KR20080020441A - Image monitoring system for object identification - Google Patents

Abstract

An image processing and monitoring system and a method for identifying an object are provided to track a route of a particular person by extracting features of an extracted image of the person, comparing still images recorded by a plurality of cameras, and searching a still image having similar features. An image extracting unit(3) extracts a motion image by subtracting images captured by a plurality of cameras(1) in real time and a background image created in a previous stage. A determining unit(4) discriminates whether the image extracted by the motion image extracting unit(3) is one person, converts an RGB image into an HSI(Hue,Saturation,Intensity) image, creates an image with a saturation value into an image obtained by employing a hybrid median filter, discriminates the head of the person of the created image, and checks the person based on his position and size. A management unit(5) manages variables such as a specific identifier, the name of a controller for monitoring a captured image, the number of the cameras, a generated time, an expired time, a hue value of an upper body of an identified object a hue value of a lower body of the identified object, an X-axis position of the camera of the identified object, an Y-axis position of the camera of the identified object, and the direction in which the identified object moves, and outputs them.

Description

Image Monitoring System for Object Identification and Monitoring Method {Image Monitoring System for Object Identification}

1 is a block diagram of an image processing surveillance system capable of object identification according to the present invention;

2 is a processing flowchart for identifying a person from a moving image according to the image processing monitoring system of the present invention;

3 is an exemplary view illustrating a background image creation method during a human object extraction step in an image processing surveillance system according to the present invention;

4 is an exemplary diagram of an image of converting an RGB image into an HSI image and labeling the converted saturation image during a human object extraction step in the image processing surveillance system of the present invention;

5 is an exemplary view illustrating a process of extracting a human object in an image processing surveillance system according to the present invention;

6 is a processing flowchart for searching for stored images in the image processing monitoring system of the present invention;

7 is an exemplary view of a folder stored during an image retrieval step in the image processing monitoring system according to the present invention;

8 is an exemplary view of an object image stored during an image retrieval step in an image processing surveillance system according to the present invention;

9 is an exemplary view of a search result during an image search step in an image processing surveillance system according to the present invention;

FIG. 10 is an exemplary view of collecting and displaying a search result as a video during an image search step in an image processing monitoring system according to the present invention; FIG.

11 is a view showing a list of object management variables in the image processing monitoring system of the present invention;

12 is a diagram illustrating an object management database table schema in an image processing surveillance system according to the present invention;

* Description of the symbols for the main parts of the drawings

1: camera 2: video capture unit

3: motion image extraction unit 4: determination unit

5: management unit 6: database unit

7: output unit

The present invention relates to an image processing surveillance system capable of object identification, and more particularly, to an image processing surveillance system and a surveillance method capable of object identification capable of quickly searching for a desired image from recorded image information in order to quickly identify a cause of an accident. It is about.

The recording method used in the current security system records the entire image taken by the camera without image processing. Such a security system includes an analog CCTV security system using a VCR. In the analog CCTV security system, a plurality of surveillance cameras (usually four, nine, or sixteen) are connected to a multiplexer, and a video signal transmitted from the multiplexer displays a split screen on a monitor.

The multiplexer captures the video from multiple cameras, creates a split screen (4 splits, 9 splits, 16 splits) and sends it to the monitor. In this case, the image transmission from the multiplexer to the monitor is not to transmit a split screen, but to sequentially transmit one image transmitted from each camera to the monitor. This video signal is also transmitted to the VCR. The VCR is transmitted from the camera and received and recorded on the split screen by the multiplexer.

Recording is performed by recording on magnetic tape, which is a storage medium of an analog system. In addition, the VCR can play back the recorded video. When the play button of the VCR is pressed, the video recorded on the magnetic tape is transmitted through the playback channel.

The multiplexer is configured to reconstruct sequentially recorded images into split screens and transmit them to the monitor, thereby reproducing the recorded images.

In order to identify the cause after an accident or an incident, all the recorded information should be searched and the corresponding video should be searched. Of course, if you can know the accident time and site, the search time will be shortened, but if not, you will have to spend a lot of time and effort to find the recorded accident video.

In the current security management system, a video recording method uses a method of detecting a motion or recording a video by an external sensor to shorten the recording time of the video recording.

However, most of the security-related image search is more than 95% to search for people, there is a problem that requires a lot of time and effort in the prior art described above.

An object of the present invention has been made to solve the problems described above, the image processing surveillance system and monitoring method capable of object identification that can quickly search for the desired image from the recorded image information to quickly identify the cause of the accident To provide.

Another object of the present invention is to extract a feature from the extracted human image in order to shorten the search time to compare the still images recorded by a plurality of cameras to search for still images having similar characteristics to track the path of a specific person It is to provide an image processing surveillance system and a surveillance method that can be identified.

Another object of the present invention is to track the object using the identified person object information and the ability to track the object in conjunction with the camera equipped with a pan-tilt, when invading the area given a specific boundary value It is to provide an image processing surveillance system and monitoring method capable of object identification that can be applied for alarm generation function, direction detection function, and object accumulation.

In order to achieve the above object, an object-identifiable image processing monitoring method according to the present invention includes: (a) creating a background image for extracting a human object; (b) subtracting the image captured in real time by the camera and the background image created in the previous step to extract a motion image; (c) determining whether the extracted motion image is a single person; (d) converting an RGB image to an HSI image; (e) creating an image to which the hybrid median filter is applied to the image having the saturation value; (f) identifying the head in position and size by identifying the head in the image created in step (e); (g) When the motion image extracted in step (c) is identified as a person, a unique identifier, a captured image surveillance controller name, a camera number, a generated time, an expired time, and an upper body color of the identified object (Hue). Managing variables such as a value of), a lower body color (Hue) value of the identified object, the X axis position of the camera of the identified object, the Y axis position of the camera of the identified object, and the moving direction of the identified object; (h) storing the identified variable information in the database when the identified object is deleted from the camera.

In the image processing monitoring method according to the present invention, (i) selecting a motion image and performing a search for a similar object, converting the selected motion image into upper body color (Hue) and lower body color (Hue) information; (j) searching the stored information in the database by the extracted Hue values of the upper and lower bodies to display image information similar to the selected motion image with the controller name, camera number, image capturing time information and the extracted image; It further comprises.

In the image processing monitoring method according to the present invention, (k) selecting a stored motion image refers to a time stamp of a video recorded in a corresponding MPEG4 or MJPEG with reference to generation time information, controller name, and camera number stored in a database. The method may further include automatically searching for a moving picture position where the selected moving picture is recorded.

In order to achieve the above object, the image processing surveillance system according to the present invention is an image processing surveillance system for capturing images captured by a plurality of cameras in an image capturing unit and extracting and displaying a human object. A motion image extractor which extracts a motion image by subtracting an image captured in real time and a background image created in the previous step, determines whether the image extracted by the motion image extractor is one person, and converts an RGB image into an HSI image. The image having the saturation value as a hybrid median filter image, and a judging unit for identifying a human head by position and size by identifying a human head from the created image, a unique identifier, a name of a captured video surveillance controller, Number of the camera, created time, expired time, upper body color (Hue) value of the identified object, lower of the identified object In that it comprises a color value, of the identified object, the camera X-axis position, the Y-axis position of the identified object, camera, and outputting the management of the variable such as the moving direction of the identified object management features.

In addition, the image processing monitoring system according to the present invention, characterized in that it further comprises a database unit for storing the identified variable information when the object identified for the information managed by the management unit is deleted from the camera.

The above and other objects and novel features of the present invention will become more apparent from the description of the specification and the accompanying drawings.

First, the concept of the present invention will be described.

The most important factor in the present invention is an algorithm for extracting image information for identifying a person from image information captured in real time.

Quickly cause the problem by identifying the object called human from the image being captured and extracting only the image of the person and storing it as a still image. It is an invention that can help to deal with cases by identification.

In other words, the present invention combines an algorithm that extracts a feature that can distinguish a person from another person's image and an algorithm that extracts a similar image by comparing the extracted characteristics. It is possible to provide a system for retrieving an image that can be solved.

EMBODIMENT OF THE INVENTION Hereinafter, the structure of this invention is demonstrated according to drawing.

1 is a block diagram of an image processing surveillance system capable of object identification according to the present invention.

As shown in FIG. 1, the image processing monitoring system capable of object identification according to the present invention captures images captured by a plurality of cameras 1 by an image capturing unit 2, and a background image for extracting a human object. Write in advance.

The motion image extractor 3 extracts a motion image by subtracting an image captured in real time by the plurality of cameras 1 and a background image created in the previous step.

The determination unit 4 determines whether the image extracted by the motion image extraction unit 3 is one person, converts the RGB image into an HSI image, and applies the hybrid median filter to the image having the saturation value. And identify the human head by location and size by identifying the human head in the created image.

When the moving image extracted by the determination unit 4 is determined as one person, the management unit 5 determines a unique identifier, a captured image surveillance controller name, a number of the camera 1, a generated time, an expired time, and identification. Upper body color (Hue) value of the identified object, lower body color (Hue) value of the identified object, X axis position of camera 1 of identified object, Y axis position of camera 1 of identified object, identified object Variables such as the direction of movement of the control is output to the output unit (7). The determination unit 4 and the management unit 5 are implemented as a microprocessor, and the output unit 7 is configured as a monitor for outputting a split screen (for example, 4 divisions, 9 divisions, 16 divisions) or sequential full screens. do.

In addition, the information managed by the management unit 5 stores the identified variable information in the database unit 6 when the identified object is deleted from the camera. This database unit 6 is made of a digital processing device such as a normal HDD.

Next, operations of the image processing surveillance system capable of identifying the object shown in FIG. 1 will be described with reference to FIGS. 2 to 12.

FIG. 2 is a diagram illustrating an algorithm for extracting a human image by detecting a motion in the captured image. In FIG. 2, in order to extract an image of a person, a background image photographed by the camera 1 is extracted (S101), and a motion is extracted based on a difference between the background image extracted by the motion image extractor 3 and a real-time image. (S102). The accuracy of the extracted object depends on how accurate the background image is.

Therefore, the algorithm for extracting the background is more important in the outdoors than in the interior. The identification of differences in day and night brightness should be prevented.

In the present invention, a background image is created by using two methods. First, the background image is stored at regular intervals. At this time, the image is acquired when there is no movement. Second, if it is determined that there is no motion by comparing the acquired image with the previous image as a moving image, the background image is managed as the average value of the sum of the RGB values of the current background image and the two images.

3 and 201 show algorithms for extracting a background screen. The initial background image is input by accumulating a certain number of frames manually and accumulating the average.

When the background image is generated, a motion image is extracted by subtracting the image afterwards. If there is no motion image in step S103, the background image is accumulated in the current background image to obtain a more real-time background image. In addition, if it is determined that there is no movement after a predetermined time, the corresponding image is accumulated in the background image (S103).

The motion image extracting unit 3 detects the motion of the background image and the currently captured image.

Next, the determination unit 4 compares the size of the extracted motion image (S104). At this time, if the size of the extracted motion image is larger than a certain size, it is likely that there are two or more or completely different objects. Therefore, in this case, only the image is kept and object management is not performed (S108).

When it is determined that the extracted motion image is one person by the determination unit 4, the manager 5 converts the extracted motion image into an HSI value (S105). An image having a saturation value is generated (S106) by applying hybrid median filtering suitable for boundary preservation (S106), and then a labeling operation is performed (S107). Hair colors vary widely from country to country to define the hair color for a particular country. Koreans usually have black hair. Therefore, if the labeling work consisting of saturation values is performed, the head value is distinguished as 250 or more among 0-255 values.

4 illustrates an image of a person (object) extracted as a motion image, an image obtained by HSI conversion of the image, and an image labeled with a chroma image.

The reason why the labeling is applied to the saturation image is that the saturation image shows the shape of the head more clearly than the image of the color (Hue), and it can be seen that the human head has a similar size and shape for each object. Because it is easier to extract more hair.

FIG. 5 sequentially shows an image 403 obtained by subtracting an image 402 captured from a background image 401, an image 403 obtained by extracting a motion image, an HSI-converted image S404, and a content 405 of labeling chroma images. Drawing.

When it is determined that the size of the extracted motion image is one person, the determination unit 5 determines the position of the head to determine the presence or absence of a person. If it is determined that the person, the management unit 5 is to give a unique ID and to manage various management variables (S108). The reason is that the moving object in one camera 1 must be managed continuously, so that the extracted object must be managed by giving a unique ID to track the object.

 At this time, the memory variables for managing the extracted human objects include the unique ID shown in FIG. 11, the created time, the expired time (the time disappeared from the screen), the average value of the hue values of the upper body and the lower body, the imaging position of the camera, and the immediately before. The heading direction, the photographed camera number, the photographed controller number, and the size of the acquired image are created and managed. As shown in FIG. 7, the image is stored in a conventional memory means or a database unit 6 by creating a folder for each position, camera, and date and time. This is for the convenience of the search later. When it is extinguished (disappears from the screen), the managed variable information and the corresponding motion video path information are stored in the database table of FIG. 12.

 In this case, the unique identifier (ID) needs to have a unique file name in order to be useful when searching. For this purpose, it is necessary to put the imaging position, the number of the camera 1, the date and time, the object number (number given at the moment shown), etc. in the unique identifier ID.

 Next, a method of searching for and displaying information of a person who is similar to the person to be searched using the stored information will be described.

In the image retrieval algorithm of FIG. 5, the first thing to do is to select an image to be searched for and read information on the selected image (S501). This can be selected from among the images stored for the camera 1 installed at each position, as shown in FIG.

In FIG. 7, the main reason for the image search is to find people who have moved at a specific location when an incident or accident occurs at a specific location. Therefore, as the search condition, the location information 601 should be made higher. However, since the position-specific camera number 602 and the date 603 do not have to be prioritized, the camera number does not need to be prioritized over the date as shown in FIG.

When the folder is searched for in FIG. 7, all images stored on the selected date are shown as shown in FIG. 8. In this case, when the image to be searched is selected, the corresponding image is HSI-converted and the resulting Hue image is set as a reference value (S502). The color image of the reference value is separated into an upper body and a lower body, and the values are compared with upper and lower color fields stored in a database.

Since the color values of all the frames are not the same, a threshold value is set to search for all images having a difference within this error limit.

 The retrieved images are displayed together with the information of the respective images as shown in FIG. 9. FIG. 9 may show other images when only the same images are generated or when color information is similarly coincidentally. Therefore, in this image retrieval algorithm, only the necessary images can be selected and viewed, and only the desired images can be reproduced and confirmed. As a result, as shown in FIG. 10, the image can be reproduced and viewed at a desired position and time zone. The timestamp of the video information recorded in MPEG4 or MJPEG associated with the generation time information, the controller information, and the camera information of the identified object information may be searched for information on a location where the selected object information is recorded.

The overall configuration of the image search algorithm according to the present invention is as follows. The image search algorithm finds the desired location and time zone in the saved folder (Fig. 6), selects the desired object (People) in that time zone (Fig. 7), and searches the most similar objects within the margin of error with the selected object (People). FIG. 8 collects only desired images and plays them as a video (FIG. 9).

As mentioned above, although the invention made by this inventor was demonstrated concretely according to the said Example, this invention is not limited to the said Example and can be variously changed in the range which does not deviate from the summary.

As described above, according to the image processing monitoring system and monitoring method capable of object identification in accordance with the present invention, by managing a human-oriented information in the database during the movement of the image information to be searched to search a variety of image information, events, accidents It is possible to provide the efficiency of the security management task by allowing the cause to be identified quickly.

In addition, the function can track the object by linking the camera with the path tracking and pan-tilt using the identified person object information, and the alarm function when the trespassing into the area given a specific threshold value. The effect is that it can provide the infrastructure to build a surveillance system that can be applied to the direction detection function, the presence of object accumulation, etc.

Claims (5)

(a) creating a background image for extracting a human object; (b) subtracting the image captured in real time by the camera and the background image created in the previous step to extract a motion image; (c) determining whether the extracted motion image is a single person; (d) converting an RGB image to an HSI image; (e) creating an image to which the hybrid median filter is applied to the image having the saturation value; (f) identifying the head in position and size by identifying the head in the image created in step (e); (g) When the motion image extracted in step (c) is identified as a person, a unique identifier, a captured image surveillance controller name, a camera number, a generated time, an expired time, and an upper body color of the identified object (Hue). Managing variables such as a value of), a lower body color (Hue) value of the identified object, the X axis position of the camera of the identified object, the Y axis position of the camera of the identified object, and the moving direction of the identified object; and (h) storing the identified variable information in a database when the identified object is deleted from the camera.  The method of claim 1, (i) converting the selected motion image into upper body color (Hue) and lower body color (Hue) information when the motion image is selected and searched for a similar object; (j) searching the stored information in the database by the extracted Hue values of the upper and lower bodies to display image information similar to the selected motion image with the controller name, camera number, image capturing time information and the extracted image; Video surveillance method further comprising. The method of claim 2, (k) If you select the saved motion video, it searches the time stamp of the video recorded with MPEG4 or MJPEG by referring to the creation time information, controller name and camera number stored in the database, and automatically selects the video location where the selected motion video is recorded. The video surveillance method further comprising the step of searching. In the image processing surveillance system for capturing images captured by a plurality of cameras in the image capture unit, extracting and displaying a human object, A motion image extracting unit extracting a motion image by subtracting an image captured in real time by the plurality of cameras and a background image created in the previous step; The image extracted by the motion image extractor determines whether one person, converts an RGB image into an HSI image, prepares an image having chroma values as an image using a hybrid median filter, and draws a human head from the created image. A judging unit that identifies and identifies a person by location and size, Unique identifier, the name of the image surveillance controller, the number of the camera, the time of creation, the time of expiration, the upper body color value of the identified object, the lower body color value of the identified object, the camera's X of the identified object And a management unit for managing and outputting variables such as an axis position, a Y axis position of the camera of the identified object, and a moving direction of the identified object. The method of claim 4, wherein And a database unit for storing the identified variable information when the object identified for the information managed by the management unit is deleted from the camera.

Images