KR20130104068A - Wide area survillance system and method of photographing a moving object - Google Patents

Abstract

PURPOSE: A broad area monitoring system and a moving object photographing method are provided to shorten the movement route prediction time of a moving object, and are provided to optimally photograph the object moving at high speed within the fast time by setting a pan/tilt/zoom (PTZ) value and an auto focus (AF) value of a camera before the moving object comes out to the predicted position. CONSTITUTION: When an event is sensed in a broad area monitoring image inputted from a camera, a controller of a broad area monitoring system predicts the position of an event subject after the time out (S803-S811). The controller determines a PTZ value of the camera. Based on a partial magnification image of the event subject, the controller sets an AF value of the camera, which aims a predicted area and is in standby status, before the event subject comes out to the prediction area (S813). [Reference numerals] (AA) Event subject information; (BB) Circumstance information; (CC) Start; (S801) Monitor a broad area; (S803) Has an event been detected ?; (S805) Take pictures of the event; (S807) Extract the subject of the event; (S809) Analyze the subject's moving patterns; (S811) Predict the future position of the subject; (S813) Set a PTZ/AF camera and wait; (S815) Has the event's subject appeared ?; (S817) Take pictures

Description

Wide area survillance system and method of photographing a moving object}

The present invention relates to a wide area monitoring system and a moving object photographing method, and more particularly, to a moving object photographing method in a wide area monitoring system using a PTZ camera and a wide area monitoring system.

In the conventional wide area monitoring system, when an event generation signal is input while a surveillance camera monitors a wide range, a PTZ and AF (auto focus) are performed on a region or a target where an event occurs and photographed.

As another example of a wide-area surveillance system, an image is taken when the object enters during an event while the surveillance camera waits in advance on a path through which the moving object passes. For example, when an event such as a collision occurs in the parking lot while the camera is waiting at the parking lot exit / entrance, the surveillance camera captures the event vehicle when the event vehicle enters the exit.

Another example of a wide area monitoring system is an example of zooming in while tracking a moving object while monitoring a wide range using a camera having an auto-tracking function.

Korean Patent No. 0879623 discloses a wide area monitoring system for automatically PTZ tracking when a moving object is detected in an area set by a user. As shown in FIG. 10, the wide area monitoring system of the patent sets a detection area and a detection rule of an event to be detected using a PTZ camera image in step S101. In operation S102, the wide area monitoring system analyzes an image signal received from a PTZ camera to detect an event occurring in the surveillance area. In this case, the event detection may be classified into an event detection method (S1021) by preset touring (S1021) and an event detection method (S1022) by by continuous scan, a preset touring method (S1021) and a continuous scan method. The event may be detected according to one of the methods selected by the user in S1022. In step S103, the wide-area monitoring system determines whether an event is detected, and returns to step S102 when an event is not detected, whereas proceeds to step S104 when an event is detected, and detects an event. Information is sent to the control / monitoring terminal to notify the user that an event has been detected. In step S105, the wide area monitoring system performs automatic PTZ tracking of the moving object that generated the event. In step S106, the wide area monitoring system determines whether the automatic PTZ tracking is terminated, and if the automatic PTZ tracking is not terminated, returns to step S105 to repeat the automatic PTZ tracking of the moving object that generated the event. On the other hand, when the automatic PTZ tracking is terminated, the process returns to step S102 to continue event detection for the surveillance area.

The present invention provides a wide area monitoring system and an object photographing method capable of predicting a moving path of an object moving at a high speed and optimally photographing a moving object at a fast time at the predicted position.

According to an embodiment of the present invention, a method of photographing a moving object in a wide area monitoring system includes: estimating a location of an event subject after a predetermined time when an event is detected in a wide range surveillance image input from a camera; Determining a pan / tilt / zoom value of the camera to direct a predicted area corresponding to the predicted position and to perform a predicted area monitoring; And setting an autofocus value of the camera waiting in the prediction area before the event subject appears in the prediction area based on the partial enlarged image in which the event subject is enlarged to the zoom value of the camera. It can include;

The location estimation of the event subject may include extracting the event subject from a wide-area event image in which the event is detected; Analyzing a movement pattern of the event subject based on the event subject information and environment information of the surveillance region; And predicting the location of the event subject based on the movement pattern.

The analyzing of the movement pattern of the event subject may include: setting a movable area based on a current position of the event subject; Determining a moving direction of the event subject based on a line position defining the movable region through which the event subject passes; And calculating a moving speed of the event subject based on the number of pixels passed by the event subject per time.

The setting of an autofocus value of the camera may include: extracting an area including the event subject from a wide-range event image in which the event is detected; And setting an autofocus value of the camera based on an event subject of the partially enlarged image in which the extracted area is enlarged by the zoom magnification of the camera.

When there are a plurality of event subjects, a prediction area monitoring order of each event subject may be set based on the movement speed of the event subjects.

According to an aspect of the present invention, there is provided a wide area monitoring system comprising: a moving object tracking unit predicting a position of an event subject after a predetermined time elapses when an event is detected in a wide area surveillance image input from a camera; And determining a pan / tilt / zoom value of the camera to direct a predicted area corresponding to the predicted position and to perform a predicted area monitoring, and based on a partial magnified image in which the event subject is enlarged to the zoom value of the camera. And a camera controller configured to set an autofocus value of the camera waiting in the prediction area before the event subject appears in the prediction area.

The present invention can shorten the moving path prediction time of the moving object, set the PTZ value and the AF value of the camera before the moving object appears in the predicted position, and thus can optimally photograph the moving object at a high speed.

1 is a view schematically showing a wide area monitoring system according to an embodiment of the present invention.
2 is a diagram illustrating an example of a wide-area event image according to an embodiment of the present invention.
3 is a diagram illustrating an example of extracting an event subject from a wide area event image according to an embodiment of the present invention.
4 is a diagram illustrating an example of setting a movable region for pattern analysis in a wide area event image according to an embodiment of the present invention.
5 is an exemplary view for explaining a method of analyzing a movement pattern of an event subject according to an embodiment of the present invention.
6 is a diagram illustrating an example of setting a prediction region in a wide-area event image according to an embodiment of the present invention.
7 is an exemplary diagram for describing a pre-autofocusing method of a prediction region according to an embodiment of the present invention.
8 is a flowchart schematically illustrating a moving object photographing method using a PTZ camera in a wide area monitoring system according to an embodiment of the present invention.
9 is an exemplary view of an image photographed by a conventional wide area monitoring system.
10 is a flowchart illustrating a monitoring method of a conventional wide area monitoring system.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

In the case of a conventional surveillance system that captures an object by controlling a PTZ (Pan / Tilt / Zoom) value of a camera when an event occurs during wide-area surveillance, as shown in FIG. If the object moves while the object may not be captured properly. In order to prevent this, when photographing an object in a wide area, obtaining clear information about a moving object is not easy.

As another conventional surveillance system, when a camera enters a waiting object before the moving object must pass when an event occurs, it can be used only where the moving object passes in advance, such as a parking lot or a building. There is a problem that the utilization range is narrow. In addition, it is impossible to respond when an exception occurs, such as when a moving object passes through an undefined location. In addition, it is essential to check whether the vehicle is an event vehicle for all vehicles passing from the event until the accident vehicle exits. Therefore, a module that can determine whether or not it is an event object is needed.

As another conventional monitoring system, a system using an auto-tracking function of a camera, such as Patent No. 0879623, has a long processing time for tracking, making it difficult to cope with an object moving at a high speed. However, it is possible for an escaped criminal to miss an object's movement faster than tracking.

The present invention predicts a future position of a moving object from an image of which an event is detected, and presets a PTZ value and an autofocus (AF) value of a camera before the moving object appears at the predicted position. The camera checks and photographs the moving object in the air at the predicted position with the PTZ value and the AF value set. Accordingly, the present invention enables accurate shooting even of a fast moving object, which is advantageous for shooting a moving object at high speed.

1 is a view schematically showing a wide area monitoring system according to an embodiment of the present invention.

Referring to FIG. 1, the wide area monitoring system 1 according to an embodiment of the present invention includes a camera 10, a controller 20, and a display unit 90.

The wide area monitoring system 1 of the present invention predicts a moving path of an object, and photographs an object quickly and accurately when the moving object is found while the camera 10 is waiting at a predicted position where the object will appear after a predetermined time. The wide area surveillance system 1 monitors the widest possible range with the image size required for detection to increase the efficiency of the surveillance system until an event occurs. This can minimize the number of cameras 10 installed in the surveillance area.

The camera 10 may be a PTZ camera which is disposed at a fixed position at a specific place and has a pan / tilt / zoom function. The camera 10 may receive various images while moving and / or panning and / or tilting rotation by the driving device, or adjusting the zoom magnification by zooming in or zooming out the zoom lens. The camera 10 may detect a moving object by dividing a specific area into a plurality of monitoring areas and monitoring each monitoring area by a preset method. At this time, in the preset method, each surveillance area is monitored while moving or rotating according to a preset preset. The camera 10 may be connected to the controller 20 by wire or wirelessly to transmit and receive images and control data.

The controller 20 detects an event from the image input from the camera 10 to track the moving object and controls the camera 10 to photograph the moving object. The controller 20 may include a moving object tracking unit 30 and a camera controller 80.

The moving object tracking unit 30 predicts a future position of the event subject after a predetermined time when an event is detected from an input image captured in a wide area (hereinafter, referred to as a 'wide surveillance image'). The moving object tracking unit 30 may include an event detector 40, a subject extractor 50, a pattern analyzer 60, a database 65, and a position predictor 70.

The event detector 40 checks whether an event has occurred from the wide area surveillance image. The event type can be defined differently depending on the installation place of the monitoring system. For example, an event such as intrusion, theft, accident or fire by a person or a car can be set. The event detector 40 determines whether an event occurs according to the detailed rule set according to the event type. The event detector 40 may borrow and use various types of events previously defined. The event detection method is not particularly limited and various conventional methods may be used.

When an event is detected, the camera 10 stops moving or rotating, and photographs the wide area monitoring region where the event is detected at the current preset position. The captured image of the wide-area surveillance region where an event is detected (hereinafter, referred to as a 'wide-event image') may be stored in a separate storage unit and used as evidence for future event occurrence. 2 is a diagram illustrating an example of a wide-area event image according to an embodiment of the present invention. 2 is a wide-area event image in which a human invasion event is detected in the wide-area surveillance image.

The subject extractor 50 identifies and extracts an event subject from the wide-area event image. The event subject may be a person, a car, or the like. The identification method of the event subject is not particularly limited, and various conventional methods may be used.

3 is a diagram illustrating an example of extracting an event subject from a wide area event image according to an embodiment of the present invention. In FIG. 3, the extracted event subject is represented by a box (first box, OB) like a rectangle in the wide area event image.

The database 65 includes prior information for predicting the future location of the event subject. By using the prior information, it is possible to minimize the errors that can occur in the future position prediction and to improve the processing speed. The dictionary information may include event subject information and environment information. The event subject information may include a movement type, a movement speed, and the like of the event subject as attributes for each event subject. For example, in the case of a car, two-dimensional (2D) movement, that is, three movements of forward / reverse / rotation are possible. People can move freely, but the maximum speed a person can move is limited compared to a car. The environmental information is spatial information of the surveillance area, and is information on an area in which an event subject cannot move (move) among the surveillance areas to be photographed, such as a wall or a building. It is possible to reduce the range to be considered when predicting future location based on event subject information and environment information.

The pattern analyzer 60 calculates the position, movement speed, and movement direction of the event subject based on the dictionary information stored in the database 65. Accordingly, the pattern analyzer 60 may analyze the movement pattern of the event subject. The pattern analyzer 60 sets the detected event subject as the center position and sets an area that can be moved around. The pattern analyzer 60 sets a movable area based on the event subject information and the environment information from the database 65. The movable area may be automatically set according to a preset rule or manually set according to an operator's input.

4 is a diagram illustrating an example of setting a movable region for pattern analysis in a wide area event image according to an embodiment of the present invention. Referring to FIG. 4, the pattern analyzer 60 finds a non-moveable area (black processing area of FIG. 4) based on the environment information and restricts the moving area, and defines the movable area as an event subject, that is, the first box OB. ) May be set to a check box-shaped box (second box, MB) of a size larger than the first box (OB). The second box MB may be composed of a plurality of lines defining a movable area.

The pattern analyzer 60 determines a moving direction of the event subject from the information regarding which line of the second box the event subject passes in the wide area event image for a predetermined time, and the event subject determines the line of the second box. The moving speed of the event subject can be calculated from the passing time information. Here, the movement speed of the event subject may be calculated based on the number of pixels moved per unit time.

5 is an exemplary view for explaining a method of analyzing a movement pattern of an event subject according to an embodiment of the present invention. Referring to FIG. 5, the pattern analyzer 60 may include a line of a plurality of lines constituting a second box MB, which is a movable area in which a first box OB defining an event subject is set around the event subject. Recognizing that the right side of (a) and the lower side of the line (b) is passed, it can be determined that the event subject is moving in the direction of the arrow. The pattern analyzer 60 may calculate a moving speed of the event subject based on the number of pixels passing by the event subject, that is, the first box OB moves for a predetermined time.

The pattern analyzer 60 is not limited to the above-described method, and may apply various conventional methods to analyze the movement pattern of the event subject by calculating the current position, the moving speed, and the moving direction of the event subject. For example, the pattern analyzer 60 may apply a moving object detection algorithm every frame to analyze the moving pattern of the object by the detected position tracking. As another example, the pattern analyzer 60 sets an area that can be moved around the current object position, and then determines whether or not the object enters the set area by comparing the characteristics of the object to determine the optical flow. ), The movement pattern of the object can be analyzed by an object tracking method using a particle filter.

The position predictor 70 predicts a future position based on the movement pattern of the event subject. The position predictor 70 calculates a position (hereinafter, referred to as a 'prediction area') in which the event subject moves after a predetermined time, based on the current position, the moving speed, and the moving direction of the event subject calculated by the pattern analyzer 60. do.

6 is a diagram illustrating an example of setting a prediction region in a wide-area event image according to an embodiment of the present invention. In FIG. 6, the prediction area of the wide area event image is indicated by a box (third box, FB).

The camera controller 80 adjusts the photographing range of the camera 10 based on the prediction area information. The camera controller 80 determines a PT (Pan / Tilt) value of the camera 10 to direct the prediction area, and determines a zoom value (zoom magnification) for switching from wide area monitoring to prediction area monitoring. The camera controller 80 sets an AF value of the camera 10 that is waiting and predicts the prediction area based on the partial enlarged image in which the event subject is enlarged by the zoom value of the camera 10. That is, since the camera 10 assumes the presence of the event subject in the prediction region and waits after performing AF before the event subject appears in the prediction region, when the event subject appears in the prediction region, the camera 10 may optimally photograph the event subject in a short time. Can be.

The AF value is set based on the background before the event subject reaches the prediction area. In this case, when the actual event subject enters the prediction area, resetting the AF value based on the event subject is inefficient and impossible to capture a high-speed object. According to the embodiment of the present invention, the AF value is set in advance before the event subject appears in the prediction area, and the camera 10 can wait in the prediction area. Accordingly, when the event subject appears in the prediction region, the autofocusing operation can be omitted or minimized, so that optimal shooting of the moving object can be performed at a high speed.

The camera controller 80 extracts a predetermined region including the event subject from the wide-area event image, enlarges the extracted event subject by increasing the resolution to a zoom level set for monitoring the prediction region, and then when an object exists in the background. To focus.

7 is an exemplary diagram for describing a pre-autofocusing method of a prediction region according to an embodiment of the present invention. Referring to FIG. 7, for pre-autofocusing, the camera controller 80 extracts a predetermined region including an event subject from a wide area event image. The camera controller 80 may extract an area having a size including at least the first box OB. The camera controller 80 may extract a region including the event subject by using the image used in the movement pattern analysis step. The camera controller 80 enlarges the extracted region by a zoom value (zoom magnification) set for monitoring the prediction region and corrects the enlarged event subject image. Here, the zoom value may be determined by the number of pixels of the entire image with respect to the number of pixels of the extracted region. The image correction method is not particularly limited in the exemplary embodiment of the present invention, and an interpolation method of increasing sharpness compared to neighboring pixels and a super resolution technique of increasing resolution by fusing a plurality of low-resolution images and The same conventional various correction techniques can be applied. The camera controller 80 sets the AF value of the camera 10 based on the enlarged and corrected event subject image.

The camera 10 moves to the prediction region according to the pan / tilt value, zooms in the prediction region according to the zoom value, performs autofocusing on the prediction region zoomed in according to the AF value, and performs the event subject as the prediction region. Wait until appears. When the event subject appears in the prediction area within a predetermined time, the camera 10 photographs the event subject after correcting the AF value. In this case, the camera 10 assumes that an object exists in the background, and since the AF value is set in advance, the correction time of the AF value may be shortened when an event subject appears in the prediction area.

The display unit 90 displays a wide area surveillance image photographed by a preset method, a wide area event image, and an area surveillance image of a predicted region zoomed in by predicting a future position of an event subject. The first box OB for tracking the event subject may be displayed and moved together with the movement of the event subject in the wide area event image and the area surveillance image.

The display unit 90 may provide visual information and / or audio information to an operator, and may include a liquid crystal display panel (LCD), an organic light emitting display panel (OLED), an electrophoretic display panel (EPD), and the like. The display unit 90 may be provided in the form of a touch screen to receive an input through the operator's touch.

8 is a flowchart schematically illustrating a moving object photographing method using a PTZ camera in the wide area monitoring system 1 according to the exemplary embodiment of the present invention. In the following, detailed description of the contents overlapping with FIGS. 1 to 7 will be omitted.

Referring to FIG. 8, one or more cameras 10 of the wide area monitoring system 1 monitor the wide area while moving or rotating the surveillance area in a fixed or preset manner (S801).

The control unit 20 of the wide area monitoring system 1 determines whether an event is detected from the wide area monitoring image (S803). If an event is not detected, the controller 20 continuously monitors the surveillance area in a wide area while moving or rotating the surveillance area in a fixed or preset manner.

When an event is detected, the camera 10 stops moving or rotating, and photographs the surveillance region where the event is detected (S805).

The controller 20 extracts an event subject from the wide area event image (S807).

The controller 20 analyzes the movement pattern of the event subject based on the event subject information and the environment information of the monitoring area (S809). The controller 20 sets the movable area based on the current position of the event subject. The controller 20 determines a moving direction of the event subject based on a line through which the event subject passes among a plurality of lines defining the movable area. The controller 20 calculates a moving speed of the event subject based on the number of pixels passed by the event subject per time.

The controller 20 predicts a future position of the event subject based on the analyzed movement pattern (S811). The controller 20 predicts a position at which the event subject will appear after a predetermined time elapses from the current position, the moving direction, and the moving speed of the event subject.

The controller 20 sets the PTZ value and the AF value of the camera 10 so that the camera 10 faces the prediction area corresponding to the predicted position and performs the prediction area monitoring (S813). The camera 10 moves to the prediction area by performing PTZ according to the PTZ value, while waiting for the prediction area with the preset AF value while waiting in the prediction area.

In order to set the pre-AF value of the camera 10, the controller 20 extracts an area including the event subject from the wide area event image. The controller 20 sets the AF value of the camera 10 based on the event subject of the partially enlarged image in which the extracted area is enlarged by the zoom value of the camera 10. The camera 10 faces the prediction area with the PTZ value and the AF value set, and waits for photographing.

When the event subject appears in the predicted region within a predetermined time in the air (S815), the camera 10 captures the event subject (S817). When the event subject does not appear in the prediction region within a predetermined time or when another object appears (S808), the controller 20 determines that the prediction has failed and predicts the future position (S811) or the camera 10 again. Can again monitor the same surveillance zone or the surveillance zone for the next preset position.

Although the above-described embodiment of the present invention has been described as an event subject as an example, the present invention is not limited thereto, and the present invention may be applied to a case where a plurality of event subjects or a plurality of event subjects occur simultaneously.

When a plurality of event subjects are extracted, the control unit 20 of the wide area monitoring system 1 predicts the movement pattern analysis and the future position for each event subject. At this time, the controller 20 may set the monitoring order for the prediction region of each event subject based on the movement speed. For example, if two event subjects having different movement speeds are detected, the controller 20 predicts a movement pattern analysis and a future position for each event subject. Then, the controller 20 first sets the PTZ value and the AF value for the event subject with a faster moving speed, the camera 10 photographs the image after waiting in the corresponding prediction area, and when the photographing is completed, the controller 20 The PTZ value and the AF value are set for the event subject having a slow moving speed, and the camera 10 captures the image after waiting in the corresponding prediction area. If the movement speeds of the two event subjects are similar, the controller 20 may set an area for monitoring and capturing both event subjects simultaneously as a common prediction region.

As another example, the wide area surveillance system 1 may be linked with other surrounding surveillance systems, and the predicted region and PTZ / AF value of one of the two detected event subjects may be transferred to the other surveillance system to monitor and predict the predicted region. You can shoot.

While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

Claims (10)

Predicting a location of an event subject after a predetermined time elapses when an event is detected from a wide range surveillance image input from a camera;
Determining a pan / tilt / zoom value of the camera to direct a predicted area corresponding to the predicted position and to perform a predicted area monitoring; And
Setting an autofocus value of the camera waiting in the prediction area before the event subject appears in the prediction area based on the partial magnified image in which the event subject is enlarged to the zoom value of the camera; Method of photographing a moving object in a wide area surveillance system comprising a. The method of claim 1, wherein the location estimation of the event subject comprises:
Extracting the event subject from the global event image in which the event is detected;
Analyzing a movement pattern of the event subject based on the event subject information and environment information of the surveillance region; And
Predicting the location of the event subject based on the movement pattern. The method of claim 2, wherein the analyzing of the movement pattern of the event subject comprises:
Setting a movable area based on a current position of the event subject;
Determining a moving direction of the event subject based on a line position defining the movable region through which the event subject passes; And
And calculating a moving speed of the event subject based on the number of pixels passed by the event subject per hour. The method of claim 1, wherein the setting of the autofocus value of the camera comprises:
Extracting an area including the event subject from the global event image in which the event is detected; And
And setting an autofocus value of the camera based on an event subject of the partial enlarged image in which the extracted area is enlarged by the zoom magnification of the camera. The method of claim 1,
And a plurality of the event subjects, the moving object photographing method in the wide area monitoring system in which the predicted area monitoring order of each event subject is set based on the movement speed of the event subjects. A moving object tracking unit predicting a position of an event subject after a predetermined time elapses when an event is detected from a wide-area surveillance image input from a camera; And
Determining a pan / tilt / zoom value of the camera to direct a predicted area corresponding to the predicted position and to perform a predicted area monitoring, and based on a partial magnified image in which the event subject is enlarged to the zoom value of the camera, And a camera controller configured to set an autofocus value of the camera waiting in the prediction area before the event subject appears in the prediction area. The method of claim 6, wherein the moving object tracking unit,
A subject extracting unit which extracts the event subject from the wide-range event image in which the event is detected;
A pattern analyzer configured to analyze a movement pattern of the event subject based on the event subject information and environment information of a surveillance region; And
And a position estimator for predicting the position of the event subject based on the movement pattern. The method of claim 7, wherein the pattern analysis unit,
Set a movable region based on the current position of the event subject, determine the movement direction of the event subject based on the line position defining the movable region through which the event subject passes, the event subject per hour A wide-area monitoring system that calculates a moving speed of the event subject based on the number of pixels that have passed. The method of claim 6, wherein the camera control unit,
An area including the event subject is extracted from the wide area event image in which the event is detected, and an autofocus value of the camera is determined based on the event subject of the partial enlarged image in which the extracted area is enlarged by the zoom magnification of the camera. Wide area monitoring system to set up. The method according to claim 6,
And a plurality of event subjects, wherein a predicted area monitoring order of each event subject is set based on the movement speed of the event subjects.

Images