits the information on the number of people passing through the entrance to the control monitoring server 400 (S500).

Next, the CCTV image information processing device 300 remotely controls the operation of each CCTV 100 (S600).

On the other hand, in the case of the CCTV image information analysis method for real-time monitoring of occupants according to the embodiment of the present invention, the pre-processing to remove noise and the features for comparing and matching the changes in the input contents are extracted and used as a structure for analysis. applied deep learning-based CNN.

Here, a convolutional neural network (CNN) is the most used algorithm in deep learning, a type of machine learning in which a model directly classifies an image, video, text, or sound. , especially useful for finding patterns to recognize scenes. In particular, these CNNs learn directly from data, classify images using patterns and do not need to manually extract features, and are mainly used in fields that require object recognition and computer vision, for example, face recognition applications. .

Meanwhile, FIG. 11 is a diagram illustrating extracting the number and direction of occupants of an input image in the CCTV image information analysis method for real-time monitoring of occupants according to an embodiment of the present invention, and FIG. It is a diagram illustrating a list of corner features for an input image when analyzing CCTV image information for real-time monitoring of occupants, and FIG. 13 is a moving direction of occupants in the CCTV image information method for real-time monitoring of occupants according to an embodiment of the present invention. It is a diagram illustrating counting the number of people entering or leaving by analyzing.

In the CCTV image information analysis method for real-time monitoring of occupants according to an embodiment of the present invention, a deep learning CNN structure is used, and the first to third steps are largely performed in order.

As the first CNN application process, as shown in FIG. 11, pre-processing and optical flow for extracting input video frame features are applied to obtain a list of corner features, and these features Store the list in a feature list structure. Specifically, elements that are unnecessary or interfere with normal detection are often included in the vector list obtained by optical flow analysis, and thus, they should be removed. For example, when the length of the vector is extremely short, when the distribution in the vector direction protrudes with a large difference from the surrounding, or when the pixel value (RGB) of the area indicated by the motion vector and the pixel value (RGB) of the base image If there is no difference, it is excluded from the vector list.

In this case, the size of the feature list structure is set to be the same as the input video frame image, and all corner feature extraction can be expressed by the optical flow of the input image. For example, the edge feature list may be given as shown in FIG. 12 .

Next, as a second CNN application process, a second feature list structure corresponding to a simplified high-level summary of the first feature list structure, which is an analysis result of the first CNN application process, is generated. Specifically, the generation of the second feature list structure is a process of obtaining and mapping a representative value of a certain region of the first CNN application process, and reduces the noise of the original image and reduces the range of calculation to be analyzed, resulting in processing time Thresholds such as the size of the convolution kernel and the minimum number of items are used in the process of generating the second feature list structure.

Next, as the third CNN application process, the movement information of the occupants can be known from the analysis of the second feature list structure arrangement detected from the image list, and in the case of the same, the consistency of the proceeding direction must be maintained. At this time, if the consistency of the optical flow process is interrupted, it is recognized as the starting point of a new motion, and it is recognized as the appearance of a new occupant. Also, as a condition for the appearance of a new occupant, it is possible to check the consistency of the directional distribution of motion vectors included in the second feature list structure arrangement.

On the other hand, as a function of the CCTV monitoring program, for example, when a "motion detection area" is set, it is transmitted to an IP camera and configured in a manner to detect the movement of occupants in the corresponding area. In addition, if there is a movement, the IP camera stores the detected information in a specific folder of the memory implemented in the camera, and the CCTV monitoring program periodically polls the specific folder of the IP camera to obtain the related information file. use In addition, the CCTV monitoring program processes the entry counting (People-In Counting) as shown in FIG. 13 a) when movement into the predetermined "motion detection area" is confirmed, and "motion detection" in the vicinity of the boundary When it is confirmed that movement out of the "area", as shown in b) of FIG. 13 , people-out counting is performed and the corresponding result is stored.

Accordingly, it is determined that the number of occupants is calculated by subtracting the people-out counting from the people-in counting. That is, it may be given as the number of occupants = (People-In Counting) - (People-Out Counting).

Meanwhile, in the case of the CCTV image information analysis method for real-time monitoring of occupants according to the embodiment of the present invention, as a result of the implementation of the CCTV monitoring program, the user can set the IP camera for CCTV using the management program. In addition, it is possible to adjust the resolution, adjust the number of frames per second (FPS), and specify the storage location of the video. In addition, the screen layout can be designated and used by the administrator according to the situation such as the monitor, for example, screen division such as 3x2, 3x3, 4x2 or 4x3 is possible. In addition, a "motion detection area" can be designated on the user interface (UI), and a plurality of CCTVs 100 in the field can be collectively changed or only a specific CCTV can be changed.

On the other hand, FIG. 14 is a diagram illustrating a program for detection and analysis in which a CCTV image information method for real-time monitoring of occupants according to an embodiment of the present invention is implemented, and FIG. 15 is a detailed example of the program for detection and analysis shown in FIG. It is a drawing showing

The program for detection and analysis in which the CCTV image information method for real-time monitoring of occupants according to an embodiment of the present invention is implemented may be implemented as shown in FIG. 14 .

Specifically, a) of FIG. 15 is a current stored image extracted from a video stream, and the indicated number indicates the numbering of the motion vector in the program. In addition, FIG. 15 b) shows the result of converting the current input image into a grayscale image, which is a step before extracting the optical flow feature from the current input image. Also, FIG. 15 c) shows the grayscale image of the previous stage, and compares it with the current input image to detect the moving direction. In addition, in FIG. 15 d), if there is no passer-by prior to analysis, it is stored as a basic image, and the basic image is used for verification to increase the accuracy of recognition detection. Therefore, as a measurement result, it can be confirmed that the recognition measurement result appears in various ways by specifying the detection area of the entrance and setting various thresholds for the CNN configuration.

After all, according to an embodiment of the present invention, according to the present invention, when a disaster such as a fire occurs in a high-rise building, in order to monitor the location and number of occupants moving to an exit such as an emergency exit or an elevator in real time, CCTV images are automatically collected and visualized. can be analyzed quickly and accurately.

The above description of the present invention is for illustration, and those of ordinary skill in the art to which the present invention pertains can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a dispersed form, and likewise components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention. do.

100: CCTV 200: Repeater (Hub)
300: CCTV image information processing device 400: control monitoring server
500: Entrance (emergency exit and elevator) 310: Video stream collection unit
320: entrance/exit detection unit 330: occupant counting unit
340: data transmitter 350: remote control
321: image sequence extraction unit 322: image analysis unit
323: movement direction detection unit 324: entry / exit recognition unit

Claims (22)

A plurality of CCTVs (100) installed at the entrance 500 for each floor and each area of the building to generate a video stream (Video Stream);
a repeater (Hub: 200) for collecting and transmitting video streams generated from each of the plurality of CCTVs (100);
According to the CCTV monitoring program and automatic doorway occupant counting program, a video stream is collected from the repeater 200, and the collected video stream is processed and analyzed to count the occupants entering or exiting the doorway, and the doorway passage according to the counting data CCTV image information processing device 300 for generating personnel information and remotely controlling the plurality of CCTVs 100; and
Including a control monitoring server 400 to control and monitor the information on the number of people passing through the entrance processed by the CCTV image information processing device 300,
Each of the plurality of CCTVs 100 has a built-in CMOS image sensor (CMOS Image Sensor: CIS) and a Dynamic Vision Sensor (DVS) or a CMOS image sensor (CIS) and a Thermal Imaging Sensor (TIS). Dome Type IP (Internet Protocol) camera;
The CCTV image information processing device 300 is implemented with a user interface for processing a CCTV monitoring program and an automatic doorway occupant counting program, so that occupants can be monitored in real time based on camera motion detection. automatically detect the entrance and exit personnel according to the traffic flow;
The CCTV image information processing apparatus 300 includes: a video stream collecting unit 310 for collecting video streams captured from the plurality of CCTVs 100 via the repeater 200 in real time; A doorway that processes and analyzes a video stream collected from the video stream collecting unit 310, and distinguishes and recognizes the entrance or exit of the occupant 500 according to the occupant's passage based on camera motion detection. Entry/exit detection unit 320; an occupant counting unit 330 for automatically counting the occupants recognized by the entrance/exit detection unit 320 to generate information on the number of people passing through the entrance; a data transmission unit 340 for transmitting the information on the number of people passing through the entrance and exit counted by the occupant counting unit 330 to the monitoring server 400; And CCTV video information analysis system for real-time monitoring of occupants comprising a remote control unit 350 for remotely controlling the operation method including rebooting, resolution, and frame per second (FPS) designation of each of the plurality of CCTVs 100 . delete According to claim 1,
In each of the plurality of CCTVs 100, when sufficient illumination is secured, the CMOS image sensor (CIS) operates to collect images, and when illumination is insufficient, the dynamic visual sensor (DVS) or thermal image sensor (TIS) CCTV image information analysis system for real-time monitoring of occupants, characterized in that it works instead to generate a video stream. According to claim 1,
The CCTV image information processing device 300 designates a recognition zone in the CCTV monitoring program, and determines whether a occupant passing through the entrance or exit is made based on the designated recognition zone, real-time monitoring of the number of occupants CCTV video information analysis system for 5. The method of claim 4,
CCTV image information analysis for real-time monitoring of occupants, characterized in that when "motion detection area" is set through the CCTV monitoring program, a remote control signal is transmitted to the IP camera for CCTV 100 to detect the movement of occupants in the area system. 5. The method of claim 4,
Through the CCTV monitoring program, "when there is movement", the IP camera for CCTV 100 stores the detection information in a specific folder of the memory implemented in the IP camera, and periodically polls the specific folder of the IP camera for related information CCTV image information analysis system for real-time monitoring of occupants, characterized in that the file is imported. 5. The method of claim 4,
The CCTV monitoring program processes entry counting (People-In Counting) when movement into a predetermined "motion detection area" is confirmed, and when movement out of the "motion detection area" is confirmed near the boundary, exit counting (People-In Counting) -Out Counting) and storing the result, the number of occupants is determined by deducting People-Out Counting from People-In Counting CCTV for real-time monitoring of occupants Video information analysis system. delete According to claim 1,
CCTV image information analysis system for real-time monitoring of occupants, characterized in that the entrance/exit detection unit 320 distinguishes and recognizes occupants based on camera motion detection even when several people move at once. According to claim 1, wherein the entrance / exit detection unit 320,
an image sequence extractor 321 for automatically extracting an image sequence by extracting a video frame from each CCTV video stream collected in real time;
an image analyzer 322 for analyzing the image extracted by the image sequence extractor 321;
a movement direction detection unit 323 that automatically detects the movement direction of the occupants according to the occupant's movement line based on camera motion detection in response to the image analysis of the image analysis unit 322; and
CCTV image information analysis system for real-time monitoring of occupants, including an entry/exit recognition unit 324 for distinguishing and recognizing the entrance or exit of the occupant. 11. The method of claim 10,
The moving direction detecting unit 323 stores the no-passage state as an initial state and automatically detects the moving direction when the occupants move. a) generating a video stream by photographing the entrance 500 by each CCTV 100 installed at the entrance 500 for each floor and each zone of the building;
b) the CCTV image information processing device 300 collecting the video stream generated by each CCTV 100 in real time via a repeater (Hub);
c) the step of recognizing, by the CCTV image information processing device 300, the entry/exit of a occupant passing through the entrance/exit 500 based on the camera motion detection;
d) generating, by the CCTV image information processing device 300, automatically counting occupants entering or exiting the doorway 500 to generate information on the number of people passing through the doorway;
e) transmitting the CCTV image information processing device 300 to the control monitoring server 400 the information on the number of people passing through the entrance; and
f) comprising the step of remotely controlling the operation of each CCTV 100 by the CCTV image information processing device 300,
Each of the plurality of CCTVs has a CMOS image sensor (CIS) and a dynamic vision sensor (DVS) or a CMOS image sensor (CIS) and a thermal imaging sensor (TIS) in a dome type ( Dome Type) IP (Internet Protocol) camera;
The CCTV image information processing device 300 is implemented with a user interface for processing a CCTV monitoring program and an automatic doorway occupant counting program, so that the number of occupants is automatically calculated according to the occupant's movement line based on camera motion detection for real-time monitoring of the number of occupants. detect;
The step c) includes: c-1) extracting each video frame from a video stream collected in real time from a plurality of CCTVs 100; c-2) if the video frame is the first frame, storing a first extracted image assuming no traffic as a basic image; c-3) converting the extracted frame image into a grayscale format to facilitate analysis; c-4) extracting an optical flow feature from the grayscale image; c-5) sequentially performing primary filtering for removing shake of the grayscale image, secondary filtering for removing flicker, and tertiary filtering for removing noise; c-6) performing an RGB comparison operation that is a pixel value of a comparison area on the stored basic image and the grayscale image on which the filtering is sequentially performed; c-7) extracting a current feature list with respect to the image on which the comparison operation of comparing pixel values (RGB) has been performed, and comparing the current feature list with a previous feature list to extract a movement direction; And c-8) CCTV image information analysis method for real-time monitoring of occupants comprising the step of recognizing whether each occupant passing through the entrance or exit 500. 13. The method of claim 12,
processing and analyzing the video stream collected in step c), and classifying and recognizing the entrance or exit of the occupant according to the occupant's passage based on camera motion detection; And CCTV image information analysis method for real-time monitoring of occupants, characterized in that generating data automatically counting the occupants entering or leaving the entrance (500) obtained as a result of the recognition in step d). delete 13. The method of claim 12,
In each of the plurality of CCTVs 100, when sufficient illumination is secured, the CMOS image sensor (CIS) operates to collect images, and when illumination is insufficient, the dynamic visual sensor (DVS) or thermal image sensor (TIS) CCTV image information analysis method for real-time monitoring of occupants, characterized in that it works instead to generate a video stream. 13. The method of claim 12,
In step c), the CCTV image information processing device 300 designates a recognition area through the CCTV monitoring program, and determines whether an occupant passing through the entrance/exit is entered with the designated recognition area as the center. CCTV image information analysis method for real-time monitoring of occupants with 13. The method of claim 12,
CCTV image information analysis for real-time monitoring of occupants, characterized in that when "motion detection area" is set through the CCTV monitoring program, a remote control signal is transmitted to the IP camera for CCTV 100 to detect the movement of occupants in the area Way. 13. The method of claim 12,
Through the CCTV monitoring program, "when there is movement", the IP camera for CCTV 100 stores the detection information in a specific folder of the memory implemented in the IP camera, and periodically polls the specific folder of the IP camera for related information CCTV image information analysis method for real-time monitoring of occupants, characterized in that the file is imported. 13. The method of claim 12,
The CCTV monitoring program processes entry counting (People-In Counting) when movement into a predetermined "motion detection area" is confirmed, and when movement out of the "motion detection area" is confirmed near the boundary, exit counting (People-In Counting) -Out Counting) and storing the result, the number of occupants is determined by deducting People-Out Counting from People-In Counting CCTV for real-time monitoring of occupants Image information analysis method. 13. The method of claim 12,
CCTV image information analysis method for real-time monitoring of occupants, characterized in that even when several people move at once in step c), each occupant is recognized separately based on camera motion detection. delete 13. The method of claim 12,
In the step c-2), the state in which there is no passage of the entrance 500 is stored as an initial state and stored as a basic image, and in the step c-7), the direction of movement is automatically detected when the occupants move in real time. CCTV image information analysis method for monitoring.

Images