KR101465933B1 - Detecting method for color object in image, detecting apparatus for color object in image and detecting method for a plurality of color object in image - Google Patents

Abstract

A method (100) for detecting a color object from an image comprises the steps of: receiving, by an image processing apparatus, an input of a source image (110); normalizing, by the image processing apparatus, the source image into an RGB image (120); converting the normalized RGB image into a YCbCgCr image to generate a first temporary image with an object detected (130); converting the normalized RGB image into an HSV image to generate a second temporary image with the object detected (140); and combining the first and second temporary images to finally detect the object (150).

Description

TECHNICAL FIELD [0001] The present invention relates to a method of detecting a color object in an image, a method of detecting a color object in an image, and a method of detecting a plurality of color objects in an image OF COLOR OBJECT IN IMAGE}

The techniques described below relate to a method and apparatus for detecting color objects in an image.

Techniques for detecting objects in digital images have been used in many applications such as intelligent traffic systems, surveillance cameras, and interactions between detected objects and computers. In addition, due to the introduction of various smart devices with high performance, researches for detecting and recognizing objects through a camera and for interacting with various smart devices using recognized objects are being actively carried out.

On the other hand, the image processing algorithm using color for object detection in the image mainly uses the feature information represented by the three-dimensional color information of R (red), G (green), and B (blue). The main reason that color information is used is that it provides relatively more information than gray or binary image, and it is possible to extract data easily and speed of image processing is fast.

Korean Laid-Open Patent No. 2013-0121344 (2013.11.06) Korean Laid-Open Patent No. 2013-0064556 (2013.06.18)

Detection objects and computer-to-computer interaction systems include Microsoft's Kinect and Sony's PlayStation Move, but there are problems with high system cost and specific applications.

 Although there are a lot of methods to acquire and track moving objects using only cameras without additional equipment, there are limitations in tracking multiple objects in a complex environment and various lighting and analyzing them in real time. In addition, the conventional multi-object recognition system has a problem in that it requires a lot of image processing time because of searching the entire input image for object detection.

Conventional YCbCr model for color object detection in image extracts only color difference information of B (blue) and R (red) except for brightness value for RGB three-dimensional color information expressed in image, It is difficult to carry out the operation.

The technique described below is to detect an object in a source image using a YCbCgCr image and an HSV image for a source image, respectively. The YCbCgCr image uses a new YCbCgCr color model by adding G (green) color difference information to the YCbCr color model.

The solutions to the technical problems described below are not limited to those mentioned above, and other solutions not mentioned can be clearly understood by those skilled in the art from the following description.

A method of detecting a color object in an image includes the steps of receiving a source image by an image processing apparatus, RGB normalizing a source image, converting a normalized RGB image into a YCbCgCr image, and generating a first temporary image Converting the normalized RGB image into an HSV image to generate a second temporary image in which the object is detected, and finally detecting the object by combining the first temporary image and the second temporary image.

In the normalizing step, the image processing apparatus normalizes the value of each component by dividing the value of each of the R, G, and B color components by the sum of the values of the three RGB components in the source image.

The step of generating the first temporary image may include converting the normalized RGB image into a YCbCgCr image, extracting an object having a threshold value or more in each color difference channel of Cb, Cg, and Cr in the YCbCgCr image, extracting the object, And Cr; performing binarization on at least one of the images; and performing combining for each image when the binarized images are a plurality of images.

The step of generating the second temporary image may include converting the normalized RGB color image into an HSV image, extracting a pixel having a first threshold value or more and a pixel having a second threshold value or less in the S channel of the HSV image, A step of extracting a pixel having a reference threshold value or more from the channel and a step of generating a second temporary image by combining the image extracted from the S channel and the image extracted from the V channel.

An apparatus for detecting a color object in an image detects an object using a threshold value in a YCbCgCr image and an HSV image with respect to a source image, and detects a color object in a source image by combining an YCbCgCr image and an HSV image, And a computing device for detecting at least one color object in the source image using a memory device and software for storing the software.

The arithmetic unit converts the normalized RGB image into a YCbCgCr image using the software to convert the first temporary image and the normalized RGB image into an HSV image, And finally detects the object by combining the first temporary image and the second temporary image.

The arithmetic unit converts the normalized RGB image into a YCbCgCr image, extracts an object having a reference threshold value or more from each color difference channel of Cb, Cg, and Cr in the YCbCgCr image, binarizes the extracted image, .

The arithmetic unit converts the normalized RGB color image into an HSV image, extracts a pixel having a first threshold value or more and a pixel having a second threshold value or less in the S channel of the HSV image, And combines the image extracted from the S channel and the image extracted from the V channel to generate a second temporary image.

There is provided a method of detecting a plurality of color objects in an image, the method comprising: receiving a source image including a plurality of target objects by the image processing apparatus; detecting, for each of the R color value, the G color value, and the B color value of the source image, Normalizing the source image by dividing the value of each component by the sum of the values of the three RGB components, and converting the normalized RGB image into a YCbCgCr image to convert the first temporary image and the normalized RGB image, HSV image to generate a second temporary image in which the object is detected, and detecting the target object by combining the first temporary image and the second temporary image.

The technique described below uses the YCbCgCr image and the HSV image, respectively, to detect objects efficiently and has a low computational complexity. Therefore, the technique described below is suitable for detecting a plurality of objects from a source image using popular equipment such as a mobile terminal.

The effects of the techniques described below are not limited to those mentioned above, and other effects not mentioned can be clearly understood by those skilled in the art from the following description.

1 is an example of a flowchart for a method of detecting a color object in an image.
2 is another example of a flowchart for a method of detecting a color object in an image.
3 is an example of a process for a method of detecting a color object in an image.
Figure 4 is an example of a block diagram for an apparatus for detecting color objects in an image.
5 is an example of a flowchart for a method of detecting a plurality of color objects in an image.
6 shows an example of detecting a plurality of color objects in an image.

The following description is intended to illustrate and describe specific embodiments in the drawings, since various changes may be made and the embodiments may have various embodiments. However, it should be understood that the following description does not limit the specific embodiments, but includes all changes, equivalents, and alternatives falling within the spirit and scope of the following description.

The terms first, second, A, B, etc., may be used to describe various components, but the components are not limited by the terms, but may be used to distinguish one component from another . For example, without departing from the scope of the following description, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

As used herein, the singular " include "should be understood to include a plurality of representations unless the context clearly dictates otherwise, and the terms" comprises & , Parts or combinations thereof, and does not preclude the presence or addition of one or more other features, integers, steps, components, components, or combinations thereof.

Before describing the drawings in detail, it is to be clarified that the division of constituent parts in this specification is merely a division by main functions of each constituent part. That is, two or more constituent parts to be described below may be combined into one constituent part, or one constituent part may be divided into two or more functions according to functions that are more subdivided. In addition, each of the constituent units described below may additionally perform some or all of the functions of other constituent units in addition to the main functions of the constituent units themselves, and that some of the main functions, And may be carried out in a dedicated manner. Accordingly, the presence or absence of each component described in the present specification will be interpreted as a function. For this reason, a method 100 for detecting a color object in an image of a technique described below and an apparatus for detecting a color object in an image 500 can be different from the corresponding drawings within the scope of achieving the object of the following description.

Also, in performing a method or an operation method, each of the processes constituting the method may take place differently from the stated order unless clearly specified in the context. That is, each process may occur in the same order as described, may be performed substantially concurrently, or may be performed in the opposite order.

The technique described below relates to a technique for detecting a specific color object in a digital image. As described above, the technique of detecting an object in an image is used in various application fields. The technology described below is intended to quickly detect an object in an image using a relatively simple and inexpensive device such as a camera built in a smart phone, a camera built in a smart TV, or a camera (web cam) connected to a computer. The techniques described below are useful for detecting a plurality of objects in an image using a low complexity object detection method. A method 100 for detecting a color object in an image described below and an apparatus 500 for detecting a color object in an image may detect an object in an image or may detect a plurality of multiple objects in an image .

A major feature of the technique described below is that a new YCbCgCr model and HSV model generated by adding G (green) color difference information to a conventional YCbCr color model are simultaneously used.

Hereinafter, a method 100 for detecting a color object in an image and an apparatus 500 for detecting a color object in an image will be described in detail with reference to the drawings.

Figure 1 is an example of a flowchart for a method 100 for detecting a color object in an image. A method 100 for detecting a color object in an image includes a step 110 of inputting a source image by an image processing apparatus, a step 120 of normalizing a source image to RGB, a step of converting a normalized RGB image into a YCbCgCr image, A step 130 of generating a first temporary image, a step 140 of converting a normalized RGB image into an HSV image to generate a second temporary image in which an object is detected, and a step 140 of combining the first temporary image and the second temporary image And finally detecting (150) the object.

A method 100 for detecting a color object in an image is a method for generating a first temporary image and a second temporary image by performing image processing on a source image and combining the generated two images to finally detect an object to be.

The source image means a color image first input through an image acquisition device of the image processing apparatus. An image processing apparatus refers to a device that detects a specific object in an image received through a camera, an image stored in a memory, or an image received through a network. The image processing apparatus includes a video device with a built-in camera, a computer device, a mobile device such as a smart phone, and a smart TV with a built-in camera. The image acquisition device includes a camera that acquires an external image in real time, an interface device that acquires an image stored in the memory, and a communication module that receives image data through a network.

The step 110 of receiving the source image receives the image directly from the camera, retrieves the image stored in the memory, or receives the image received through the network or the like.

The RGB normalization of the source image 120 is a process of normalizing the input RGB color image. Since the input color image is composed of RGB models, since the general RGB color model is sensitive to the change in illumination and it is difficult to perform accurate calculation, it is difficult to perform the normal operation by dividing the component values of r, g and b by the sum of the three component values. Go through the process. The normalization process is shown in Equation 1 below.

In the above equation, R, G, and B are RGB component values that are normalized, respectively.

2 is another example of a flowchart for a method 100 for detecting a color object in an image. FIG. 2 illustrates a step 130 of transforming a normalized RGB image into a YCbCgCr image to generate a first temporary image in which an object is detected 130 and a second temporary image in which an object is detected by converting the normalized RGB image into an HSV image Step 140 will be described in more detail. The step 130 of generating the first temporary image and the step 140 of generating the second temporary image may be performed sequentially, or may be performed in parallel, irrespective of the order.

First, a step 130 of generating a first temporary image will be described. The normalized RGB color image is converted into a YCbCgCr image (131). The YCbCgCr image is generated as a YCbCgCr model in which G (green) color difference information is added to the YCbCr color model.

The process of extracting the chrominance information Cb, Cg and Cr for each of B (blue), G (green) and R (red) excluding the brightness component Y in the RGB color model is represented by the following Equations 2 to 4 same.

In the above equation, R, G, and B represent the respective component values for r (red), g (green), and b (blue) in which the normalization process is performed in the RGB color model.

It is preferable to extract a pixel (a color having a larger color) of a predetermined threshold value or more in the color difference channels of Cb, Cg, and Cr extracted from the above expression (132). For example, a value of 150 or more can be extracted and used. However, the threshold setting can be changed in consideration of various environmental variables such as system performance, image information of an object to be detected, illuminance at which an image is acquired, and the like. The threshold used in the process of generating the first temporary image is called the YCbCgCr threshold value. The YCbCgCr threshold value may be different depending on each chrominance channel.

In this process, the maximum value of the specific color difference information among the color difference channels of Cb, Cg and Cr may be extracted. For example, if the chrominance information of the target object to be extracted is known, only a specific channel capable of detecting the corresponding object may be used.

In order to detect a specific object in the YCbCgCr image, it is preferable to perform binarization on the YCbCgCr image (133). Binarization simplifies the object of the image and improves the efficiency of object detection.

After the corresponding color is detected (132) in each chrominance channel, it is divided into object regions, and a binarization step is performed on the object region with respect to the background. The binarization process is shown in Equation (5) below.

는 이진화를 위한 임계치를 의미한다. 즉, Cg, Cg 및 Cr 각각에 대하여 추출한 색차 정보가

이상인 경우는 255인 값(검은색)을 갖고,

미만인 경우는 0인 값(흰색)을 갖게 된다. 이진화를 위한 임계치

는 Cg, Cg 및 Cr을 추출하기 위해 사용하였던 제1 영상 임계치와는 다른 것이다.
"Region of Object" means an area where a specific object is located in the image.

Denotes a threshold value for binarization. That is, the color difference information extracted for each of Cg, Cg, and Cr

(Black) in a case where the number of pixels is equal to or larger than 255,

(White), the value of 0 is obtained. Threshold for binarization

Is different from the first image threshold used to extract Cg, Cg and Cr.

The first temporal image may be generated by binarizing all the images of the Cg, Cg, and Cr chrominance channel, and the first temporal image may be generated by binarizing only the image of the Cg, Cg, and Cr chrominance channels, It might be. If the chrominance information of the target object to be detected is known in advance, only the image of the specific channel can be binarized.

Furthermore, the binarization target may be selected by analyzing the image of each chrominance channel or by analyzing the binarized image of each chrominance channel. However, basically, it is preferable that a method of binarizing all the images of three chrominance channels and combining them.

After performing the binarization step 150, it is preferable to perform a step of removing unnecessary objects or noise or the like for more accurate detection of the object region of interest (134). For example, an erosion operation according to the following expression (6) and an expansion operation (Erosion) according to the following expression (7) can be performed.

B는 수학식 6과 같이 정의할 수 있다. 침식연산은 객체 내부의 돌출부는 감소하고 외부의 돌출부는 증가시켜서 물체 내부나 배경에서 발생한 구멍과 같은 공간을 채우거나 짧게 끊어진 영역을 연결하는데 주로 사용되며, 이진 영상에서 침식연산은 입력 화소가 균일한 곳에서는 변화가 없으나 흑백 화소가 같이 있는 영역에서 수행된다.
When the erosion operation assumes that A and B are a set of pixels, A

B can be defined as shown in Equation (6). The erosion operation is mainly used to fill a space such as a hole formed in the inside of the object or the background or to connect a short-cut region by decreasing the protrusion inside the object and increasing the protrusion of the outside. In the binary image, But there is no change in the area where black and white pixels are present.

B는 수학식 7과 같이 정의할 수 있다. 여기서

는 형태소 B를 이동한 결과로, 이 중 A집합에 완전하게 포함되는

의 집합이 침식으로 일어난 결과를 의미한다. 즉, B를 A위로 이동하면서 B가 완전하게 포함되는 장소를 찾은 뒤 각 장소에서 원점에 해당하는 점을 모아 만든 집합을 팽창연산이라 정의할 수 있다.
Assuming that A and B are a set of pixels,

B can be defined as Equation (7). here

Is the result of moving the morpheme B,

The result of erosion. In other words, a set that is obtained by finding the place where B is completely contained while moving B over A and gathering the points corresponding to the origin in each place can be defined as an expansion operation.

A first temporal image for the source image can be prepared by extracting an object having a threshold value or more from the YCbCgCr color model for each color difference information and performing a binarization process and a noise removal process. The first temporary image generation step 130 corresponds to a process of converting a source image into a YCbCgCr image and detecting a specific object in the image.

A process 140 for generating a second temporary image using the source image will also be described. The step 140 of generating a second temporary image 140 includes converting a normalized RGB color image into an HSV image 141, extracting a pixel having a first threshold value or more and a pixel having a second threshold value or less in the S channel of the HSV image In operation 142, a step 143 extracts a pixel having a reference threshold value or more from the V channel of the HSV image and a step 144 generates a second temporary image by combining the image extracted from the S channel and the image extracted from the V channel. .

In the process 140 of generating the second temporary image, only the S (Saturation) channel and the V (Value) channel are used except for the H-sensitive channel. Of course, the information of the H channel may be used in some cases. In FIG. 2, it is described that the maximum area and the minimum area of the S channel are extracted (142) in the HSV, which means a process of extracting a pixel having a first threshold value or more and a pixel having a second threshold value or less in the S channel. The first threshold value and the second threshold value used in the S channel are called an S channel threshold value. In the V-channel, a maximum area exceeding a specific reference threshold value is extracted. The threshold used in the V-channel is called the V-channel threshold.

Like the above-described YCbCgCr threshold value, the S-channel threshold value and the V-channel threshold value can also be changed according to the environment variable in which the object is detected. For example, the first threshold value of the S channel may be 205, the second threshold value of the S channel may be 35, and the threshold value of the V channel may be 190.

The second temporary image is generated by performing an AND operation on the image extracted from the S channel and the image extracted from the V channel (144).

Then, the first temporary image and the second temporary image are combined (AND) (150). An object detected through the first temporary image and an object detected through the second temporary image are combined to enhance the object detection effect. And the target object is detected from the source image based on the combined final image.

3 is an example of a process for a method of detecting a color object in an image.

(a) is a screen for receiving a background image and a finger image with a marker attached thereto, and (b) is a RGB normalized image of the input image.

(c) is an example of a screen for converting a normalized RGB image into a first temporary image. (c), (c-1) converting the normalized RGB image into a YCbCgCr image, (c-2) extracting a maximum area exceeding a predetermined threshold value in each channel of Cb, Cg and Cr A process (c-3) of binarizing the image of the channel and a process (c-4) of removing noise from the binarized image.

(d) is an example of a screen for converting a normalized RGB image into a second temporary image. (d) is a process (d-1) of converting the normalized RGB image into an HSV image, (d-2) extracting a maximum area and a minimum area in the S channel among the HSVs, (D-3) of extracting the S-channel image and the V-channel image and (d-4) combining the S-channel image and the V-channel image.

(e) is a process of finally combining the first temporary image and the second temporary image, and (f) is an example of detecting an object on the screen in which the first temporary image and the second temporary image are combined.

4 is an example of a block diagram for an apparatus 500 for detecting a color object in an image.

An apparatus 500 for detecting a color object in an image detects an object using a threshold value in each of a camera 510 that receives a source image, a YCbCgCr image and a HSV image of a source image, a YCbCgCr image that detects an object, A memory device 520 for storing software that combines the HSV images to detect color objects in the source image, and a computing device 530 for detecting at least one color object in the source image using software.

FIG. 4 shows an example of receiving a source image through a camera. An apparatus 500 for detecting a color object in an image is a form in which software for detecting a color object in a source image is stored in a memory device 520. The device 500 for detecting a color object in an image may be a device such as a smart phone equipped with an app for object detection, or the embedded software for object detection may be an object detection only device stored in a flash memory or ROM.

The object detection software 525 basically corresponds to a program that performs a method 100 for detecting a color object in the above-described image.

The computing device 530 is a configuration for driving the object detection software 525 stored in the memory device 520. The apparatus 500 for detecting a color object in an image may include a display device 550 for outputting a source image input to the camera 510 and / or a target object detected from the source image. And an interface device 540 for receiving information about the target object from the user. The user may input the above-described YCbCgCr threshold value, H channel threshold value, V channel threshold value, and the like through the interface device 540. The display device 550 and the interface device 540 may be various devices used in a computer or a mobile terminal device.

The arithmetic operation unit 530 converts the normalized RGB image into a YCbCgCr image using the software to convert the first temporary image and the normalized RGB image into an HSV image, The second temporary image is generated, and the object is finally detected by combining the first temporary image and the second temporary image.

The arithmetic operation unit 530 normalizes the source image by dividing the value of each component by the sum of the values of the three RGB components for each of the R color value, the G color value, and the B color value of the source image, as shown in Equation (1) .

The arithmetic unit 530 converts the normalized RGB image into a YCbCgCr image using Equations (2) to (4). The arithmetic unit 530 then extracts an object having a reference threshold value or more in each of the chrominance channels Cb, Cg, and Cr in the YCbCgCr image, and binarizes the extracted image to generate a first temporary image.

The arithmetic unit 530 converts the normalized RGB color image into an HSV image, extracts a pixel having a first threshold value or more and a pixel having a second threshold value or less in the S channel of the HSV image, And a second temporary image is generated by combining the image extracted from the S channel and the image extracted from the V channel.

The computing device 530 finally combines the first temporary image and the second temporary image to detect the target object.

5 is an example of a flowchart for a method 200 for detecting a plurality of color objects in an image. 6 shows an example in which a plurality of color objects are detected in an image.

A method 200 for detecting a plurality of color objects in an image includes a step 210 of receiving a source image from an image processing apparatus, a step 220 of selecting a plurality of target object regions from a source image, A step 230 of transforming the normalized RGB image into a YCbCgCr image to generate a first temporary image in which a target object is detected 240, transforming the normalized RGB image into an HSV image, An image generating step 250 and a step 260 of combining the first temporary image and the second temporary image to finally detect the target object.

A method 200 for detecting a plurality of color objects in an image has the same process as the method 100 for detecting a color object in the image described above. However, it is intended to clearly show that a plurality of objects can be detected from a source image. Therefore, the same configuration as the method 100 for detecting a color object in an image will be briefly described or a description thereof will be omitted.

It detects a specific object specified by the user among many objects existing in the source image. The object to be detected is called the target object. There may be many objects in the source image input in operation 210. Referring to FIG. 6, there are color markers of different colors at five fingertips. The user can select a target object which is a color object to be detected from the input source image. To be more precise, the region in which the target object is located in the source image is selected (220).

A method of selecting a target object region from a source image by a user may use a method of designating the target object region as a region of interest. For example, if the display device of the mobile terminal is a touch panel, the target object area can be touched to designate the area as a region of interest. Alternatively, a user may designate a plurality of target object regions as a region of interest by using a cursor or the like. Furthermore, a plurality of target object regions may be extracted from the source image using an image processing technique.

Thereafter, the image processing apparatus performs RGB normalization on the source image (230), YCbCgCr transforms the normalized RGB image to generate a first temporary image (240), and HSV-transforms the normalized RGB image to generate a second temporary image 250). Finally, the first temporary image and the second temporary image are combined to detect the target object in the source image (260). Fig. 6 shows an example of detecting five color target objects.

It should be noted that the present embodiment and the drawings attached hereto are only a part of the technical idea included in the above-described technology, and those skilled in the art will readily understand the technical ideas included in the above- It is to be understood that both variations and specific embodiments which can be deduced are included in the scope of the above-mentioned technical scope.

500: a device for detecting a color object in an image
510: camera 520: memory
525: Object detection software 530:
540: Interface device 550: Display device

Claims (16)

A method of detecting a color object in an image by an image processing apparatus,
Receiving the source image from the image processing apparatus;
Performing RGB normalization on the source image;
Converting the normalized RGB image into a YCbCgCr image and generating a first temporary image in which an object is detected in the YCbCgCr image;
Converting the normalized RGB image into an HSV image and generating a second temporary image in which the object is detected in the HSV image; And
And finally detecting the object by combining the first temporary image and the second temporary image,
Wherein the step of generating the first temporary image and the step of generating the second temporary image are sequentially performed irrespective of the order, or a method of detecting a color object in an image performed in parallel. The method according to claim 1,
The normalizing step
Wherein the image processing apparatus normalizes the value of each component by the sum of the values of the three RGB components for each of the R color value, the G color value, and the B color value in the source image. The method according to claim 1,
The step of generating the first temporary image
Converting the normalized RGB image into a YCbCgCr image;
Extracting an object having a threshold value or more in each color difference channel of Cb, Cg and Cr in the YCbCgCr image;
Extracting the object, performing binarization on at least one image of Cb, Cg, and Cr; And
And performing combining for each image when the binarized images are a plurality of images. The method of claim 3,
The converting step
Wherein the image processing apparatus extracts B color difference information (Cb), G color difference information (Cg), and R color difference information (Cr) from the RGB image as shown in the following equation.

,

,

(R, G, and B in the above formula are the values of the RGB component subjected to the normalization, respectively) 5. The method of claim 4,
The step of extracting an object having a threshold value or more
And extracting only the Cb, Cg, and Cr values that are greater than or equal to a reference threshold value. The method according to claim 1,
The step of generating the second temporary image
Converting the normalized RGB image into an HSV image;
Extracting a pixel having a first threshold value or more and a pixel having a second threshold value or less in an S channel of the HSV image;
Extracting a pixel in the V channel of the HSV image that is greater than or equal to a reference threshold value; And
And combining the extracted image in the S channel and the extracted image in the V channel to generate the second temporary image. A camera for receiving a source image;
A memory device for storing software for detecting an object using a threshold value in a YCbCgCr image and an HSV image for the source image, and detecting a color object in the source image by combining the YCbCgCr image and the HSV image, ; And
And an arithmetic unit for detecting at least one color object in the source image using the software. 8. The method of claim 7,
The computing device
Performing RGB normalization on the source image using the software,
Converting the normalized RGB image into a YCbCgCr image and converting the first temporary image and the normalized RGB image into an HSV image by detecting the object to generate a second temporary image in which the object is detected, A device for detecting a color object in an image that finally detects an object by combining temporary images. 9. The method of claim 8,
The computing device
Wherein the color image is normalized by dividing a value of each component by the sum of values of three RGB components for each of the R color value, the G color value, and the B color value of the source image. 9. The method of claim 8,
The computing device
Extracting a normalized RGB image into the YCbCgCr image, extracting an object having a reference threshold value or more in each color difference channel of Cb, Cg, and Cr in the YCbCgCr image, performing binarization on the extracted image, An apparatus for detecting a color object in an image generating a temporary image. 11. The method of claim 10,
The computing device
A device for detecting a color object in an image for converting the RGB image into the YCbCgCr image using the following equation with respect to B color difference information Cb, G color difference information Cg, and R color difference information Cr in the RGB image, .

(R, G, and B in the above formula are the values of the RGB component subjected to the normalization, respectively) 9. The method of claim 8,
The computing device
Extracting a pixel having a first threshold value or more and a pixel having a second threshold value or less from the S channel of the HSV image and extracting a pixel having a reference threshold value or more from the V channel of the HSV image; And combining the image extracted in the S channel and the image extracted in the V channel to generate the second temporary image. A method of detecting a plurality of color objects in an image by an image processing apparatus,
Receiving a source image including a plurality of target objects by the image processing apparatus;
Normalizing the source image by dividing the value of each component by the sum of the values of three RGB components for each of the R color value, the G color value, and the B color value of the source image; And
The image processing apparatus converts the normalized RGB image into a YCbCgCr image, converts the first temporary image and the normalized RGB image, which have detected the object in the YCbCgCr image, into an HSV image, and detects an object in the HSV image Generating a second temporary image, and combining the first temporary image and the second temporary image to detect the target object. 14. The method of claim 13,
In the detecting step
The image processing apparatus converts the normalized RGB image into the YCbCgCr image, extracts an object having a reference threshold value or more from each of the color difference channels Cb, Cg, and Cr in the YCbCgCr image, and binarizes the extracted image And generating a first temporary image by performing a predetermined process on the detected first color image. 14. The method of claim 13,
In the detecting step
Wherein the image processing apparatus converts the normalized RGB image into an HSV image, extracts a pixel having a first threshold value or more and a pixel less than a second threshold value in the S channel of the HSV image, And combining the image extracted in the S channel and the image extracted in the V channel to generate the second temporary image. 14. The method of claim 13,
Further comprising receiving information for selecting a target object region in which a plurality of target objects are located in the source image, before the step of detecting the target object,
Wherein the detecting of the target object comprises detecting only the object located in the selected target object region.

Images