KR101840042B1 - Multi-Imaginary Fence Line Setting Method and Trespassing Sensing System - Google Patents

Abstract

The present invention relates to a method of setting a multi-imaginary fence line. The method includes: a step of receiving and storing an image taken through a camera; a step of generating fence line 2 by connecting P1 and P2, which are two points of an area to be monitored in the received image, through a mouse or touch point, and then, generating imaginary fence line 1 and imaginary fence line 3 at a predetermined distance from both sides of the fence line 2; a step of setting a crossing direction to the set imaginary fence lines; a step of detecting and tracking a moving tracking detection object from the image taken and transmitted through the camera; a step of selecting a tracking detection object, which is a trespassing candidate, by filtering the moving tracking detection object; a step of determining whether tracking pass information of the tracking detection object is to pass through the imaginary fence lines; a step of finally determining whether trespassing has occurred by making the tracking detection object check a crossing sequence of a multi-imaginary fence line if the passage through the imaginary fence lines is determined; and a step of providing alarm information to a user in the case of trespassing.

Description

TECHNICAL FIELD [0001] The present invention relates to a multi-imaginary fence line setting method and a trespassing sensing system,

The present invention relates to a method for setting a plurality of hypothetical lines in a surveillance area and judging whether or not an intruder is intruding in consideration of passage and direction of a plurality of lines of the sensed object, thereby reducing false positives. Generally, in the field of intelligent CCTV system, object area (people, automobile) detected and tracked from the input image is regarded as an intrusion when the user passes a predetermined line, and information is provided to the customer. However, the conventional method as described above causes inconvenience to customers because of frequent misreporting.

The prior art related to the present invention is disclosed in Korean Patent Laid-Open No. 10-2016-0037180 (published on April 6, 2016). FIG. 1 is a flowchart illustrating a method of setting a region of interest for the conventional intelligent image analysis. In FIG. 1, a conventional method of setting a region of interest for intelligent image analysis is performed by an image analysis apparatus (S100). At this time, the analysis target image may be received from the image capturing apparatus or may be selected from the image transmitted from the image capturing apparatus. When the analysis target image is input, the image analysis device sets a boundary line for detecting an event in the analysis target image according to a user input signal (S110). The boundary line may be set at a designated position of the image to be analyzed, and may be represented by a graphic shape in which one or more line segments, one or more curves, a combination of one or more line segments and curves, and one or more line segments or curves are connected. In addition, the image analysis apparatus sets the size of the object to be detected or tracked before setting the region of interest (S120), and sets a distance value d that defines the range of the region of interest based on the size of the object (S130). The size of the object may be set according to an input signal of a user (manager), and the distance value d may be set to X times the size (a large value of the width or the length) of the object. Here, X may have a range of 2 to 5, for example. In step S120, the image analysis apparatus sets an area within a certain distance as an area of interest based on one or more line segments or curves included in the boundary line among the analysis target images. Here, the predetermined distance based on the one or more line segments or curves corresponds to the distance value set in step S130. The step S120 may be implemented by various methods using a graphic form or a distance map.

The conventional method of setting a region of interest for intelligent image analysis has a problem in that when noises due to changes in environment, climate, or illumination change a line set by a user, a false intrusion situation is detected and an alarm is provided. SUMMARY OF THE INVENTION It is, therefore, an object of the present invention to provide a method for intrusion detection that is robust against noise in consideration of a plurality of virtual line settings and a moving direction of a tracking sensing object.

A method for setting a composite virtual fringe line according to the present invention includes the steps of: receiving and storing an image captured by a camera; detecting a point P 1 and a point P 2 of a region to be monitored in the received image, Generating virtual pans line 1 and virtual pans line 3 at a predetermined distance from both sides of the pans line 2 to form a virtual pans line 3; Detecting and tracking a movement tracking sensing object in an image captured and transmitted by a camera, selecting a tracking sensing object as an intrusion candidate using movement tracking sensing object filtering, Determining whether the tracking path information of the sensing object passes through a virtual fancy line, If it has passed through the comprising the steps of, if the user of the intrusion alarm provides information to the final determination whether an intrusion detection by tracking the object is checked for crossing the order of the compound virtual paenseu line.

The method of setting the composite virtual fancy line of the present invention and the intrusion detection system using the method as described above have the effect of reducing the amount of noise and reducing the amount of false information. In addition, another effect of the present invention is to determine whether or not an intrusion is detected by grasping the moving direction of the tracking sensing object. In addition, the present invention has an effect of delivering an alarm message to a VMS, a DVR, an NVR, and a mobile phone when the tracking detection object is determined as an intrusion object, and stores an image and a time stamp during an event section of the intrusion object, The intrusion event section can be easily retrieved from the security image.

FIG. 1 is a flowchart of a conventional method of setting a region of interest for intelligent image analysis,
FIG. 2 is an explanatory diagram applied to the virtual punch line setting method of the present invention,
FIG. 3 is an explanatory diagram applied to the method of setting a composite virtual fancy line of the present invention,
FIG. 4 is an explanatory diagram of object filter parameter setting applied to the method of setting a composite virtual fancy line of the present invention;
FIG. 5 is an explanatory diagram for a candidate selection of a tracking sensing object by a linear perspective size filter using an imaginary fancy line;
FIG. 6 is a sensitivity explanatory diagram according to a similarity degree of a virtual fancy line direction and a movement direction of a tracking sensing object,
FIG. 7 is a first explanatory diagram for calculating the presence or absence of a virtual fas line cross in path information of a track detection object applied to the present invention;
FIG. 8 is a second explanatory diagram for calculating the presence / absence of a virtual fas line cross in path information of a track detection object applied to the present invention;
FIG. 9 is an explanatory diagram of a sequence in which a tracking sense object applied to the present invention crosses a virtual fancy line.
FIG. 10 is a flowchart illustrating an intrusion detection method control using a complex virtual fuse line setting of the intrusion monitoring area of the present invention.
11 is a block diagram of an intrusion detection system using a method of setting a composite virtual fancy line of the present invention.

The method for setting a composite virtual fancy line of the present invention and the intrusion detection system using the same will now be described with reference to FIGS. 2 to 11. FIG.

FIG. 2 is an explanatory diagram applied to the virtual punch line setting method of the present invention. 2, (a) shows a down / right direction mode, (b) shows a up / left direction mode, and (c) shows a bidirectional mode. In the virtual pans line setting method of the present invention, (1) and line (3) so as to be spaced a certain distance from the line (2). The generation of the line 2 can be generated using a mouse or a touch point and the distance between the line 2 and the line 1 and the distance between the line 2 and the line 3 can be selected by the user , And if the distance (distance) is 0, it can be regarded that there is no line (1) and line (3). In addition, a virtual fas line crossing direction for determining whether or not the tracking sensing object is intruded is automatically generated and generated and stored in a down / right direction mode, a left / right direction mode, and a bidirectional mode, A counterclockwise orthogonal vector of the vector can be obtained in the down / right direction mode, and a clockwise orthogonal vector can be obtained in the up / left direction mode. In addition, the above-described crossing direction mode can be formed such that the user has a certain angle in addition to the orthogonal direction, and the crossing direction can be set to be the same for the line (1) and line (3) It is. In addition, the user can set and store a virtual fas line crossing sequence in order to detect an event on the composite virtual fas line. In the down / right direction mode, the crossing order is the line 1 → line 2 → line 3 Directional mode is to cross in the order of line 3 → line 2 → line 1 and the bidirectional mode is one of the down / right direction mode and the up / left direction mode If they match, it can be determined that they can be crossed in both directions. Although the setting for the crossing direction has been described, it is possible to set the crossing lines independently for the line 1, the line 2 and the line 3 without the crossing order, and for the lines 1 and 3, If the distance to line 2 is zero, line 1 and line 3 need not be considered for the crossing direction.

3 is an explanatory diagram applied to the method for setting a composite virtual fancy line of the present invention. 3, the user can select a plurality of points of the surveillance area received from the camera and select and store the composite virtual fancy line. In FIG. 3 (a), points P1 and P2 Can be set to the virtual pans line 1, and the line between point P 2 and point P 3 indicates that the virtual pans line 2 can store and store the composite virtual pans line, and (b) The virtual fancy line automatic generation and the setting of the crossing direction can be performed as described in FIG. 2 by indicating that the fancy line can be set.

4 is an explanatory diagram of object filter parameter setting applied to the method of setting a composite virtual fancy line of the present invention. In FIG. 4, the object filter parameter setting applied to the method of setting the composite virtual fancy line of the present invention can set a linear perspective size filter of the intrusion detection object using the set virtual fancy line. As shown in FIG. 4, when the object filter is set, two pairs of maximum and minimum filters having arbitrary sizes are set and stored by using the starting point (P1) and the ending point (P2) of the virtual fancy line as a base point will be. In addition, the user can set the position and size of the minimum / maximum size filter on the virtual pans line, and when the object size filtering is applied, the relationship between the size ratio of the two pairs of maximum / minimum filters and the virtual pans line of the lines of P1 and P2 And a linear perspective size filter according to the position of the intrusion detection object tracked can be applied. For example, the Linear perspective size filter of the intrusion detection object can be set using a predetermined virtual fancy line. In FIG. 4, the start point and the end point of the virtual fancy line are set as a foot-point, (Where the user can change the position and size of the max / min size filter on the virtual fancy line), and when applying object size filtering, the ratio of the sizes of the two pairs of max / min filters A linear fuzzy line relation between P1 and P2 line is derived and a linear perspective size filter according to the position of the intrusion detection object tracked can be applied.

FIG. 5 is an explanatory diagram for selecting candidates of a tracking sensing object by a linear perspective size filter using an imaginary fancy line. FIG. In FIG. 5, the candidate selection of the tracking sensing object by the linear perspective size filter using the virtual fancy line is performed by using a y-axis based perspective filter as an example. First, the y-axis position (pixel position) The size change according to the following equation is calculated as 1,

……[아래 1]

... ... [1 below]

Here, Max_W_Ratio is the width ratio of the maximum size filter, Max2 _width is The width (value) of the front full size filter, Max1 _width is P2 _y is the y-axis value of the front reference point P2, P1 _y is the y-axis value of the rear reference point P1, Max_H_Ratio is the height ratio of the maximum size filter, Max2 _height is The height (value) of the front maximum size filter, Max1 _height is The height (value) of the rear maximum size filter, Min_W_Ratio is the width ratio of the minimum size filter, and Min2 _width is the _width The width (value) of the front minimum size filter, Min1 _width is The width (value) of the rear minimum size filter, Min_H_Ratio is the height ratio of the minimum size filter, and Min2 _height is The height (value) of the front minimum size filter, Min1 _height The height (value) of the rear minimum size filter.

Second, the perspective size of the tracking sensing object is estimated as below 2 using the position of the tracking object (TrkObj) position (x, y) size filter,

……[아래 2]

... ... [Below 2]

Here, Pos_Gap is the distance between the tracking sensing object on the y axis and the front reference point P2, TrkObj is the tracking sensing object, Max_TrkObjwidth is the maximum _width of the tracking sensing object, TrkObj _width is Max_TrkObj _height is the maximum _height of the tracking detection object, TrkObj _height is the _height of the tracking detection object, Min_TrkObj _width is the minimum _width of the tracking detection object, and Min_TrkObj _height is the minimum height of the tracking detection object.

Third, based on the perspective magnitude value of the tracking object, object size filtering is performed as shown in 3 below to detect the tracking object candidate, which is an event candidate, to obtain the tracking object.

……[아래 3]

... ... [3 below]

Here, Max2 _width is the _width of the maximum size filter, Max2 _height is the _height of the maximum size filter, Min2 _width is the _width of the minimum size filter, and Min2 _height is the _height of the minimum size filter. In the above, the linear perspective filter is an example filter for candidate selection of the tracking sensing object and can select the tracking sensing object candidate using Calibration or other relational expression.

FIG. 6 is an explanatory diagram of sensitivity according to the direction of the virtual fancy line and the movement direction of the tracking sensing object according to the present invention. In FIG. 6, the degree of similarity between the direction of the virtual fence crossing and the direction of movement of the tracking sensing object may be set and stored as sensitivity. In FIG. 6, when the dot product of the virtual fringe line direction vector and the movement tracking sensing object direction vector is obtained, the degree of coincidence of directions can be obtained and the reference value for the degree of coincidence can be set to the sensitivity. For example, when the angle between the virtual fuselage direction vector and the movement tracking sensing object orientation vector is 15 degrees or more, the reference value is a direction in which the movement direction of the tracking sensing object is different from the direction of the virtual fancy line, It can be judged by crossing. In addition, when the tracking detection object detection and tracking model is based on the background learning model, the sensitivity can be set based on the similarity between the input image of the tracking sensing object area received from the camera and the preset background image.

FIG. 7 is a first explanatory diagram for calculating whether or not path information of a tracking sensing object applied to the present invention is a virtual fas line crossing. In FIG. 7, in order to determine whether there is a virtual fas line crossing of the tracked sensing object, it is necessary to select an intrusion detection candidate, and a tracked sensing object region and tracking path information for intrusion detection are acquired through a movement sensing object detection and tracking model An example of the movement tracking detection object detection and tracking model may be based on a background learning model, and in this case, a foreground classification image and a background restoration image may be obtained and provided. Also, there is a method of selecting invasion candidates based on the object size filter and a method of selecting invasion candidates based on the object appearance filter. The method of selecting an intrusion candidate according to the object size filter may be such that if the size of the detected object being tracked does not fall within the range of the set maximum / minimum size filter, the tracking object can be set to be excluded from the intrusion candidate. If the size filter mode set on the panning line is a fixed size filter, it is possible to simply select the intrusion candidate according to the degree of similarity by comparing the size of the trace area with the width / width or width of each size filter, In the case of applying the maximum size / minimum size filters set by the user, the information of the physically restored information in the 3D coordinate system and the area of the detected object being tracked are stored in the 3D coordinate system, It is possible to select intrusion candidates according to similarity by comparing vertical values If the size filter mode set on the virtual pans line is a linear perspective size filter, the user assigns the relationship of the maximum / minimum size filter set on the virtual pans line to the size of the tracking sense object The intrusion candidate can be selected according to the degree of similarity. In addition, the intrusion candidate selection method using the object outline filter can classify the appearance of the tracking detection object in advance and can classify people, vehicles, noise, etc., and obtain a reference value (score value) The selected intrusion detection object is compared with the threshold value registered and stored, and if the selected intrusion detection candidate object is different from the threshold value, it can be excluded from the intrusion candidate. For example, if the tracking detection object model is based on the background learning model, the feature difference between the background image and the input image for the tracking area is calculated. If the calculated difference value is larger or smaller than the preset threshold value stored by the user, . In the above, characteristics of the background image and the input image may be contrast correction, edge / edge angle similarity, foreground contrast edge ratio, histogram similarity, and the like. In addition, the object outline filter using the feature between the background image and the input image can be classified into a feature vector and a feature vector by using a learning device such as SVM (Support Vector Machine) It can be classified. The path of the object being tracked in FIG. 7 can be selected by the user as the path is updated and saved every frame processed. In FIG. 7, the number of tracking paths is 10, and when the path is updated, the last path position (first path position on the left side) is deleted and the current position information is newly stored. In order to determine whether the path of the tracked object is a crossing of the virtual pans line, a vector is created for the starting points of the virtual pans line vector and the virtual pans line vector set by the user and the position points of the path, It is possible to judge whether or not the path of the tracked sensing object has crossed the virtual fancy line. If the number of the traced paths crossing the virtual fancy line satisfies the following conditional expression 1, it can be determined that the object is a virtual fancy line crossing first-

Total number of paths / 2 -1 ≤ Number of crossing points ≤ Total number of paths / 2 +1 ... (Conditional Expression 1)

FIG. 8 is a second explanatory diagram for calculating the presence or absence of a virtual fas line cross in the path information of the track detection object applied to the present invention. In FIG. 8, the presence or absence of the virtual fas line crossing of the path information of the tracking sensing object according to the present invention determines the virtual fas line crossing of the primary candidate tracking sensing object, A global motion vector is obtained using the position t, and when it is determined that the global motion vector has passed the virtual path line, it is determined that the tracking sense object has crossed the virtual fuse line. In addition, it is possible to calculate the agreement degree of the crossing direction by internally crossing the global movement vector of the tracking sense object crossed with the virtual pans line and the crossing direction vector set by the user, and the crossing direction agreement If it is larger than the threshold value set by the user and registered and registered, the tracking detection object is finally determined as the crossing tracking detection object of the virtual fas line. The global motion vector (Trk_P2-Trk_p1) of the tracking sensing object is the tracking object position vector before t-10 and the current tracking position vector. Here, 10 denotes the number of tracking position paths, which can be changed by the user.

FIG. 9 is an explanatory diagram of a sequence in which a tracking sensing object applied to the present invention crosses a virtual panning line. In FIG. 9, if a crossing direction of the crossing direction of the track sensing object is established and registered, the crossing of the composite virtual fanning line determines that an intrusion detection object exists in the monitoring area. For example, if an intrusion candidate tracking object is set and stored and registered in the order of virtual px line (1) → virtual px line (2) → virtual px line (3), it is determined as an intrusion detection object. In the case of the bi-directional mode, the virtual punch line 1 → the virtual punch line 2 → the virtual punch line 3 or the virtual punch line 3 → the virtual punch line 2 → the virtual punch line 1 ) As the intrusion detection object. In FIG. 9, a virtual punch line having a distance of 0 in the virtual punch line at the center is ignored in the crossing order. When the distance of the virtual punch line 3 is 0, The virtual pose line 1 and the virtual pose line 3 judge that the virtual pose line 2 is a virtual pose line when the distance from the center virtual pose line 2 is 0 It is judged that the tracking detection object is an intrusion detection object.

When the trace detection object is finally determined as the intrusion detection object, the alarm signal is generated as much as the user sets and the generated alarm signal can be transmitted to the VMS / DVR / NVR or the mobile phone to cope with the event. It is possible to easily retrieve an image of an intrusion event section from a security image recorded in a large capacity and to store it separately.

FIG. 10 is a flowchart illustrating an intrusion detection method control using a complex virtual fuse line setting of the intrusion monitoring area of the present invention. 10, the intrusion detection method using the complex virtual fuse line setting of the intrusion monitoring area according to the present invention includes a step S11 of generating and storing a virtual fuse line in a surveillance region image received from a camera, A step S12 of generating and storing an object filter based on the virtual fancy line, a step S13 of storing an object filter based on the virtual fancy line, (S14) of judging whether or not the candidate tracking object has crossed the virtual fancy line (S15); a step (S15) of generating an alarm signal in case of intrusion; And transmitting the signal to the user terminal (S16). In the step S11, a virtual fuse line is formed or a center virtual fuse line and a center virtual fuse line are separated from each other using a mouse or a touch pointer in an input image received from the camera to form a respective auxiliary line So that it can be stored. In addition, the virtual fancy lines can be generated in a complex manner or can be generated and stored in a plurality of ways. In the step S12, a direction orthogonal to the virtual fancy line is set as a moving direction of the intrusion tracking sensing object. In step S12, the moving direction of the intrusion tracking sensing object may be set to have an angle other than a direction orthogonal to the virtual fancy line. If there is a virtual fancy auxiliary line, the center virtual fancy line and the auxiliary line You can set the direction of movement of the intrusion detection object in down / right direction mode, up / left direction mode or bi-directional mode. In step S13, the object filter setting of the intrusion detection object can set a linear perspective size filter. In the linear perspective size filter, the starting point (P1) and the end point (P2) point, and sets two pairs of maximum and minimum filters of an arbitrary size. In step S13, the candidate tracking target object can be selected using the outline filter of the tracking target object. If a linear perspective size filter is applied, the step S14 may exclude the intrusion detection object from the candidate tracking object if the intrusion detection object is not within the range of the maximum and minimum filters, When an object is selected, feature differences between the background image and the input image for the tracking area are calculated and candidates are selected based on the similarity of feature differences. In the above, the feature similarity degree between the background image and the input image may be a degree of similarity between contrast, edge / edge angle, foreground edge ratio, and histogram. In addition, the feature difference similarity determination can be made using an SVM learning device. Another method of selecting the candidate tracking object in step S14 is to create a vector for the starting points of the virtual pans line vector and the virtual pans line vectors set by the user and the position points of the paths of the tracking sensing object, ), It is possible to judge whether or not the path of the tracked object has crossed the virtual fancy line. The number of the traced paths crossing the virtual fancy line is the total number of paths / 2 -1 ≤ the number of crossing points ≤ the total number of paths / 1, it can be judged to be a virtual fas line crossing primary candidate tracking object. In step S15, a global motion vector is obtained using the current estimated position (t) from the past last tracking path point (t-10). If it is determined that the global motion vector has passed the virtual fuse line, The virtual panning line is determined to be the crossing of the virtual panning line, and the other method is characterized in that the virtual panning line is crossed and the global movement vector of the crossing detection object crossed and the crossing direction vector set by the user Direction) and the crossing direction can be calculated internally. If the crossing direction agreement degree of the crossing direction object crossed with the virtual pans line is larger than the threshold value registered and stored by the user, And a final determination is made as a crossing tracking detection object.

11 is a block diagram of an intrusion detection system using a method of setting a composite virtual fence of the present invention. 11, the intrusion detection system using the composite virtual fence setting method of the present invention includes a camera 10 for photographing a surveillance region and transmitting the surveillance region to a server in real time, a camera 10 for storing an input image received from the camera, Sets and stores a virtual fancy line for determining the intrusion of the tracking object, stores the virtual flesh line crossing movement direction of the tracking object, which can be regarded as an intrusion, and sets the virtual fancy line The object filter is set and stored and the object outline filter is set and stored. If there is a tracking object in the scope of the object filter for the tracking object in the monitoring area received from the camera, it is determined that there is a candidate tracking object. Background image (person, car, noise, etc.) and mouth And determines whether or not a candidate tracking object is present based on the path of the tracking object. If the candidate tracking object exists, it is set and stored in the past A global motion vector is obtained from a past tracking path position using the current estimated position, and when the global motion vector passes the virtual fas line, it is determined that the tracking sensing object crosses the virtual fuse line of the monitoring area, A management server 20 for generating an alarm signal and transmitting the generated alarm signal to a user terminal, a user terminal 30 for receiving and providing an alarm signal from the management server, and an Internet network Or a wireless communication network (40). The determination that the tracking sensing object has crossed the virtual fancy line indicates that it is determined to be an intrusion into the surveillance region, and the global movement vector of the tracking sensing object and the crossing of the virtual fancy line in the case of the intrusion set by the user The directional coincidence is internally calculated with respect to the direction vector, and if the direction match degree is larger than the threshold value registered and stored by the user, it can be determined that the tracked sensing object has crossed the virtual fancy line and entered.

10: camera, 20: management server,
30: user terminal, 40: internet network / wireless communication network

Claims (16)

An intrusion detection method using a virtual virtual panning line setting of an intrusion monitoring area,
An intrusion detection method using a combined virtual fuse line setting of the intrusion monitoring area,
Generating and storing a virtual fancy line in the surveillance region image received from the camera (S11);
(S12) setting and storing a moving direction of the intrusion tracking sensing object in advance based on the virtual fancy line;
Creating and storing an object filter based on the virtual fancy line (S13);
Tracking of the surveillance zone received by the camera A vector is created for the virtual pose line vector set by the user and the start point of the virtual pose line vector and the position points of the path of the track detection object for the sensing object and calculated as a cross product It is possible to judge whether or not the path of the tracking detection object has crossed the virtual fancy line and if the number of tracking paths crossing the virtual fancy line satisfies the total number of paths / 2 -1? Crossing point number? Total number of paths / 2 +1 Selecting (S14) a virtual fancy line crossing primary candidate tracking sensing object;
And determining (S15) whether or not the candidate tracking object has been infiltrated by crossing the virtual fancy line (S15).
The method according to claim 1,
(S11) of creating and storing a virtual fancy line in the surveillance region image received from the camera,
It is possible to form and store one auxiliary virtual line or a plurality of auxiliary lines spaced a certain distance around the center virtual fuse line and the center virtual fuse line by using a mouse or a touch pointer in the input image received from the camera A method for intrusion detection using a compound virtual fuse line configuration of an intrusion monitoring zone.
The method according to claim 1,
The virtual fuse line may include:
Can be generated in a complex manner or can be generated and stored in plurality
Intrusion detection method using multiple virtual pans line setting of intrusion monitoring area.
The method according to claim 1,
(S12) of setting and storing a moving direction of the intrusion tracking sensing object on the basis of the virtual fancy line,
And setting a direction orthogonal to the virtual fancy line as a direction of movement of the intrusion tracking sensing object.
The method of claim 1, wherein
(S12) of setting and storing a moving direction of the intrusion tracking sensing object on the basis of the virtual fancy line,
It is possible to set the moving direction of the intrusion tracking detection object to have an angle other than the direction orthogonal to the virtual fancy line. If there is a virtual fancy auxiliary line, the order of the center virtual fancy line and the auxiliary line is set in the moving direction setting of the intrusion detection object Wherein the direction of movement of the intrusion detection object can be set to a down / right direction mode, a up / left direction mode, or a bi-directional mode.
The method according to claim 1,
In the step S13 of creating and storing an object filter based on the virtual fancy line, the object filter setting of the intrusion-
Linear Perspective The size of the filter can be set up. The linear perspective filter has two pairs of maximum and minimum filters of arbitrary size, with the start point (P1) and end point (P2) of the virtual pans line as foot- Wherein the intrusion detection area is configured to detect the intrusion of the intruder.
The method according to claim 1,
In the step S13 of creating and storing an object filter based on the virtual fancy line, the object filter setting of the intrusion-
Wherein the candidate tracking object can be selected using an external filter of the tracking object. ≪ RTI ID = 0.0 > 8. < / RTI >
delete delete The method according to claim 1,
The step (S15) of determining whether the candidate tracking object crosses the virtual fancy line and has entered the step
If the global motion vector is found using the current estimated position (t) from the past last track path point (t-10) and the global motion vector is determined to have passed the virtual fas line, The intrusion monitoring method comprising the steps of: (a)
The method according to claim 1,
The step (S15) of determining whether the candidate tracking object crosses the virtual fancy line and has entered the step
You can calculate the degree of agreement of the crossing direction by crossing the global movement vector of the trace detection object crossed with the virtual fas line and the crossing direction vector set by the user, and if the crossing direction agreement of the tracking sense object crossed with the virtual pans line is set by the user And if it is larger than the registered threshold value, the trace detection object is finally determined as the crossing trace detection object of the virtual fuse line.
The method according to claim 1,
An intrusion detection method using a combined virtual fuse line setting of the intrusion monitoring area,
Further comprising: generating an alarm signal when the intruder is intruded and transmitting the generated alarm signal to the user terminal (S16).
delete delete An intrusion detection system using a combined virtual fuse line configuration of an intrusion monitoring area,
An intrusion detection system using a combined virtual fuse line configuration of the intrusion monitoring area,
A camera (10) for photographing a surveillance area and transmitting the surveillance area to a server in real time;
A virtual fancy line for storing an input image received from a camera and determining an intrusion of a tracking sensing object in a monitoring area image of the stored input image is set and stored and a virtual fancy line crossing movement of the tracking sensing object, Set direction and store it, set object filter based on virtual pans line set in the surveillance area, set or save the object outline filter, and track the surveillance area received from the camera. If there is a detected object, it is determined that there is a candidate tracking object, or a similarity between a background image (human, automobile, noise, etc.) and an input image is compared through an object outline filter to determine whether a candidate tracking object exists Zone of the candidate tracking detection object based on the path of the tracking detection object And if the candidate tracking object exists, a global movement vector is obtained using the current estimated position from the past tracking path position stored in the past, and if the global movement vector passes through the virtual fancy line, A management server 20 for determining that the object has crossed the virtual fuse line of the surveillance region, storing the determination result, generating an alarm signal, and transmitting the alarm signal to the user terminal;
A user terminal 30 for receiving and providing an alarm signal from the management server;
And an Internet network or a wireless communication network (40) for connecting a camera, a management server, and a user terminal via a network.
16. The method of claim 15,
The determination by the management server that the tracked sensing object has crossed the virtual fancy line,
And the directional agreement is internally calculated with respect to the crossing direction vector of the virtual fence line in the case of the intrusion set by the user, and the direction match degree is determined by the user And if it is larger than the preset threshold value, the trace detection object can be determined to have crossed the virtual fancy line and enter the intrusion monitoring area.

Images