KR101893875B1 - Behavior Detection System Of Discards Garbage And Behavior Detection Method Of Discards Garbage - Google Patents

Abstract

The present invention relates to a discarding behavior detection system and a discarding behavior detection method that enable accurate discrimination between a normal situation and a problem situation. The discarding behavior detection system of the present invention comprises an image processor for receiving an image packet according to an elapse of time from the outside, converting the image packet into an image frame, and storing the image frame; a background frame generating unit for generating one of the output image frames received from the image processor as a background frame; a difference frame generating unit for receiving the background frame and the image frame, dividing the latest image frame and the background frame among the image frames into a plurality of blocks, and generating a block having a difference between the background block of the background frame and the image block of the latest image frame as a difference frame; an object storage detecting unit for storing a normal object and detecting a first object matching with the normal object in the difference frame or a second object mismatching with the normal object by comparing the background frame and the difference frame; a scoring unit for scoring a block of a difference frame including the first object when the first object is detected in the object storage detecting unit; and a discarding object sensing unit for adding the score of the scored block to obtain an added value if a second object is detected in the object storage detecting unit and a difference frame including the second object is generated in the difference frame generating unit over a reference time, and sensing the second object as a discarding object if the added value becomes less than a reference score upon comparison of the added value with the reference score.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to an organic behavior detection system,

The present invention relates to an organic behavior detection system and a organic behavior detection method.

In recent years, closed circuit television (CCTV) has been used in various places such as banks, public institutions, and environmental monitoring areas such as alleys. Since it is used in various places of closed-loop television, development of technology capable of more accurately detecting objects to be photographed for efficient monitoring is actively under way.

Particularly, an image sensing system for analyzing image information and automatically detecting an abnormal behavior and sending an alarm to the manager is actively being developed. The Intelligent Video Surveillance (IVS) system consists of an image preprocessing step, a background region separation step, an object identification step, an object tracking step, and an event detection step for detecting an event based on a predefined rule.

The intelligent image detection system consisting of these steps must accurately detect the event in order to obtain the detection efficiency and high reliability.

However, most of the currently developed intelligent image detection systems are generating frequent errors when detecting events.

For example, even when an object is caught in a place where the shape, position, etc. are changed by the wind, such as a tree or a door panel, or a place where there is no change due to a severe illumination change, an event is detected. This problem also causes a problem of lowering the reliability of event detection.

Korean Patent No. 10-1519342 (Publication date 2015.05.13)

SUMMARY OF THE INVENTION Accordingly, the present invention has been made keeping in mind the above problems occurring in the prior art,

More specifically, the present invention allows a person to roam in a certain area and more accurately detect an organic behavior, that is, a problem behavior, of an object.

The problems to be solved by the present invention are not limited to the above-mentioned problems, and other technical problems which are not mentioned can be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention,

An image processor for receiving an image packet from the outside and converting the image packet into an image frame according to passage of time and storing the image frame;

A background frame generator for generating one of the output image frames received from the image processor as a background frame;

A block having a difference between a background block of the background frame and an image block of the latest image frame is divided into a plurality of blocks, A differential frame generator for generating a differential frame;

An object storage detector for storing a normal object and detecting a first object matching the normal object or a second object mismatched with the normal object in the differential frame by comparing the background frame and the difference frame;

A score assigning unit for assigning a score to a block of the differential frame including the first object when the first object is detected in the object storage detecting unit;

When the second object is detected by the object storage detection unit and the difference frame including the second object is generated by the difference frame generation unit over a reference time, the score of the block to which the score is assigned is added to obtain an addition value And an organic object sensing unit for comparing the added value with a reference score and sensing the second object as an organic object when the added value becomes the reference score or more.

When the second object is generated, the organic object detecting unit may add the points given to each block before and after the difference frame including the second object, and then compare the score with the reference score.

The scoring unit may assign a negative value until the score of the block becomes a zero point at a predetermined time interval when the first object disappears in the block.

The organic object detecting unit may generate an event signal when an added value of the points given to each of the blocks before and after the up and down and the next and after the difference frame including the second object is equal to or greater than the reference point.

(A) receiving an image packet from the outside according to the passage of time, converting the image packet into an image frame, and storing the converted image frame in the image processing unit;

(B) generating at least one image frame output from the image processing unit as a background frame;

The background frame and the newest image frame are divided into a plurality of blocks, and a block having a difference between the background block of the background frame and the image block of the latest image frame, (C) generating a difference frame;

(D) detecting a first object matching the normal object or a second object mismatched with the normal object in the differential frame by comparing the background frame and the difference frame in an object storage detection unit storing a normal object, step;

A step (E) of assigning a score to a block of the differential frame including the first object when the object storage detecting unit detects the first object;

When the second object is detected by the object storage detection unit and the difference frame including the second object is generated by the difference frame generation unit over a reference time, the score of the block to which the score is assigned is added to obtain an addition value Comparing the addition value with a reference score, and (F) sensing the second object as an organic object when the addition value is equal to or greater than the reference score.

INDUSTRIAL APPLICABILITY The organic behavior detection system and organic behavior detection method according to the present invention can accurately distinguish between a situation where garbage is discharged and a situation where no garbage is discharged. In this way, it is possible to accurately distinguish the situation in which garbage is generated in the front or the alley of another person, and to generate the notification information accurately.

Therefore, by generating the notification information only when the garbage is generated, the reliability of the generated notification information can be increased.

Further, through the highly reliable notification information, when the notification information is generated, the user is relieved of the situation that the garbage is evacuated and the people are seen by the alley or the like, so that the garbage can be recognized and the quick response can be made .

1 is a block diagram of an organic behavior detection system according to an embodiment of the present invention.
2 is a diagram showing an operation state of the image processing unit of FIG.
FIG. 3 is a diagram showing an operation state of the background frame generation unit and the difference frame generation unit of FIG. 1;
4 is a diagram illustrating a normal object stored in the object storage detection unit of FIG.
FIG. 5 is a diagram illustrating a state in which a first object and a second object are detected in the object storage detection unit of FIG. 3. FIG.
FIG. 6 and FIG. 7 are views showing the score addition and operation states of the organic object detection unit of FIG. 1;
8 is a diagram illustrating an event signal generation according to the operation of the organic object detection unit of FIG.
9 is a block diagram of an organic behavior detection system according to an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention and methods for achieving them will become apparent with reference to the embodiments described in detail below with reference to the accompanying drawings.

The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the invention to those skilled in the art. It is provided to inform the person completely of the scope of the invention.

The claims of the present invention can be defined by the claims supporting the claims, including the claims. In addition, like reference characters refer to like elements throughout the specification.

In addition, the 'first object' described in the present specification is an example of a person as shown in the figure for convenience of explanation, the 'second object' is an example of a tree and trash dampening bag, But is not limited to a person and may be changed by automobiles, motorcycles, etc. depending on the situation and the user's setting. And the second object can be changed to any number of boxes, bags, and the like.

Hereinafter, the organic behavior detection system according to an embodiment of the present invention will be described in detail with reference to FIGS. 1 to 8, and the organic behavior detection method will be described in detail with reference to FIG.

First, an organic behavior detection system according to an embodiment of the present invention will be described with reference to FIG.

1 is a block diagram of an organic behavior detection system according to an embodiment of the present invention.

The organic behavior detection system 1 according to an embodiment of the present invention divides a background frame and an image frame into a plurality of blocks and extracts a difference frame between the background frame and the image frame. Then, after determining whether the object included in the extracted difference frame is a normal object, a score is given to the block when it is determined as a normal object. Then, the score assigned to the block is added and then compared with the reference score to accurately detect whether or not the waste is organic.

In other words, the organic behavior detection system 1 according to an embodiment of the present invention accurately distinguishes a normal situation from a problem situation.

The organic behavior detection system 1 includes an image processing unit 10, a background frame generation unit 20, a difference frame generation unit 30, an object storage detection unit 40, a score addition unit 50, 60 are connected so as to transmit and receive mutual data communication, so that it is possible to accurately distinguish between a situation in which garbage is emitted and a situation in which garbage is not emitted, by transmitting and receiving data.

Hereinafter, the components constituting the organic behavior detection system and the operation of each component will be described with reference to Figs. 2 to 8. Fig. FIG. 2 is a diagram illustrating an operation state of the image processing unit of FIG. 1, FIG. 3 is a diagram illustrating an operation state of a background frame generation unit and a difference frame generation unit of FIG. 1, FIG. 5 is a diagram illustrating a state in which a first object and a second object are detected by the object storage detection unit of FIG. 3, and FIGS. 6 and 7 are views showing a state in which the score addition and the operation of the organic object detection unit Fig. And FIG. 8 is a view illustrating an event signal generation according to the operation of the organic object detection unit of FIG.

The image processing unit 10 receives the image packet 101 as shown in FIG. 2A and converts it into an image frame 102 as shown in FIG. 2B. Then, the converted image frame 102 is transmitted to the background frame generation unit 20.

The image processing unit 10 may include a video packet receiving module and a decoding module. Accordingly, the video processing unit 10 can sequentially receive the video packets 101 from an external network such as a camera over time. Then, the received video packet can be sequentially decoded and converted into an image frame having a specific resolution, that is, a picture.

The background frame generation unit 20 sequentially receives the image frames 102 from the image processing unit 10. The background frame generation unit 20 may generate one of the received image frames 102 as a background frame 201. [

The background frame generation unit 20 can generate the background frame 201 as the image frame 102 before the first object 311 and the second object 312 to be described later are detected. That is, the background frame generation unit 20 can generate the background frame 201 of the past image frame 102.

The difference frame generation unit 30 receives the background frame 201 from the background frame generation unit 20 and receives the image frame 102 from the image processing unit 10 to receive the background frame 201 and the image frame 102 Can be divided into a grid (Grid) form as shown in FIG. 3 (a) and FIG. 3 (b).

Here, the grid may be a type in which sixteen blocks are formed in the horizontal direction and nine blocks are formed in the vertical direction as shown in the figure, and 32 blocks are formed in the horizontal direction and 18 blocks are formed in the vertical direction Type. These grids may be formed in various types and may be applied to the image frame 102 and the background frame 201.

Through this grid, a plurality of background blocks are formed in the background frame 201, and image blocks are formed in the image frame 102. Here, the background block and the image block may be formed to have the same size.

The difference frame generation unit 30 generates a difference frame 310 in which the difference between the image block 302 of the most recently transmitted image frame, i.e., the latest image frame, and the background block 202 is different.

The difference frame generation unit 30 transmits the generated difference frame 310 to the object storage detection unit 40.

The object storage detection unit 40 stores a normal object in advance before the image processing unit 10, the background frame generation unit 20, and the difference frame generation unit 30 operate.

Here, the normal object is separated from the background of the camera image as shown in (a) of FIG. 4, the shadowed object as shown in (b) of FIG. 4 is removed, and the part of the object is learned based on Convolutional Neural Net It is possible to obtain the extracted object through the process.

Such a normal object is intended to be a person so that the description herein will be concise and clear. However, a person is not defined as a normal object, and a car or a motorcycle can be a normal object as needed.

The object storage detecting unit 40 compares the background frame 201 and the difference frame 310 in a state where the normal object is stored and stores the normal object in the difference frame 310 as shown in FIG. And detects the first object 311 or the second object 312 that does not match the normal object. Here, the first object 311 may be a person who matches a normal object, and the second object 312 may be a tree or a mouth deck that does not match a normal object.

When the first object 311 is detected by the object storage detection unit 40, the score addition unit 50 assigns a number greater than 0 to the block including the first object 311. [ For example, as shown in FIG. 6A, the score addition unit 50 may assign a point to the block in which the first object 311 is detected. At this time, the point assigning unit 50 assigns one point to each block in which the first object 311 exists for every 0.03 second, for example, as shown in FIG. 6 (b) .

In addition, when the first object 311 disappears in the block, the score assigning unit 50 may add a number less than 0 until the score of the block becomes zero at a predetermined time interval, and subtract the score assigned to the block.

Here, the constant time may be set to various times such as 1 second or 0.5 seconds. For example, as shown in FIG. 7, a block to which three points are assigned may be subtracted by one point corresponding to a predetermined time t after the first object 3111 disappears.

In this way, the score addition unit 50 can obtain a score in which the first object 311 does not exist, that is, a block in which no person exists, and a score approximate to 0 or 0 .

When the second object 312 is detected in the object storage detection unit 40, the score addition unit 50 assigns a zero point to the block including the second object 312, And maintains the added value at the zero point state. The organic object detection unit 60 receives data from the object storage detection unit 40 and the difference frame generation unit 30 and calculates whether the object is organic or not.

More specifically, the organic object detection unit 60 detects the second object 312 after the first object 311 has disappeared at the position where the first object 311 has stayed, If a difference frame including the second object 312 is generated for more than 10 seconds, for example, 10 seconds, the score added to the neighboring block of the second object 312 is added to generate an added value.

Here, the organic object sensing unit 60 calculates the addition value by adding the points given to each block before and after the difference frame including the second object 312. Then, the added value thus calculated is compared with the reference score.

The organic object detecting unit 60 compares the added value with the standard score and detects the second object 312 as the organic object A, that is, the garbage bag, when the added value becomes equal to or greater than the reference score.

At this time, the organic object detection unit 60 may generate an event signal by sensing the second object 312 as an organic object. For example, the event signal generated by the organic object detection unit 60 may be transmitted to the notification device S. Then, the notification device can be operated as shown in FIG. 8 (b). Here, the generated notification signal is generated only when the garbage is generated, and it is possible to accurately inform people that the garbage is generated when the notification signal is generated.

Therefore, the organic behavior detection system 1 can reliably generate the notification signal by inducing a person inside the building to recognize the notification signal and to exit the building when the notification signal is generated based on the high reliability of the notification signal You can not.

Hereinafter, the organic behavior detection method according to an embodiment of the present invention will be described in detail with reference to the organic behavior detection system 1 for detecting the objects described so far. The description of the components constituting the organic behavior detection system and the characteristics of the components can be directly applied to the organic behavior detection method.

The organic behavior detection method is based on the flowchart of FIG.

9 is a flowchart of a method of detecting organic behavior in an embodiment of the present invention.

The organic behavior detection method divides the background frame 201 and the image frame 102 into a plurality of blocks and extracts a difference frame between the background frame and the image frame in a series of processes. Then, if it is determined that the object included in the extracted difference frame is a normal object, when it is determined as a normal object, the score is given to the block, the score given to the block is added, Can be accurately detected.

Hereinafter, steps of the organic behavior detection method will be described in detail.

The organic behavior detection method starts with step (A) of receiving an image packet from the outside according to the elapse of time, converting the image packet into an image frame, and then storing the image frame in the image processing unit 10 (S110). Here, the image processing unit 10 may be formed as a queue, and may store the image data in the order in which they are transmitted, and may output image packets in the image frame 102 in the stored order.

Thereafter, the organic behavior detection method proceeds to step (B) of generating at least one image frame output from the image processing unit as a background frame (S120). At this time, the background frame may be one of the input image frames.

After step (B), after receiving the latest image frame from the background frame 201 and the image processing unit, the background frame and the latest image frame are divided into a plurality of blocks, and then the background block and the image block The process proceeds to step (C) (S130) of generating a difference frame as a difference frame.

Thereafter, in the object storing and detecting unit 40 storing the normal object, the first object 311 matching the normal object or the second object 312 mismatched with the normal object in the difference frame is compared with the background frame and the difference frame, (D) (step S140).

When the first object is detected in the object storage detection unit 40, the process proceeds to step (E) of assigning a score to a block of a differential frame including the first object 311.

Here, step (E) may be performed in detail (step S150) of identifying the first object and assigning a score to the block.

After the step (E), if the second object 312 is detected in the object storage detection unit 40 and a difference frame including the second object is generated in the difference frame generation unit longer than the reference time, (S60). If the added value is greater than the reference score, the process proceeds to step (F) (step S60). In step (F), a differential frame including a second object is generated for 10 seconds or more, for example, at least a reference time in the differential frame generation unit (S180). When a differential frame including the second object is generated, Adding the scores of the assigned blocks to obtain an added value, comparing the added value with the reference score, and detecting the second object as an organic object when the added value is equal to or greater than the reference score.

If the second object 312 is detected as an organic object in step (F), an organic signal may be generated in step S190.

Further, the event signal may be transmitted to the notification device S to output a sound.

The organic behavior detection method performs the above-described steps in a series of steps, and accurately distinguishes between a situation in which garbage is generated and a situation in which no garbage is generated.

The data output from whether or not the garbage is collected can accurately operate the notification device and increase the reliability of the notification signal for the garbage collection.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, You will understand. It is therefore to be understood that the embodiments described above are in all respects illustrative and not restrictive.

1: Organic behavior detection system 10: Image processing unit
101: video packet 102: video frame
20: background frame generation unit 201: background frame
202: background block 30: differential frame generator
301: latest image frame 302: image block
310: differential frame 311: first object
312: second object
40: object storage detecting unit 401: normal object
50: Scoring part 60: Organic object detection part
A: Organic object S: Notification device

Claims (5)

An image processor for receiving an image packet from the outside and converting the image packet into an image frame according to passage of time and storing the image frame;
A background frame generator for generating one of the output image frames received from the image processor as a background frame;
A block having a difference between a background block of the background frame and an image block of the latest image frame is divided into a plurality of blocks, A differential frame generator for generating a differential frame;
An object storage detector for storing a normal object and detecting a first object matching the normal object or a second object mismatched with the normal object in the differential frame by comparing the background frame and the difference frame;
A score assigning unit for assigning a score to a block of the differential frame including the first object when the first object is detected in the object storage detecting unit;
When the second object is detected by the object storage detection unit and the difference frame including the second object is generated by the difference frame generation unit over a reference time, the score of the block to which the score is assigned is added to obtain an addition value And an organic object detecting unit for comparing the added value with a reference score and detecting the second object as an organic object when the added value becomes the reference score or more. The method according to claim 1,
Wherein the organic object detection unit adds a score given to each block before and after the difference frame including the second object when the second object is generated and then compares the score with the reference score, . The method according to claim 1,
Wherein the scoring unit assigns a negative value until a score of the block becomes zero at a predetermined time interval when the first object disappears in the block. 3. The method of claim 2,
Wherein the organic object detecting unit generates an event signal when an added value of the points given to each of the blocks before and after the up and down and the left and right of the difference frame including the second object is equal to or greater than the reference point. (A) receiving an image packet from the outside according to the passage of time, converting the image packet into an image frame, and storing the converted image frame in the image processing unit;
(B) generating at least one image frame output from the image processing unit as a background frame;
And a controller for receiving the background frame and the latest image frame from the image processor, dividing the background frame and the latest image frame into a plurality of blocks, and generating a difference between the background block of the background frame and the image block of the latest image frame (C) generating a difference frame as a difference frame;
(D) detecting a first object matching the normal object or a second object mismatched with the normal object in the differential frame by comparing the background frame and the difference frame in an object storage detection unit storing a normal object, step;
A step (E) of assigning a score to a block of the differential frame including the first object when the object storage detecting unit detects the first object;
Wherein when the second object is detected in the object storage detection unit and the difference frame including the second object is generated in a difference frame generation unit performing the step (C) over a reference time, (F) detecting the second object as an organic object when the added value is greater than or equal to the reference score by comparing the added value with a reference score, Detection method.

Images