KR20140048539A - Method and apparatus for recognizing object using surrounding image of vehicle - Google Patents

Abstract

Disclosed are a device and a method for recognizing an object using a surrounding image of a vehicle. The device recognizes an interest object within a surrounding image of a vehicle using a plurality of cameras preinstalled in the vehicle. The device for recognizing an object comprises an image acquisition unit for acquiring a photograph image photographed by a plurality of cameras; a preprocessing unit for correcting a radiation distortion and a junction error for the photographed image based on image selection information inputted by operation and generating a correction image; an image synthesis unit for generating the correction image as a single synthesis image corresponding to the image selection information; a background separation unit for determining a foreground pixel corresponding to a moving object using a background pixel designated for the background having no movement according to a certain standard among the pixels within the synthesis image and generating a foreground image including the foreground pixel; an interest object tracking unit for tracking an interest object using the foreground image; and an output unit for outputting surrounding image result information including the tracked interest object. [Reference numerals] (220) Image acquisition unit; (230) Preprocessing unit; (240) Image synthesis unit; (260) Front view background separation unit; (270) Interest object tracking unit; (280) Surrounding image information generation unit

Description

Object recognition apparatus and method using surrounding image of vehicle {Method and Apparatus for Recognizing Object Using Surrounding Image of Vehicle}

The present embodiment relates to an object recognition apparatus and method using a surrounding image of a vehicle. More specifically, by compensating the radiation distortion and the joint error of the captured image taken by using a high-quality wide-angle camera, synthesized into a single image such as panoramic view, corner view and top view, and synthesized synthesized image in the foreground and background image The present invention relates to an object recognition apparatus and method using a surrounding image of a vehicle that recognizes an object of interest.

The contents described in this section merely provide background information on the present embodiment and do not constitute the prior art.

Conventional vehicle surrounding image systems correct each image photographed by a camera installed in a vehicle in a top view form, and generate an overlay processed image using weights of pixels for a plurality of corrected images. To display. In addition, the conventional system has a problem in that the recognition distance and resolution are limited in recognizing the surrounding image by using a D1-class (640x480) camera. In addition, the conventional vehicle peripheral imaging system has a problem that should be used only for limited image processing such as image registration and lane transplantation.

Therefore, using the HD (720p) camera to overcome the limitations of the existing recognition distance and resolution, and also to recognize the moving objects around the vehicle at the same time requires a technology that automatically informs the collision avoidance when parking or driving. .

The present embodiment synthesizes a single image such as a panoramic view, a corner view, and a top view by correcting a radiation distortion and a joint error of a captured image photographed using a high-quality wide-angle camera, and synthesizes the synthesized image in the foreground and background images. The main object of the present invention is to provide an apparatus and method for recognizing an object using a surrounding image of a vehicle that recognizes an object of interest.

According to an aspect of the present embodiment, an apparatus for recognizing an object of interest in a surrounding image of a vehicle by using a plurality of cameras pre-installed in a vehicle, the apparatus comprising: an image acquisition unit configured to acquire a captured image photographed by the plurality of cameras; A preprocessing unit generating a corrected image by correcting a radiation distortion and a bonding error of the photographed image based on the image selection information inputted through the manipulation; An image synthesizer configured to generate the corrected image as one composite image corresponding to the image selection information; A foreground background separator configured to determine a foreground pixel corresponding to a moving object by using a background pixel designated for a background without movement according to a predetermined criterion among the pixels in the composite image, and generate a foreground image including the foreground pixel; An object of interest tracking unit that tracks the object of interest using the foreground image; And an output unit configured to output surrounding image result information including the tracked object of interest.

According to another aspect of the present embodiment, a method of recognizing an object of interest around a vehicle, the method comprising: an image acquiring process of acquiring a captured image photographed by a plurality of cameras attached to the vehicle; A preprocessing step of correcting a radiation distortion of the photographed image and correcting a junction error of the photographed image based on a previously stored junction reference image; An image synthesizing process of synthesizing a plurality of the corrected captured images into one image selected from a panoramic view, a corner view, and a top view; A foreground background separation process of determining a foreground pixel by using a background pixel designated according to a predetermined criterion among pixels in the synthesized image, and generating a foreground image including the foreground pixel; An object of interest tracking process of tracking the object of interest using the foreground image; And an output process of outputting surrounding image result information including the tracked object of interest.

As described above, according to the present exemplary embodiment, the radiation distortion and the bonding error of the photographed image photographed using the high-quality wide-angle camera are corrected and synthesized into a single image such as a panoramic view, a corner view, and a top view. By separating the image into a foreground and background image to recognize the object of interest, the recognition distance and resolution of the surrounding image can be increased, and the surrounding image and moving objects (eg, pedestrians, vehicles, motorcycles, and animals) can be recognized simultaneously. This has the effect of providing more accurate alarms for collision avoidance.

1 is a block diagram schematically illustrating a system for recognizing an object by using surrounding image information of a vehicle according to the present embodiment.
2 is a block diagram schematically showing an object recognition apparatus according to the present embodiment.
3 is a flowchart illustrating a method of recognizing an object by using surrounding image information of a vehicle according to the present embodiment.
4 is an exemplary diagram for describing an operation of generating a recognition reference image according to the present embodiment.
5 is an exemplary diagram for describing an operation of generating a reference image for correcting radiation distortion according to the present embodiment.
6 is an exemplary view for explaining a captured image of a vehicle according to the present embodiment.
7 is an exemplary view for explaining a method of detecting an object by the object recognition apparatus using the top view image according to the present embodiment.

Hereinafter, the present embodiment will be described in detail with reference to the accompanying drawings. It should be noted that, in adding the reference parts to the constituent elements of the respective drawings, the same constituent elements are denoted by the same reference numerals as much as possible even if they are displayed on different drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

In describing the components of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are intended to distinguish the constituent elements from other constituent elements, and the terms do not limit the nature, order or order of the constituent elements. When a component is described as being "connected", "coupled", or "connected" to another component, the component may be directly connected to or connected to the other component, It should be understood that an element may be "connected," "coupled," or "connected."

1 is a block diagram schematically illustrating a system for recognizing an object by using surrounding image information of a vehicle according to the present embodiment.

The vehicle surrounding image recognition system according to the present exemplary embodiment includes an image photographing apparatus 110, an object recognition apparatus 120, and a monitoring apparatus 130. Here, the image capturing apparatus 110 may be implemented by including a plurality of cameras such as the first camera 112, the second camera 114, the third camera 116, and the fourth camera 118.

The image capturing apparatus 110 refers to a device for capturing a captured image of a vehicle surroundings. The image capturing apparatus 110 according to the present embodiment may capture an image for recognizing and detecting an object moving around the vehicle. It is preferable to implement such as. Here, the photographed image may include a plurality of frames photographed by the plurality of cameras 112, 114, 116, and 118. Also, the image capturing apparatus 110 may be configured as a wide-angle camera capable of capturing an image of HD quality, and the captured image may be an image having a resolution of at least 720p, that is, a resolution of 1280x720. Here, the 720p resolution refers to a photographed image having 720 pixels vertically in the photographed image and simultaneously outputting even and odd lines 60 times per second without intersecting the scan lines.

The image capturing apparatus 110 includes a first camera 112, a second camera 114, a third camera 116, and a fourth camera 118 for photographing the front, rear, left, and right sides of the vehicle. Each camera photographs a different direction with respect to the vehicle, and a plurality of cameras 112, 114, 116, and 118 may have a region of an image photographed in duplicate.

The object recognizing apparatus 120 acquires a plurality of captured images captured by the image capturing apparatus 110, and corrects the radiation distortion and the joint error of each obtained captured image to synthesize the single image. . Here, the object recognizing apparatus 120 corrects a radiation distortion in which the image is expanded or contracted radially from the center of the captured image by the performance of a wide-angle camera having a wide viewing angle or an external environment such as the atmosphere or the sun. In addition, the object recognizing apparatus 120 performs a correction so as to accurately connect between the captured images by reducing the error of the junction portion of each of the captured images when combining the image of one of the panoramic view, corner view and top view. Here, the object recognizing apparatus 120 may correct the splicing error by using the splicing reference image such that the spliced error of the plurality of captured images is less than 8 cm.

The object recognition apparatus 120 performs an operation of tracking the object of interest determined by separating the foreground and the background from the combined image such as a panoramic view, a corner view, and a top view. Here, the object recognizing apparatus 120 designates a plurality of nodes having a weight greater than a predetermined criterion as a background pixel among pixels including a predetermined number of Gaussian nodes in the composite image. The foreground pixel is determined within the synthesized image based on the designated background pixel, and the foreground image of the determined foreground pixel is separated. Here, the object recognizing apparatus 120 compares at least one or more of color information, histogram, shape of the object, etc. of the plurality of objects included in the separated foreground image to determine the same object, and the determined object of interest. To track.

The monitoring device 130 outputs surrounding image information including the object of interest recognized by the object recognizing device 120 and includes a display for monitoring. Here, the monitoring device 130 is preferably connected to the object recognition device 120 in a wired manner, but is not necessarily limited thereto, and the surrounding image information may be short-range communication such as a wireless LAN signal or Bluetooth, Zigbee, or the like. Can also be received using. Meanwhile, although the monitoring device 130 is described as being implemented separately from the object recognition device 120 in FIG. 1, the monitoring device 130 is not necessarily limited thereto and may be implemented as one device including a display unit in the object recognition device 120. It may be.

In addition, the monitoring device 130 may display and display the object of interest including the surrounding image information so that the object of interest is highlighted in the form of a rectangle or a circle.

2 is a block diagram schematically showing an object recognition apparatus according to the present embodiment.

The object recognition apparatus 120 according to the present exemplary embodiment includes an image processor 210 and an object recognizer 250. Here, the image processor 210 includes an image acquirer 220, a preprocessor 230, and an image synthesizer 240, and the object recognizer 250 includes a foreground background separator 260 and an object of interest tracker. 270 and the surrounding image information generator 280.

The image processor 210 performs an operation of compensating the captured image and synthesizing it into one image. Here, the components included in the image processor 210 will be described in more detail as follows.

The image acquisition unit 220 performs an operation of acquiring a captured image of the surrounding of the vehicle by the image photographing apparatus 110. Here, the photographed image is an image photographed before, after, left and right of the vehicle through a plurality of cameras included in the image capturing apparatus 110, and means a respective image for each camera. In addition, the photographed image is preferably one captured frame, but is not necessarily limited thereto, and may be an image in which a plurality of frames are combined. For example, the image acquisition unit 220 may be a first photographing of the surrounding of the vehicle from the first camera 112, the second camera 114, the third camera 116, and the fourth camera 118 previously installed in the vehicle. An image (front), a second captured image (rear), a third captured image (left side), and a fourth captured image (right side) may be acquired. Here, the plurality of cameras 112, 114, 116, and 118 is preferably a wide-angle camera capable of capturing an image of HD quality, and the photographed image may be an image having a resolution of at least 720p, that is, a resolution of 1280x720.

The preprocessor 230 performs an operation of correcting the radiation distortion and the bonding error of the plurality of captured images acquired by the image acquirer 220.

The preprocessor 230 corrects the distortion of the captured image in which the radiation distortion is caused by the performance of the wide-angle camera having a wide viewing angle or an external environment such as the atmosphere or the sun. Radiation distortion refers to a phenomenon in which the image is expanded or contracted radially from the center of the photographed image.

The preprocessor 230 may correct the radiation distortion by using at least one of statistical correction, histogram matching correction, and drop line correction. Here, the statistical correction means a method of correcting by converting the remaining portions except for a specific row or column based on the pixel mean and standard deviation of a predetermined specific row or column, and the histogram matching correction is contrast between images. It means a method of correcting the histogram matching technique that fits to a specific row or column. In addition, drop line correction refers to a method of correcting by applying an average value of pixel information adjacent to a specific row or column to a drop line in which pixel information does not exist.

The preprocessor 230 corrects a bonding error (that is, an error distance that is a difference between the actual distance and the image) of the captured image by using the previously generated bonding reference image. Here, the preprocessor 230 performs correction to reduce the error of the joint portion of each photographed image so as to accurately connect the photographed images when generating the panorama view, the corner view, and the top view in the image synthesizer 240. . Here, the previously generated bonding reference image refers to an image obtained by photographing a grid grid that is pre-installed on the bottom surface of the space where the vehicle is located, through a plurality of cameras attached to the vehicle, and the preprocessor 230 based on the captured image. Bonding correction is performed on the captured images of the front, rear, left and right sides of the vehicle. Here, the preprocessing unit 230 corrects the bonding error such that the distance between the plurality of photographed images and the actual distance is less than 8 cm using the bonding reference image. desirable.

The image synthesizer 240 synthesizes the plurality of captured images corrected by the preprocessor 230 into one image. The image synthesizing unit 240 according to the present embodiment synthesizes the plurality of photographed images corrected for the radiation distortion and the joint error into at least one selected image selected from the user among panorama view, corner view, and top view. Here, the panoramic view refers to an image generated by joining an image photographed for a part or all directions of the vehicle by using each camera attached to the vehicle, and the corner view is at least one of four corners of the vehicle selected by the user. Means an image generated by joining the images taken for the corner of the, and means the image generated by looking down from the top of the vehicle using the image taken by each camera attached to the vehicle.

The object recognizer 250 performs an operation of tracking an object of interest determined by separating the foreground and the background from the combined image such as a panoramic view, a corner view, and a top view. In addition, the object recognition unit 250 may store a recognition reference image measuring a resolution of recognizing an object by photographing a calibration chart and recognize an object existing within a predetermined distance based on the recognition reference image. .

The components included in the object recognizer 250 will be described in more detail as follows.

The foreground background separator 260 separates the composite image such as the panoramic view, the corner view, and the top view into a foreground pixel and a background pixel to separate the foreground image. In more detail, the foreground background separator 250 designates a plurality of nodes having a weight greater than a predetermined criterion as a background pixel among pixels including a predetermined number of Gaussian nodes. The foreground pixel is determined in the composite image based on the designated background pixel. Here, the foreground background separator 260 separates the foreground image by using the determined foreground pixel. For example, the foreground background separator 250 designates a background pixel according to a weight of a Gaussian node when a person passes in the composite image, which is a panoramic view, and determines that the foreground pixel is a foreground pixel because the pixel for the passing person does not correspond to the background pixel. In addition, the image of the moving person can be separated into the foreground image.

The object-of-interest tracking unit 270 tracks the object of interest by using the foreground image separated by the foreground background separator 260. Here, the foreground image may be displayed as a value of 1, and the image corresponding to the background may be displayed as a value of 0 and divided.

The object-of-interest tracking unit 270 compares at least one or more of color information, histogram, shape of the object, etc. of the plurality of objects included in the separated foreground image in the synthesized time series. If the coordinate value is located within the range, it is determined as the same object and the determined object of interest is tracked. Here, the object-of-interest tracking unit 270 may track the object of interest by storing a trajectory coordinate value at which the object of interest moves.

The surrounding image information generation unit 280 generates surrounding image information including the tracked object of interest and transmits the generated peripheral image information to the monitoring apparatus for outputting the surrounding image of the vehicle.

3 is a flowchart illustrating a method of recognizing an object by using surrounding image information of a vehicle according to the present embodiment.

The object recognition apparatus 120 obtains a plurality of captured images from the image photographing apparatus 110 (S310). Here, the plurality of captured images are photographed by a wide-angle camera capable of capturing images of HD quality, and the photographed images are images having a resolution of at least 720p, that is, a resolution of 1280x720.

The object recognizing apparatus 120 corrects the radiation distortion and the bonding error of each captured image based on the stored reference image (S320). Here, the object recognizing apparatus 120 corrects the distortion of the captured image in which the radiation distortion is caused by the performance of the wide-angle camera having a wide viewing angle or an external environment such as the atmosphere or the sun. In addition, the object recognizing apparatus 120 corrects an error of a joint portion of each captured image when generating a composite image such as a panoramic view, a corner view, and a top view, based on the previously generated joint reference image.

The object recognizing apparatus 120 synthesizes the plurality of captured images corrected in operation S320 into a selected image among a panoramic view, a corner view, and a top view (S330). Here, the panoramic view refers to an image generated by joining an image photographed for a part or all directions of the vehicle by using each camera attached to the vehicle, and the corner view is at least one of four corners of the vehicle selected by the user. It refers to an image generated by joining the image photographed for the corner of, and the synthesized image as seen from the top of the vehicle by using the image taken by each camera attached to the vehicle.

The object recognition apparatus 120 separates the synthesized image into the foreground and the background (S340). The object recognizing apparatus 120 designates a plurality of nodes having a weight greater than a predetermined value as a background pixel among the pixels including a predetermined number of Gaussian nodes in the composite image, and synthesizes them based on the specified background pixel. Foreground pixels are determined in the image, and the foreground image is separated using the determined foreground pixels.

The object recognition apparatus 120 tracks a moving object of interest using the foreground image (S350). The object recognizing apparatus 120 compares at least one or more of color information, histogram, and shape of the plurality of moving objects in the foreground image separated in operation S340 to determine the same object, and tracks the determined object of interest. .

The object recognition apparatus 120 transmits the surrounding image information including the tracked object of interest to the monitoring apparatus 130 to be displayed (S360).

In FIG. 3, steps S310 to S360 are described as being sequentially executed. However, this is merely illustrative of the technical idea of an embodiment of the present invention, and the general knowledge in the technical field to which an embodiment of the present invention belongs. Those having a variety of modifications and variations may be applicable by changing the order described in FIG. 3 or executing one or more steps of steps S310 to S360 in parallel without departing from the essential characteristics of an embodiment of the present invention. 3 is not limited to the time series order.

4 is an exemplary diagram for describing an operation of generating a recognition reference image according to the present embodiment.

The calibration chart is pre-installed on a wall located at a predetermined distance using the image capturing apparatus 110, and the object recognition apparatus 120 captures an object existing within a predetermined distance based on the captured recognition reference image. Recognize. For example, as shown in (a) of FIG. 4, a calibration chart such as (b) is photographed on a wall located at a distance of 5 m using the image photographing apparatus 110, and the object recognition apparatus 120 is photographed. Recognition Measures whether an object existing within a 5 m distance from the reference image can be recognized. Here, the object recognizing apparatus 120 measures the resolution by checking the number of pixels of the image corresponding to the effective angle of view (eg, 60 degrees) in the recognition reference image, and measures the object within a predetermined distance based on the resolution. It can also be set based on whether it can be recognized.

5 is an exemplary diagram for describing an operation of generating a reference image for correcting radiation distortion according to the present embodiment.

As shown in FIG. 5, a ground reference image is generated by photographing a surface of a grid chart installed using the image photographing apparatus 110 installed at a predetermined height. Here, the image capturing apparatus 110 is preferably installed at 1 m to 1.2 m corresponding to the height of the actual vehicle, but is not necessarily limited thereto.

Meanwhile, the object recognizing apparatus 120 corrects a bonding error of the captured image based on the bonding reference image. Here, the description of the bonding error will be described using the plurality of captured images shown in FIG. 6.

6 is an exemplary view for explaining a captured image of a vehicle according to the present embodiment.

As shown in Figure 6, using the four cameras attached to the front, rear, left and right of the vehicle to shoot the ground on which the grid chart is installed, the object recognition device 120 is photographed front image 610, The rear image 620, the right image 630, and the left image 640 are compared with previously stored bonding reference images to correct the bonding error. For example, the object recognizing apparatus 120 may include the left region of the front image 610 and the right region of the left image 640, the left region of the left image 640, and the rear image 620 based on the previously stored joint reference image. The right region, the left region of the rear image 620 and the right region of the right image 630, the left region of the right image 630 and the right region of the front image 610, the portion where two or more photographed images are joined The junction error can be corrected. Here, the camera attached to the vehicle is described as only four, but is not necessarily limited thereto, and the number of cameras may be increased or decreased as long as it can record 360 degrees around the vehicle.

7 is an exemplary view for explaining a method of detecting an object by the object recognition apparatus using the top view image according to the present embodiment.

As illustrated in FIG. 7, (a) refers to an image in which one vehicle 702 exists and the first object 710 and the second object 720 are positioned around the vehicle.

(b) refers to an image synthesized into a top view image by correcting a captured image photographed using a plurality of cameras attached to the vehicle 702. Here, the synthesized image is preferably synthesized in a time series order such as an image synthesized at t time and a synthesized image for t + 1 time. Here, the first object 710 is tracked by separating the composite image for t time into the foreground and the background, and the second object 720 is tracked by separating the composite image for t + 1 into the foreground and the background.

(c) generates the surrounding image information about the surroundings of the vehicle by matching the composite image for t time and the composite image for t + 1 time. Here, the surrounding image information may represent a plurality of objects moving around the vehicle by continuously matching the synthesized image generated in time series.

(d) notifies that the object exists in the captured image corresponding to the left side of the vehicle 702, and displays the first object 710 and the second object 720 in a square display on the image corresponding to the left side. Highlight and display.

The foregoing description is merely illustrative of the technical idea of the present embodiment, and various modifications and changes may be made to those skilled in the art without departing from the essential characteristics of the embodiments. Therefore, the present embodiments are to be construed as illustrative rather than restrictive, and the scope of the technical idea of the present embodiment is not limited by these embodiments. The scope of protection of the present embodiment should be construed according to the following claims, and all technical ideas within the scope of equivalents thereof should be construed as being included in the scope of the present invention.

As described above, the present embodiment is applied to the field to increase the recognition distance and resolution of the surrounding image of the vehicle, and simultaneously recognizes the surrounding image and the moving object (eg, a pedestrian, a vehicle, a motorcycle, and an animal). It is a useful invention that produces the effect of providing a more accurate alarm on collision avoidance.

110: image recording device 120: object recognition device
130: monitoring device 210: image processing unit
220: image acquisition unit 230: preprocessor
240: image synthesizer 250: object recognition unit
260: foreground background separator 270: object of interest tracking unit
280: surrounding image information generation unit

Claims (13)

An apparatus for recognizing an object of interest in a surrounding image of a vehicle by using a plurality of cameras installed in a vehicle,
An image acquisition unit for acquiring photographed images captured by the plurality of cameras;
A preprocessing unit generating a corrected image by correcting a radiation distortion and a bonding error of the photographed image based on the image selection information inputted through the manipulation;
An image synthesizer configured to generate the corrected image as one composite image corresponding to the image selection information;
A foreground background separator configured to determine a foreground pixel corresponding to a moving object by using a background pixel designated for a background without movement according to a predetermined criterion among the pixels in the composite image, and generate a foreground image including the foreground pixel;
An object of interest tracking unit that tracks the object of interest using the foreground image; And
Output unit for outputting the surrounding image result information including the tracked object of interest
Object recognition apparatus comprising a. The method according to claim 1,
The pre-
And correcting the radiation distortion and the joint error based on at least one of the image selection information selected from a panoramic view, a corner view, and a top view. 3. The method of claim 2,
The pre-
Correcting the radiation distortion of the photographed image in which the image is expanded or contracted radially from the center of the photographed image by the performance of a wide-angle camera having a wide viewing angle or external environmental factors such as the atmosphere and the sun. Object recognition device. The method of claim 3, wherein
The pre-
And correcting the radiation distortion using at least one of statistical correction, histogram matching correction, and drop line correction for a predetermined row or column in the photographed image. 3. The method of claim 2,
The pre-
And correcting the splicing error so as to accurately connect the respective captured images based on the previously stored splicing reference image corresponding to the image selection information. 6. The method of claim 5,
The junction reference image is,
And a reference image generated by photographing a grid chart pre-installed on the floor of the space where the vehicle is located. The method according to claim 1,
Wherein the image synthesizer comprises:
And reconstructing the corrected image into one of a panoramic view, a corner view, and a top view. The method according to claim 1,
The foreground background separator,
Object recognition, characterized in that the foreground pixel is separated using the background pixel designated according to a Gaussian model, and the foreground pixel is generated using a weight value of the pixel. Device. The method according to claim 1,
The object of interest tracking unit,
And tracking the trajectory of the object of interest using respective coordinate values corresponding to the foreground image. The method of claim 9,
The object of interest tracking unit,
Comparing at least one or more of color information, a histogram, and a shape of the object in the foreground image, the object of interest in which the coordinate value is located within a predetermined range may be determined to determine the object of interest. And an object recognition device for tracking the coordinate value. In the method of recognizing the object of interest around the vehicle,
An image acquisition process of acquiring a captured image photographed by a plurality of cameras attached to the vehicle;
A preprocessing step of correcting a radiation distortion of the photographed image and correcting a junction error of the photographed image based on a previously stored junction reference image;
An image synthesizing process of synthesizing a plurality of the corrected captured images into one image selected from a panoramic view, a corner view, and a top view;
A foreground background separation process of determining a foreground pixel by using a background pixel designated according to a predetermined criterion among pixels in the synthesized image, and generating a foreground image including the foreground pixel;
An object of interest tracking process of tracking the object of interest using the foreground image; And
An output process of outputting surrounding image result information including the tracked object of interest
Object recognition method comprising a. The method of claim 11,
The foreground background separation process,
And separating the foreground pixel by using the background pixel designated according to a Gaussian model, and generating the foreground pixel by using a weight value of the pixel. The method of claim 11,
The object tracking process of interest,
And identifying and tracking the object of interest by comparing at least one or more information among color information, histogram, and shape of the object moving in the foreground image.

Images