System and method for object detection and tracking by using edge CCTV
The present invention relates to a system and a method for object detection and tracking by using an edge CCTV, the system and method determining whether an object is identical and tracking same by collectively considering target object characteristics having a relatively small amount of data which is processable according to the specifications of an edge CCTV, wherein, if suspect objects are detected by multiple CCTVs, a control server determines one target object through precise analysis, and tracks the target object while reflecting a changed feature thereof. The present invention may comprise a CCTV node including a camera, a transceiver, and an edge processor, and a control server including a determination module and a tracking module.
1 . A system for detecting and tracking an object using an edge CCTV, the system comprising:
a plurality of CCTV nodes each comprising a camera, a transceiver configured to continuously transmit video captured by the camera to a supervising server and receive characteristic information for detecting a target object from the supervising server, and an edge processor configured to analyze the captured video based on the received characteristic information, report appearance of a suspicious object to the supervising server when the suspicious object is detected, and report departure of the suspicious object to the supervising server when the suspicious object disappears after reporting the appearance of the suspicious object; and
the supervising server comprising a determination module configured to confirm the suspicious object as the target object or determine the suspicious object as a detection error based on video having the suspicious object whose appearance is reported, and a tracking module configured to generate a movement path of the target object based on CCTV nodes each reporting appearance and departure of the target object.
2 . The system according to claim 1 , wherein:
the characteristic information comprises at least three of a top color, a bottom color, a hair length, a height, wearing a hat, wearing a tie, wearing glasses, wearing a mask, or carrying bag, and
when a detected object corresponds to a preset number or more of pieces of the characteristic information, the edge processor of the CCTV node considers the object as a suspicious object.
3 . The system according to claim 2 , wherein:
the supervising server further comprises a propagation module configured to report appearance of the target object to other CCTV nodes around a CCTV node reporting the target object when the determination module confirms the suspicious object as the target object, and
when the detected object corresponds at least three pieces of the characteristic information, and propagation of appearance of the target object in a surrounding CCTV node is received from the supervising server, the edge processor of the CCTV node determines the object as a detection error.
4 . The system according to claim 3 , wherein:
when a suspicious object is confirmed as the target object, the determination module of the supervising server further reports confirmation of the target object to a CCTV node reporting the target object,
upon occurrence of a change in characteristic information among at least one of putting on or taking off glasses, putting on or taking off a hat, changing a color of a top due to taking off clothes, or removing a bag of the target object being tracked, the edge processor of the CCTV node receiving report of confirmation of the target object further reports the changed characteristic information to the determination module of the supervising server, and
the propagation module of the supervising server propagates the reported changed characteristic information to nodes around the CCTV node reporting the change.
5 . The system according to claim 1 , wherein the supervising server further comprises a display module configured to output videos transmitted from the CCTV nodes to a display device divided into a plurality of areas in real time, display all videos in which suspicious objects are reported by switching from a common mode to an attention mode, and then switch videos determined to have objects of detection errors to the common mode again.
6 . The system according to claim 5 , wherein the supervising server further comprises an alarm module configured to compare the movement path of the target object with a digital map, and to notify the display module so that video including the target object is displayed in a warning mode or transmit a warning message to a preset administrator terminal when the target object approaches a preset security zone or dangerous zone.
7 . The system according to claim 5 , wherein:
the CCTV node further provides capturing environment information including an illuminance value of light in a capturing range and a color value of light of surrounding streetlights to surrounding CCTV nodes via the supervising server, and
when the target object departs from the CCTV node capturing the target object, each of the surrounding CCTV nodes detects the suspicious object in corrected video factoring the capturing environment information.
8 . A method of detecting and tracking an object using an edge CCTV, the method comprising:
transmitting, by a supervising server, characteristic information of a target object to a plurality of CCTV nodes;
receiving, by a transceiver of each of the CCTV nodes, characteristic information for detecting the target object from the supervising server, analyzing, by an edge processor of the CCTV node, captured video based on the received characteristic information to report appearance and departure of a suspicious object to the supervising server when the suspicious object is detected;
receiving, by a determination module of the supervising server, a report of the appearance of the object and video in which the suspicious object is detected from the CCTV node, analyzing the video, and confirming the suspicious object as the target object; and
generating, by a tracking module of the supervising server, a movement path of the target object based on CCTV nodes reporting appearance and departure of the target object.
9 . The method according to claim 8 , wherein:
the characteristic information comprises at least three of a top color, a bottom color, a hair length, a height, wearing a hat, wearing a tie, wearing glasses, wearing a mask, or carrying bag, and
when a detected object corresponds to a preset number or more of pieces of the characteristic information, the edge processor of the CCTV node considers the object as a suspicious object.
10 . The method according to claim 9 , wherein:
the supervising server further comprises a propagation module configured to report appearance of the target object to other CCTV nodes around a CCTV node reporting the target object when the determination module confirms the suspicious object as the target object, and
when the detected object corresponds at least three pieces of the characteristic information, and propagation of appearance of the target object in a surrounding CCTV node is received from the supervising server, the edge processor of the CCTV node determines the object as a detection error.
11 . The method according to claim 10 , wherein:
when a suspicious object is confirmed as the target object, the determination module of the supervising server further reports confirmation of the target object to a CCTV node reporting the target object,
upon occurrence of a change in characteristic information among at least one of putting on or taking off glasses, putting on or taking off a hat, changing a color of a top due to taking off clothes, or removing a bag of the target object being tracked, the edge processor of the CCTV node receiving report of confirmation of the target object further reports the changed characteristic information to the determination module of the supervising server, and
the propagation module of the supervising server propagates the reported changed characteristic information to nodes around the CCTV node reporting the change.
12 . The method according to claim 8 , wherein the supervising server further comprises a display module configured to output videos transmitted from the CCTV nodes to a display device divided into a plurality of areas in real time, display all videos in which suspicious objects are reported by switching from a common mode to an attention mode, and then switch videos determined to have objects of detection errors to the common mode again.
13 . The method according to claim 12 , wherein the supervising server further comprises an alarm module configured to compare the movement path of the target object with a digital map, and to notify the display module so that video including the target object is displayed in a warning mode or transmit a warning message to a preset administrator terminal when the target object approaches a preset security zone or dangerous zone.
14 . The method according to claim 12 , wherein:
the CCTV node further provides capturing environment information including an illuminance value of light in a capturing range and a color value of light of surrounding streetlights to surrounding CCTV nodes via the supervising server, and
when the target object departs from the CCTV node capturing the target object, each of the surrounding CCTV nodes detects the suspicious object in corrected video factoring the capturing environment information.