KR20040080293A - Realtime Object Extraction System and Method - Google Patents

Abstract

PURPOSE: A system and a method for extracting an object in real time are provided to separate a background region and an object region from an image using an active infrared camera. CONSTITUTION: An infrared reflection object(10) is used as background. An active infrared camera(20) has an infrared LED(Light Emitting Diode), a general camera and an infrared camera. The infrared LED irradiates infrared illumination to the infrared reflection object and an object before the infrared reflection object. The general camera photographs the infrared reflection object and the object in a state in which the infrared illumination is irradiated and outputs the photographed images. The infrared camera focuses the infrared reflection object and the object in the state in which the infrared illumination is irradiated through a lens, and outputs a clear or dark gray image according as the quantity of an infrared ray inputted through the lens is great or little. A computer(30) divides the image outputted from the infrared camera into an object part and a background part, and synthesizes a graphic image with the background part.

Description

Real-time object extraction system and method

The present invention relates to a real-time object extraction system and method, and more particularly, to a system comprising a background having an infrared reflection characteristic and an active infrared camera detecting the same, and a method of synthesizing an image taken by the active infrared camera in real time. It is about.

The main purpose of real-time image processing is to find out the shape and area of an object in order to separate the object and the background from the image input from the video camera.

The separated background part or object part obtained as a result of this process is used to create a new image by synthesizing a computer graphic image, etc., or the object-based compression algorithm proposed by MPEG-4.

Conventionally, in order to facilitate the separation process, the background is made of a single color that the object does not have, and the object and the background are separated from the image by using the fact that the background color is different from the object.

This separation method is commonly called Chroma Key, and an example of Chroma Key is that the announcer in TV weather forecast explains the weather in front of the graphic screen.

As described above, in the conventional chroma key, a single color background using a color not included in the object as a background color is used to easily separate the background from the object.

However, since the background can not be changed every time by painting a different color, once the background color is determined, the same color cannot be used for the object.

For this reason, Chromakey uses colors that are not used by many people, such as blue or green, as background colors.

In the process of separating objects, the smaller the change in background color, the easier it is to separate the background from the object. The color of the background is affected by the light that illuminates it.

In addition, even if the color of the background is constant, it is difficult to separate the background from the object because the background color of the image does not come out uniformly if the lighting used is not uniform.

Also, the shadow of an object also changes the background color, so it must be removed with proper lighting.

Therefore, in general, chroma key should be installed so that the screen is bright and uniform overall, and there is no shadow part.

As a result, in the conventional chroma key, there is a problem in that the brightness and the arrangement of the lighting are well adjusted so that the color of the background is uniform in the image, so there is a problem that the image of the non-uniform lighting cannot be made or used in a dark lighting environment.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to use an infrared reflector as a background instead of using a background color as in the prior art, and use an active infrared camera to display a background in an image. Infrared-key method that separates the area from the object area, unlike the conventional chroma key, there is no limit to the color that the object can have, and the illumination in the visible region does not need to be uniform, so that various bright and dark lighting configurations can be used. In addition, since there is no need to use a specific color for the background, not only can the background be composed of various colors, but also a white background can be used as a screen, and a real-time object extraction system and method can be used to give various effects to the image. To provide.

In order to achieve the above object, the present invention, the infrared reflector used as the background;

An infrared LED for illuminating infrared light on the infrared reflector and an object in front of the infrared reflector, a general camera which photographs and outputs an infrared reflector and an object while the infrared light is illuminated, and an infrared reflector when the infrared light is illuminated An active infrared camera composed of an infrared camera which grabs an object with a lens and outputs a bright or dark contrast image according to a large amount and a small amount of infrared rays inputted to the lens;

It is to provide a real-time object extraction system comprising a computer that separates the image output from the infrared camera into an object portion and a background portion, and synthesizes a graphic image on the background portion of the live image.

The present invention also provides a first step of taking an image with a general camera and an infrared camera in the state of illuminating the infrared light on the background and the object (object or person) of the infrared reflector;

A second step of distinguishing an object region from a background region by performing image processing for binarizing a contrast image output from the infrared camera based on a threshold;

Determining whether a general camera image or a graphic image is specified in the output image by determining whether each pixel in the contrast image of the infrared camera is larger or smaller than a predetermined threshold for each pixel;

It is to provide a real-time object extraction method comprising a fourth step of outputting a graphic image is synthesized in the background portion of the general camera image as an output image.

1 is a block diagram of a real-time object extraction system according to the present invention.

2A and 2B are schematic diagrams of the active infrared camera shown in FIG. 1.

3 is a flowchart illustrating a method of synthesizing a graphic image on a background portion of a live image using the present invention.

<Description of the symbols for the main parts of the drawings>

10: infrared reflector 20: active infrared camera

22: infrared LED 24a: half mirror

24b: mirror 26: normal camera

28: infrared camera 30: computer

40: object

Hereinafter, the configuration and operation of the present invention will be described with reference to the accompanying drawings.

The present invention comprises an infrared reflector 10 used as a background, an active infrared camera 20, and a computer 30 for processing an image received by the active infrared camera.

A schematic diagram of the overall infrared-key system can be shown in FIG. 1, where the active infrared camera 20 is a combination of an infrared LED 22 illuminating infrared light, an infrared camera 28, and a general camera 26. The camera system has a structure as shown in FIG. 2.

Figure 2a is a side view showing the internal structure, Figure 2b is a front view.

In FIG. 1, an object (object or person) 40 is positioned in front of a background made of the infrared reflector 10, and in this state, the infrared light is illuminated by the infrared LED 22 to illuminate the background and the object 40, and the scene is infrared. The camera 28 catches up.

The infrared camera 28 outputs a bright or dark gray image as the amount of infrared light input to the lens is large and small.

On the other hand, since the object or the person 40 is less reflected by the infrared reflector 10 than the infrared reflector 10, the image captured by the infrared camera 28 has a bright background area and the object area has a relatively large amount of reflected infrared light. It will come out dark.

If the image is subjected to binarization based on an appropriate threshold as shown in Equation 1, the object region and the background region can be distinguished.

here Is the contrast image output from the infrared camera 28, Denotes the x and y coordinates of the pixel, respectively.

Denotes a binarized output image, where an area indicated by 1 represents an object and an area indicated by 0 represents a background.

In order to use the above method for synthesizing graphic images as in chroma key, an image captured by the general camera 26 and an image captured by the infrared camera 28 must be photographed at the same time.

That is, the object and the background area separated from the image of the infrared camera 28 should be the same as the area of the general camera 26 so that there is no mismatch even when combining the two images.

Therefore, as shown in FIG. 2, a half mirror 24a and a mirror 24b are used in the active infrared camera 20 to arrange the same image in the two cameras 26 and 28.

The active infrared camera 20 of FIG. 2 is very economical because it can be made simply by attaching the infrared camera 28 and the infrared LED 22 to a general commercial camcorder.

Although there may be various methods for synthesizing the live image and the graphic image in the computer 30, one example of synthesizing the graphic image in the background portion of the live image is shown in a flowchart as shown in FIG. 3.

In FIG. 3, I (x, y) is an image input from a general camera, and G (x, y) is a graphic image to be synthesized (S30).

Whether the general camera input image I (x, y) is assigned to the output image O (x, y) by determining whether the input infrared camera image P (x, y) is larger or smaller than a predetermined threshold for each pixel. In operation S32, it is determined whether the graphic image G (x, y) is designated.

In this case, the threshold value is used as a criterion for distinguishing whether a pixel is a background or an object in an infrared image.

If P (x, y) is greater than the threshold value, more infrared rays are input, so this pixel becomes the background area that is an infrared reflector. If P (x, y) is less than the threshold value, an object that reflects less infrared light is used. Becomes a part.

Eventually, the output O (x, y) is an image obtained by synthesizing the graphic image G (x, y) on the background of I (x, y) (S34).

If you use Chroma Key to make a weather forecast on TV, the announcer explains that the hand points to a portion of the screen and looks away from the eye.

This means that you can see the external monitor where you can see the announcer himself composited with the graphic image. In this way, in Chroma Key, the user does not know what graphic image appears in the background. It is necessary.

In addition, when interacting with an image, such as pointing to a specific point by hand, as in weather forecasting, the point of the hand point is not known, so the contents of the resulting image are not known. The disadvantage is that the action is unnatural.

In the present invention, when the white infrared reflector is used as the background, the background can be used as a screen.

This allows you to display specific markers on the background or display the resulting graphic screen without affecting the separation of the announcer from the background in applications such as weather forecasts, so you can see and explain the background that the announcer points to. Can be more natural.

As described above, according to the infrared-key method of using an infrared reflector as a background and separating an background region and an object region from an image using an active infrared camera,

First, there is no limitation on the color that a detection object (object) can have, which is a disadvantage of a method of separating objects using a conventional chroma key.

Secondly, the illumination in the visible region does not need to be uniform, so that various lighting configurations, bright and dark, can be used, and in particular, even in the dark, objects can be distinguished in real time.

Third, since there is no need to use a specific color for the background, it is possible to configure a background of various colors.

Fourth, when using a white infrared reflector as a background can be used as a screen at the same time can display a variety of information on the background (screen) or give a variety of effects to the image.

Claims (5)

Infrared reflector used as background; An infrared LED for illuminating infrared light on the infrared reflector and an object in front of the infrared reflector, a general camera which photographs and outputs an infrared reflector and an object while the infrared light is illuminated, and an infrared reflector when the infrared light is illuminated An active infrared camera composed of an infrared camera which grabs an object with a lens and outputs a bright or dark contrast image according to a large amount and a small amount of infrared rays inputted to the lens; And a computer for separating the image output from the infrared camera into an object portion and a background portion, and composing a graphic image on the background portion of the live image. The real-time object extraction system according to claim 1, wherein the general camera and the infrared camera use a half mirror and a mirror to capture an image at the same time. The real-time object extraction system according to claim 1, wherein the infrared reflector is made white and used as a screen. A first step of capturing an image with a general camera and an infrared camera in the state of illuminating infrared light on the background and the object (object or person) of the infrared reflector; A second step of distinguishing an object region from a background region by performing image processing for binarizing a contrast image output from the infrared camera based on a threshold; Determining whether a general camera image or a graphic image is designated as an output image by determining whether the infrared image of the infrared camera is larger or smaller than a predetermined threshold for each pixel; And a fourth step in which a graphic image is synthesized on the background portion of the general camera image and output as an output image. The method of claim 4, wherein the object region and the background region are divided by Equation 2 in the second step. here Is the contrast image of the infrared camera Are the x and y coordinates of the pixel, respectively. Is a binary output image, the area marked with 1 is the object and the area marked with 0 is the background.