KR101557059B1 - Image monitoring system and method for removing needle - Google Patents

Abstract

The present invention relates to an image monitoring system and method for a foot and a foot so as to prevent a problem caused by incomplete footing by capturing an image of a procedure area to confirm the position and removal of the hand, An upper photographing unit configured to take a photograph of a patient's front face portion and an hourly needles to confirm the position of the acupuncture needle and how to perform a stroke, a fixed form on the side of the body where the patient is positioned for the hour hand, A first and a second side radiographing part for radiating the body surface of the patient's side part and the hourly needles, and an image picked up by the upper part radiographing part and the first and second side radiographs, And a tracking result display unit for displaying a procedure position and whether or not to perform a stroke based on the RGB color information of the part or the intensity of the needle head part will be.

Description

TECHNICAL FIELD [0001] The present invention relates to an image monitoring system and method for removing an image,

The present invention relates to a system for confirming the position and removal of an acupuncture needle. More specifically, the present invention relates to a system for confirming the position and removal of the acupuncture needle, And more particularly, to a video monitoring system and method.

In general, Oriental Medicine, which is called Oriental Medicine, is a medicine that has been handed down over thousands of years but has not been able to benefit from science relatively as compared with Western medicine.

Oriental medicine treatment in Oriental medicine, the oriental doctor through the photographic method to determine the status of the patient is doing acupuncture or moxibustion. The photographic method is to grasp the state of the patient by watching, listening to, asking, and tapping.

Such acupuncture or moxibustion is being performed for the purpose of treatment and for the purpose of maintenance according to the health condition of the patient.

In order to further scientificize the herbal medicine treatment, various researches have been conducted. As a method of the above-mentioned herbal treatment science, methods such as skin impedance measurement, pulse wave measurement and skin temperature measurement are applied to the diagnosis method.

In addition, bedding, moxibustion and hot water medicine are used as the above-mentioned medical treatment methods of oriental medicine, and bedding or moxibustion is used as one of the skin stimulation methods for treating diseases by operating various stimulation methods on certain parts of the body surface In recent years, a variety of new therapies have been developed and used in clinical trials in addition to traditional stimulation methods.

However, in the case of acupuncture or moxibustion, it is a fact that various problems arise due to the method of stimulating the patient directly by touching the body surface.

Among them, in the case of acupuncture, not only the acupuncture size is small, but also a large number of acupuncture points are often settled immediately for treatment, so medical accidents are often occurred because the acupuncture is not completely performed during acupuncture. These medical accidents are one of the medical accidents that anyone in the oriental hospital can agree with.

In this way, all of the needles stuck to the body of the patient after being soaked in the body by the treatment means are not recovered during the footsteps but left in the patient's body, This can threaten the patient's life. In addition, this causes a large civil lawsuit, which causes the medical institutions to incur damages.

Therefore, it is required to provide a device that can effectively prevent the occurrence of such medical accidents.

Korean Patent Publication No. 10-2013-0140459

In order to solve the problem of the conventional acupuncture sample, the present invention provides a method of detecting an acupuncture sample by capturing an image of a procedure area to check the position and removal of the acupuncture needle, And to provide a video monitoring system and method for a one-shot so as to prevent a problem.

The present invention provides a video monitoring system for a stroller so that the position of the struck needle can be accurately displayed on the screen, And a method thereof.

The present invention relates to a video monitoring system and a method for a footstep so as to prevent a problem caused by incomplete footsteps by applying fluorescence or a light emitting material to the head portion and tracking the footstep using an image photographed under ultraviolet light or general illumination The purpose is to provide.

The present invention uses a color information of RGB color information to detect the needles, or simply uses information indicating that the intensity of light is relatively high without color information, thereby tracking the outbreak of the incontinence, And more particularly, to a video monitoring system and method for video surveillance.

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

In order to achieve the above object, according to the present invention, there is provided an image monitoring system for a foot, comprising: a movable body configured to move in a predetermined direction; A first and a second side radiographing part configured to be fixed on a side surface of the main body for the patient to take an image of a body surface of a side part of the patient and a pointed needle of the patient, And a tracking result display unit for analyzing the image photographed by the two side photographing unit and displaying the position of the operation and whether or not the foot is released based on the RGB color information of the head portion and the intensity of the head portion.

Here, the trace result display unit may include a binarization processor that finds and binarizes the acupuncture candidate region using a difference from a reference RGB color component of the emitted acupuncture head portion, and a binarization processor that performs noise elimination and size component extraction on the extracted candidate region A straight line component extraction unit for removing the false candidate region and extracting two straight line components by using edge and radon transformation and a candidate region verification unit for verifying the needle through the morphological analysis of the two linear components of the final candidate region And the like.

The binarization processing unit includes a difference image acquiring unit for acquiring a difference image between an image input unit for inputting a photographed image under ultraviolet light or general illumination and a reference RGB color component for a needle head, And a binary image acquiring unit for acquiring a binary image through color component analysis of the candidate region.

The straight line component extraction unit may include a false candidate region removal unit that removes a false candidate region in consideration of the size of the noise cancellation and acupuncture candidate region, and a false candidate region removal unit that extracts, from the image of the final candidate region from which the false candidate region is removed, An edge detection unit for performing edge detection, and a linear component extraction unit for extracting two principal straight line components by applying a radon transformation to the detected edge image.

The candidate region verification unit may include a first verification unit that performs one step verification of a candidate region based on the parallelism of the two extracted lines and a final verification unit that determines a distance between the extracted two lines and a length component of the main line And a secondary verification unit.

According to another aspect of the present invention, there is provided a method for monitoring an image of a foot, comprising the steps of: capturing a body surface of a front part of a patient and a time- A method of video monitoring, the method comprising: searching and binarizing an acupuncture candidate region using a difference from a reference color component of a emitted acupuncture head; removing noise from the extracted candidate region, Extracting two straight line components by using edge and radon transformation, removing the candidate region, and verifying needle through morphological analysis of two straight line components of the final candidate region.

The step of finding and binarizing the acupuncture candidate region includes the steps of acquiring a difference image from a reference RGB color component of the acupressure head portion when an image photographed under ultraviolet light or general illumination is input, And obtaining a binary image through color component analysis of the candidate region.

The threshold value is obtained by using the standard deviation of each reference RGB color,

의 유클리드 거리를 적용하여 후보영역 중 Red와 Green 성분이 Blue 성분보다 일정비율보다 적은 영역은 제거하는 것을 특징으로 한다.

The region where the red component and the green component are less than a certain ratio than the blue component is removed by applying the Euclidean distance of the candidate region.

The step of extracting the two straight line components using the edge and radon transformation includes the steps of removing the noise and removing the false candidate region based on the size of the acupuncture candidate region and the step of removing the false candidate region from the final candidate region Detecting an edge from an image of the edge image, and extracting two main straight line components by applying a radon transformation to the detected edge image.

Then, in the step of extracting the two main straight line components by applying radon transformation to the detected edge image, a radon transformation is applied to the edge image of each obtained candidate region, and the value of the largest value and the second largest value And extracts two main straight line components.

The step of verifying the needle is performed by performing a one-step verification for the candidate region by the parallelism of the extracted two straight lines, a step of performing the final verification by the interval of the extracted two straight lines and the length component of the main line And a control unit.

According to another aspect of the present invention, there is provided a method of monitoring an image for a foot, comprising the steps of: capturing an image of a front face of a patient and a time- A method of performing image monitoring is provided. When an image is input, a smoothing process is performed using an average filter in a preprocessing process, and a morphological operation opening and closing operations are applied, A step of extracting an acute candidate region from the candidate region by using the size and brightness information of each candidate region, performing a binarization process by reflecting the acupuncture characteristics, And a step of finding a parallel straight line component of the needle head portion and verifying the needle.

delete

The step of removing the false needle candidate region may include the steps of removing the false candidate region using the size and brightness of the needle candidate region and extracting the absolute value of the brightness value difference between neighboring pixels in the smallest rectangle surrounding the needle candidate region And removing the false candidate region using Region Growing to integrate the neighboring pixels smaller than 10 into the same region.

In the step of removing the false candidate region, the coupling component labeling is performed on the binarized image to obtain the area, which is the number of pixels constituting each acupuncture candidate region, the long axis length and the short axis length of the ellipse surrounding the acupuncture candidate region ,

의 조건을 만족하는 영역들을 거짓 침 후보 영역으로 판단하여 제거하고, 침 후보 영역을 구성하는 화소값들의 평균 및 분산을 이용하여,

의 조건을 모두 만족하는 후보 영역들을 제거하는 것을 특징으로 한다.

Is determined as a false needle candidate region, and the average and variance of the pixel values constituting the needle candidate region are used,

The candidate regions satisfying all of the conditions of < RTI ID = 0.0 >

의 조건을 만족하지 않으면 해당 후보 영역에는 침이 없다고 판단하는 단계와,가장 큰 직선을 찾은 경우, 나머지 직선 l_i 의 길이 L(l_i)를 사용하여

의 조건을 만족하는 직선들을 찾고 해당하는 직선이 없는 경우 해당 후보 영역에는 침이 없다고 판단하는 단계와,가장 큰 직선 l_max 와 평행한 직선을 찾기 위하여 두 직선 간의 각도 차이가 5도 이하인 직선들을 찾으며, 해당하는 직선이 없는 경우 해당 후보 영역에는 침이 없다고 판단하는 단계와, l_max, l_i 두 직선간의 거리 D(l_max, l_i)를 이용하여

을 만족하는 직선 l_i 가 존재하는 경우 침이라 판정하며, 해당하는 직선이 없는 경우 해당 후보 영역에는 침이 없다고 판단하는 단계를 포함하는 것을 특징으로 한다.And verifying the deposition is carried out and the Radon transform in the edge image of the needle candidate regions and step length to find the largest linear elements, the largest linear length l _max of L (l _max) is

(L _i ) of the remaining straight line l _i is used when the largest straight line is found,

Determining that there is no needle in the candidate region if there is no corresponding straight line; searching for straight lines having an angle difference of less than 5 degrees between two straight lines to find a straight line parallel to the largest straight line l _max; when the non-linear to the candidate area by the step and, l _max, l _i distance D _(max l, l _i) between the two straight lines determined that there is no needle

Determined as if the needle is a straight line l _i exist that satisfy, and is characterized in that it comprises the step of determining if there is no straight line corresponding candidate region has no needle.

The video monitoring system and method according to the present invention has the following effects.

First, the image of the operation area is captured and the image analysis can accurately confirm the position and removal of the operated needle.

Second, it is possible to precisely check the position and removal of the acupuncture needle, thereby preventing problems caused by incomplete assault.

Third, it is possible to check whether the acupuncture treatment is performed in a place necessary for treatment by displaying the location of the acupuncture needle on the screen, so that more effective and stable treatment is possible.

Fourth, it is possible to prevent problems caused by imperfect striking by applying fluorescence or luminescent material to the saliva head and tracking the striking by using images taken under ultraviolet light or general illumination.

Fifth, it is possible to precisely confirm the position and removal of the acupuncture needle by using only the information that the acupuncture point is relatively high without recognizing acupuncture or color information and using the applied RGB color information have.

Brief Description of the Drawings Fig. 1 is a block diagram of an image monitoring system for foot-
FIG. 2 is a view showing a principle of light reflection for tracking
FIG. 3 is a block diagram illustrating an upper body image acquiring process using an image monitoring system for footing according to the present invention
FIG. 4 is a block diagram illustrating a process of acquiring a lower body image using an image monitoring system for footing according to the present invention
FIG. 5 is a block diagram illustrating an image acquiring process in the arm and the abdomen using the image monitoring system for foot-taking according to the present invention
FIG. 6A is a detailed configuration diagram of a video monitoring system for footing according to the present invention
FIG. 6B is a flow chart illustrating a method of recognizing an acupuncture using RGB color information in a video monitoring system for footing according to the present invention
Figs. 7A to 7E are diagrams showing the step-by-step image structure in the needle recognition process using RGB color information
FIGS. 8A and 8B are diagrams showing an image showing two main linear component extraction processes in the detected image
FIGS. 9A and 9B are diagrams showing the process of selecting a final candidate region by the first and second verification processes
FIGS. 10A and 10B are diagrams showing an image showing the result of the recognition of the lower body image
Figs. 11A and 11B are diagrams showing an image showing the result of the recognition of the upper body image
FIG. 12 is a flowchart showing an acupuncture recognition process in an intensity image according to the present invention.
Figs. 13A to 13E are diagrams showing an image of a needle candidate detection and binarization process using morphological computation
14A and 14B are diagrams showing a process of removing a false needle candidate region
15 is a diagram showing an edge extraction and masking process
16 is a diagram showing a process of extracting two main straight line components from an edge image
FIGS. 17A to 17C are diagrams showing an image of a result of a stepping-off process using a video monitoring system for foot-

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a preferred embodiment of a video monitoring system and method according to the present invention will be described in detail.

The features and advantages of the image monitoring system and method for footing according to the present invention will be apparent from the following detailed description of each embodiment.

FIG. 1 is a block diagram of an image monitoring system for footing according to the present invention, and FIG. 2 is a block diagram illustrating a principle of light reflection for tracking a foot.

The present invention is able to prevent problems caused by imperfect stooping by using the applied RGB color information to track the stitch using only the information that the stitch is recognized or the intensity of the stitch head is relatively high without color information .

As shown in FIG. 1, the image monitoring system for footing according to the present invention includes a movable body and an upper photographing unit for photographing a body surface of a patient's front part and a time- (1), a first and second side radiographing sections (2) and (3) configured to be fixed on the side surface of the main body (5) where the patient is positioned for the hour hand, And a tracing result display unit (not shown) arranged on one side of the main body 5 for analyzing images photographed by the upper photographing unit 1 and the first and second side photographing units 2 and 3 to display a procedure position 4).

Here, the tracking result display unit 4 analyzes the photographed image and displays fluorescence or luminescent material on the acupressure head and displays the acupuncture using ultraviolet rays or an image photographed under general illumination do.

The method of recognizing the needle includes a method of recognizing the needle using the applied RGB color information or using only information that the intensity of the needle head portion is relatively high without color information.

The tracking result display section 4 has image processing software for recognizing the needle from the photographed image, and recognizes the hand struck directly from the photographed image and displays the position and the number of the hands.

The image monitoring system for the footstep according to the present invention automatically displays the number of total hands and the number of hands remaining after the hands are stuck so that a medical accident due to incomplete hands can be prevented in advance.

In an embodiment of the present invention, the yellow light emitting material is applied to the head portion of the needle, and the light reflection principle of the applied light emitting material is as shown in FIG.

The needle is small in size and simple in shape, making it difficult to extract the presence and position of the needle in a general manner.

In an embodiment of the present invention, fluorescence or luminescent material is applied to the needle head portion and ultraviolet rays or images photographed under general illumination are used in order to extract the needle effectively.

Since the fluorescent material emits visible light when it receives ultraviolet light, the head portion has a brighter light than the surrounding light, and the light emitting material emits visible light and likewise becomes brighter than the surrounding portion.

As shown in FIG. 2, the principle of the reflective tape made of glass beads is that, when the incident light enters, unlike a reflected light of a rough surface or a smooth surface, Can emit light, and the applied needle head part becomes brighter than the surrounding area.

Using this principle, according to the present invention, a method of recognizing needles by using the applied RGB color information and a method of recognizing by using only information that the intensity of the beating head is relatively high without color information is recognized, Position and number.

The photographing environment of the upper photographing section 1 and the first and second side photographing sections 2 and 3 is preferably controlled as follows.

It is preferable to use the remote control to keep the camera angle at 40 to 45 degrees and to prevent the camera from shaking.

Use it as a support to minimize unnecessary camera movement, and set the camera lens to turn on automatically when shooting with a circular light on the rim.

It is preferable that the photographing is performed in a state in which all the internal fluorescent lamps are turned on and the camera is moved to the back of the camera for minimizing ambient light change during photographing, and zooming is not used.

FIG. 3 is a block diagram illustrating an upper body image acquisition process using an image monitoring system for foot striking according to the present invention, and FIG. 4 is a block diagram illustrating a lower body image acquisition process using an image monitoring system for foot striking according to the present invention.

And FIG. 5 is a block diagram illustrating an image acquiring process of the arm and the abdomen using the image monitoring system for footing according to the present invention.

A video monitoring system and method for foot-on-foot according to the present invention will be described in detail as follows.

FIG. 6A is a detailed block diagram of an image monitoring system for footing according to the present invention, and FIG. 6B is a flowchart illustrating a needle recognition method using RGB color information in an image monitoring system for footing according to the present invention.

The trace result display unit 4 includes the following configuration for acupident recognition using RGB color information.

First, a binarization processing unit 60 for finding and binarizing the acupuncture candidate region by using a difference from the reference RGB color component of the illuminated needle head portion, and a noise removal and size component for the extracted candidate region A straight line component extraction unit 70 for extracting two straight line components by eliminating the false candidate region and using edge and radon transformation and a linear region extraction unit 70 for extracting a candidate region for verifying the needle through morphological analysis of the two linear components of the final candidate region And a verification unit (80).

Here, the binarization processing unit 60 includes an image input unit 61 for inputting a photographed image under ultraviolet light or general illumination, and a difference image acquiring unit 62 for acquiring a difference image between a reference RGB color component of the needle- And a binary image acquiring unit 63 for acquiring a binary image through threshold value extraction and color component analysis of a candidate region.

The straight line component extraction unit 70 includes a false candidate region removal unit 71 for removing the false candidate region in consideration of the size of the noise cancellation and acupuncture candidate region and a false candidate region removal unit 71 for removing the false candidate region from the false candidate region removal unit 71 And a linear component extraction unit 73 for extracting two principal linear components by applying a radon transformation to the detected edge image, .

The candidate region verification unit 80 includes a first verification unit 81 for performing a one-step verification on a candidate region based on the parallelism of the extracted two straight lines, And a secondary verification unit 82 for performing final verification.

The handwriting recognition algorithm using RGB color information according to the present invention is largely composed of three steps.

First, the process of finding and binarizing the acupuncture candidate region using the difference from the reference RGB color component of the illuminated head portion (S601 to S603), considering the noise removal and size components of the extracted candidate region (S604 to S606) of extracting two straight line components by using edge and radon transforms, and a process of verifying needles by morphological analysis of the two linear components of the final candidate region (S607 to S608).

Specifically, when an image photographed under ultraviolet light or general illumination is inputted (S601), a difference image is acquired from a reference RGB color component of the head portion (S602)

Subsequently, a threshold value is extracted and a binary image is obtained through color component analysis of the candidate region (S603)

Then, the false candidate region is removed by removing the noise and considering the size of the needle candidate region (S604).

Then, an edge is detected from the image of the final candidate region from which the false candidate region is removed (S605)

Then, radon transformation is applied to the detected edge image to extract two principal line components (S606)

Then, a one-step verification is performed on the candidate region based on the parallelism of the extracted two straight lines (S607), and the final verification is performed based on the extracted two straight lines and the length component of the main straight line (S608).

Each step of the needle recognition process using the RGB color information according to the present invention will be described in more detail as follows.

FIGS. 7A to 7E are step-by-step image configuration diagrams in the needle recognition process using RGB color information, and FIGS. 8A and 8B are image configuration diagrams illustrating two main linear component extraction processes in the detected image.

9A and 9B are block diagrams illustrating a process of selecting a final candidate region by the first and second verification processes.

The image input step S601 is performed as follows.

In the present invention, fluorescence or luminescent material is applied to the head portion for effective extraction of acupuncture, and ultraviolet rays or images photographed under general illumination are used.

The image of FIG. 7A is the image of the lower end portion of the patient under the general illumination by applying the yellow light emitting material to the head portion.

When the patient complains of a feeling of refusal to shoot, the ambient light may be darkened and ultraviolet light may be used to acquire the image. As with the light emitting material, due to the applied fluorescent material, Can be applied equally.

The difference image acquisition (S602) with the reference RGB color component of the head portion is performed using the reference RGB color component of the salient head portion emitted in yellow as shown in FIG. 7B, Equation 1 is applied to the Euclidean distance to obtain the difference image.

In the step S603 of obtaining the threshold value and the color component analysis of the candidate region, the threshold value is obtained by using the standard deviation of each reference RGB color, and the candidate region extracted is not the composition ratio of the RGB color component As shown in FIG. 7C, since the Euclidean distance of Equation 1 depends entirely on the Euclidean distance, the region where the red and green components are less than a certain percentage of the blue component in the candidate region is removed (yellow light emitting material is applied to the needle portion)

This process can be expressed by Equation (2).

In the step S604 of removing the false candidate region in consideration of the size of the noise elimination and acupuncture candidate regions, an opening operation and a closing operation of the morphological operation are applied to the acquired binary image Noise is removed and candidates of more than or equal to a predetermined size are removed using the size (area) information of each candidate region.

FIG. 7D shows a binary image, and FIG. 7D shows a needle candidate region displayed on a gray image of the input image, in which a false candidate region is removed.

In operation S605, the edge of the candidate region is obtained by applying an erosion operation, which is a morphological operation, to the edge region of the obtained candidate region as in Equation (3).

This process can be also obtained by using an edge detection mask (sobel, canny, etc.) on the intensity image of the candidate region.

FIG. 8A shows edge detection of the acupuncture candidate region, and FIG. 7E shows an edge for the first needle.

In operation S606, radon transformation is applied to the detected edge image to extract two principal line components. In step S606, radon conversion is applied to the edge image of each candidate region, And extracts two principal line components.

FIG. 8B shows detection of two linear components by applying a radon transformation to the edge image of FIG. 8A.

The first step verification step (S607) for the candidate region based on the parallelism of the extracted two straight lines is to determine whether the extracted two straight line components are parallel (angle measurement of each straight line) And the first step verification for

FIG. 9A shows the image after the one-step verification of the candidate region (the false candidate region 9 disappears).

Then, in the final verification step (S608) based on the interval between the two straight lines and the length component of the main straight line, whether the interval between the extracted two straight lines and the length of the main straight line are within a certain range is verified using Equation (5) As shown in FIG. 9B.

10A and 10B and FIGS. 11A and 11B show recognition results of the lower body and upper body images by applying such an algorithm.

FIGS. 10A and 10B are diagrams showing an image capturing result for a lower body image, and FIGS. 11A and 11B are views showing an image capturing result for an upper body image.

In the video monitoring system and method for foot-on-foot according to the present invention, acupuncture recognition algorithm in the intensity image is as follows.

FIG. 12 is a flowchart illustrating an acupuncture process in an intensity image according to the present invention.

Acupuncture recognition in the intensity image recognizes the acupoint with only the intensity image without the color information. The acupuncture recognizes the acupuncture region by using a large morphological operation and finds a bright or dark area smaller than the surrounding area, S1204), removing the false needle candidate region using the size and brightness information of each candidate region (S1205 to S1206), searching the edge and performing radon transformation to find a parallel straight line component of the needle head portion And a step of verifying the needle (S1207 to 1209).

First, when an image is input (S1201), smoothing is performed using an average filter in a preprocessing step (S1202)

Subsequently, an acupuncture candidate region using the morphological operation is extracted (S1203), and the binarization process is performed reflecting the acupuncture characteristics (S1204)

Then, the false candidate region is removed using the size and brightness of the acupuncture candidate region (S1205)

Subsequently, the false candidate regions are removed by using Region Growing, which is divided into basic regions having common characteristics of images, and regions in which the differences in the predetermined characteristics are small are continuously integrated to leave only regions with large differences in characteristics (S1206 )
Here, the region growth is to integrate neighboring pixels having an absolute value of brightness difference difference of neighboring pixels smaller than 10 in the smallest rectangle surrounding the needle candidate region into the same region.

Then, edge detection of the final candidate region is performed (S1207), two main line components are extracted by applying the radon transformation to the edge image of the candidate region (S1208), and needle extraction is performed (S1209)

Here, the needle candidate detection step using the morphological operation (S1203) extracts the needle candidate using the characteristics of the open operation and the close operation of the morphological operation.

FIGS. 13A to 13E are block diagrams illustrating acupuncture candidate detection and binarization processes using morphological operations.

13A is an input image.

13A and 13B show a result of performing a warm-up operation using a circle-shaped structural element larger than the thickness D of the saliva head portion, as shown in FIG. 13B , And it can be seen that the saliva head portion on the skin has disappeared.

13C is a result of performing a close operation on the input image, and it can be seen that the head portion on the sheet has disappeared.

Before the open and close operations are performed, smoothing of the input image is performed using the average value filter of the mask size m.

In order to find the acupuncture candidate region, the image I _{close obtained} by performing the _open operation in the image I _open and the image I _{close obtained} by performing the close operation are subtracted as shown in Equation 6, and the absolute value is taken to obtain the acupunctive candidate region as shown in FIG. 13d.

In the binarization step (S1204) reflecting the characteristics of the acupuncture, as shown in FIG. 13D, the acupuncture is thicker than the other parts and is small in size. Therefore, in the present invention, the histogram of the image is obtained, The binarization is performed with the pixel value n where the number of pixels from 0 to n is 0.5% of the total number of pixels as a threshold value.

The step of removing the false needle candidate region using the size information and the brightness information (S1205) performs connection component labeling on the image of FIG. 13E, which is the binarized image, to calculate an area, which is the number of pixels constituting each needle candidate region, The length and the short axis length of the ellipse surrounding the region are determined and the regions satisfying the condition of Equation (7) using the long axis length and the short axis length are determined as a false needle candidate region and removed.

Also, since the needle portion is not generally a dark portion and the needle portion has uniform brightness, candidate regions that satisfy both of the conditions of Equation 8 using the mean and variance of the pixel values constituting the needle candidate region are removed .

여기서, Area₁ 은 면적의 상한 기준값, Area₂ 는 면적의 하한 기준값, l₁ 은 장축 길이 기준값, l₂ 는 단축길이의 하한 기준값, l₃ 는 단축길이의 상한 기준값, Avg는 밝기값 평균의 기준값, Var은 분산 기준값이다.

Wherein, Area ₁ is the upper limit reference value of the area, Area _2, the lower limit reference value of the area, l ₁ is the lower limit reference value of the long axis length reference value, l ₂ are shorter in length, l ₃ is an upper limit reference value of the shortening, Avg is the reference value of the average brightness value , And Var is a dispersion reference value.

Then, the process of removing the false needle candidate region using region growth (S1206) proceeds as follows.

14A and 14B are block diagrams illustrating a process of removing a false needle candidate region.

In the process of extracting the acupuncture candidate region by using the morphological operations of opening and closing, there are many cases where the acupuncture region is not acuminate but is smaller in size and brighter or darker than the surrounding region.

A typical case is a small area between characters on a bed sheet.

The white areas in (a) of FIG. 14 (a) and the area of (a) in FIG.

FIGS. 14A and 14B show enlargement of the width and height of the minimum included rectangle of each candidate region and extraction of the corresponding rectangular region from the binary image. FIG.

FIGS. 14A and 14B show extracted rectangular regions in the intensity image. FIG. It is the area between the characters on the sheet and continues beyond the rectangular area.

14 (d) and 14 (b) show the results of region growth for (c). It can be seen that the sheet portion is detected over the entire area within the square.

If the height or width of the area obtained by performing the area growth is equal to the height or width of the enlarged area as shown in FIGS. 14 (d) and 14 (b), it is regarded as a sheet and removed.

When applying the region growth, the seed value uses the average value of the pixel values of the acupuncture candidate region, and uses a threshold of ± 10.

The edge extraction and unnecessary edge removal step (S1207) proceeds as follows.

15 is a configuration diagram showing an edge extraction and masking process.

15 (a) is a cut-out of the least-squares rectangular region of the acupunctive candidate region from the intensity image, and (b) is the result of finding the edge by applying the canny edge operator.

It can be seen that there are many edges in the needle as well as the outer part of the needle.

(c) is the result of finding and expanding the candidate region corresponding to (a) in the binary image, and (d) is the result of erosion after inversion.

(c) and (d) show 0 for the black part and 1 for the white part.

(e) is a result of removing unnecessary edges according to Equation (9). In this equation, I _edge is the edge image, I _d is the expansion association, and I _e is the erosion image. It can be seen that only the edge of the needle part is left and the unnecessary remaining edge part is cleanly removed.

Then, the probe verification step (S1208) using radon conversion proceeds as follows.

FIG. 16 is a configuration diagram illustrating two principal linear component extraction processes in an edge image. FIG.

Since the needle head is composed of two long straight lines parallel to each other, the radon transformation is performed on the edge image of the needle candidate region and the straight line component having the longest length is found.

It does not satisfy the greatest straight line l _max length L (l _max) the condition of the equation (10) of the judges is that there is no needle corresponding candidate area.

If the largest straight line is found, straight lines satisfying the condition of the expression (11) are searched using the length L (l _i ) of the remaining straight line l _i .

If there is no corresponding straight line, it is determined that there is no needle in the candidate region.

In order to find a straight line parallel to the largest straight line l _max , straight lines with an angle difference of less than 5 degrees between two straight lines are searched. If there is no corresponding straight line, it is judged that there is no needle in the candidate area.

Then, since the distance between the two lines corresponds to the thickness of the head portion, l _max , l _{i when} the straight line l _i satisfying the expression (12) using the distance D (l _max , l _i ) If there is no corresponding straight line, it is determined that there is no needle in the candidate region.

FIG. 16 shows the result of extracting two linear components in the area having acupuncture points.

The final acupuncture extraction result using the image monitoring method for the footstep according to the present invention is as follows.

FIGS. 17A through 17C are diagrams illustrating an image of a stepping-out process using a video monitoring system for footing according to the present invention.

In the image of FIG. 17A in which the fluorescent material was applied, both the four needles on the left side and the four needles on the right side were accurately extracted.

In the image of FIG. 17B, both the left needle 3 and the right needle 3 were accurately extracted. In the image of FIG. 17C in which the luminescent material is applied, it can be seen that both the left needle needle 6 and the right needle needle 6 are accurately extracted.

The image monitoring system and method according to the present invention can be used to recognize the needles by using the RGB color information applied to the head of a person's head, or to use the information that the intensity of the head is relatively high without color information So that problems caused by imperfect strikes can be prevented.

As described above, it will be understood that the present invention is implemented in a modified form without departing from the essential characteristics of the present invention.

It is therefore to be understood that the specified embodiments are to be considered in an illustrative rather than a restrictive sense and that the scope of the invention is indicated by the appended claims rather than by the foregoing description and that all such differences falling within the scope of equivalents are intended to be embraced therein It should be interpreted.

61. Image input unit 62. Difference image obtaining unit
63. Binary image acquisition unit 71. False candidate region removal
72. Edge detection unit 73. Linear component extraction unit
81. Primary verification unit 82. Second verification unit

Claims (16)

Fluorescent or luminescent material is applied to the head portion,
An upper photographing unit configured to take a photograph of a patient's front face portion and an hourly needles to check the position of the acupuncture needle and whether or not the user performs a stroke;
A first and a second side radiographing unit configured to fix the side surface of the body, the first side and the second side radiographing the side surface of the patient and the hour hand;
The image is captured by the upper photographing unit and the first and second side photographing units on one side of the main body, and the image is displayed on the basis of the RGB color information of the needle head portion and the intensity of the needle head portion And a result display unit for displaying an image of the subject. The apparatus according to claim 1,
A binarization processor for searching and binarizing the acupuncture candidate region using a difference from a reference RGB color component of the illuminated needle head,
A straight line component extraction unit for removing a false candidate region by considering noise removal and size components of the extracted candidate region and extracting two straight line components by using edge and radon transformation;
And a candidate region verifying unit for verifying the needle through morphological analysis of the two linear components of the final candidate region. 3. The image processing apparatus according to claim 2,
An image input unit for inputting a photographed image under ultraviolet light or general illumination,
A difference image acquiring unit for acquiring a difference image from a reference RGB color component of the head portion,
And a binary image acquiring unit for acquiring a binary image through threshold value extraction and color component analysis of a candidate region. 3. The image processing apparatus according to claim 2,
A false candidate region removing unit that removes a false candidate region by considering a size of a noise cancellation candidate region,
An edge detection unit for performing edge detection from the image of the final candidate region from which the false candidate region has been removed in the false candidate region removal,
And a linear component extraction unit for extracting two principal linear components by applying a radon transformation to the detected edge image. The method of claim 4, wherein, in order to verify the needle through morphological analysis of the two principal straight line components of the final candidate region extracted by the linear component extraction unit,
The candidate region verifier,
A first verification unit for performing a one-step verification of a candidate region by the parallelism of the extracted two straight lines,
And a second verification unit for performing final verification based on the interval and the length component of the extracted two straight lines. Fluorescent or luminescent material is coated on the head of the needle and the body is configured to be movable so that the body surface of the patient's front part and the time-of-arrival needle are photographed to check the position of the needle and whether the needle is struck or not. In this case,
Searching and binarizing the acupuncture candidate region using the difference from the reference RGB color component of the emitted needle head portion;
Extracting two linear components using edge and radon transforms, removing noise candidates for the extracted candidate regions, removing false candidate regions based on size components,
And verifying the needle through morphological analysis of the two linear components of the final candidate region. 7. The method of claim 6, wherein finding and binarizing the acupuncture candidate region comprises:
Acquiring a difference image from a reference RGB color component of the head portion when an image photographed under ultraviolet light or general illumination is input;
Extracting a threshold value, and acquiring a binary image through color component analysis of a candidate region. 8. The method of claim 7, wherein the threshold value is obtained using a standard deviation of each reference RGB color,

Wherein a region where the red and green components of the candidate region are less than a certain percentage of the blue component is removed by applying the Euclidean distance of the candidate region. 7. The method of claim 6, wherein extracting the two straight line components using edge and radon transforms comprises:
Removing noise and removing a false candidate region based on the size of the acupuncture candidate region,
Detecting an edge from an image of a final candidate region from which a false candidate region has been removed;
And applying a radon transformation to the detected edge image to extract two principal straight line components. 10. The method according to claim 9, wherein, in the step of extracting two principal linear components by applying a radon transformation to the detected edge image,
And applying the Radon transform to the edge image of each candidate region to extract the largest and the second largest values to extract two principal straight line components. 10. The method of claim 9, wherein, in order to verify the needle using the extracted two principal components,
Performing a one-step verification of a candidate region based on whether or not two extracted straight lines are parallel;
And performing final verification based on an interval and a length component of the extracted two straight lines. A method for monitoring an image for footing by photographing a body surface of a front part of a patient and an hourly needle in order to confirm the position of the needle,
When the image is input, smoothing is performed using an average filter in the preprocessing process, extraction of the acupuncture candidate region by applying morphological operations such as opening and closing operations, Searching for a needle candidate region;
Removing the false needle candidate region using the size and brightness information of each candidate region;
And performing radon transformation on the edge of the image to search for a parallel straight line component of the needle head portion to verify needle penetration. delete 13. The method of claim 12, wherein removing the false needle candidate region comprises:
Removing a false candidate region using the size and brightness of the needle candidate region;
And removing the false candidate region using region growth, which merges neighboring pixels having an absolute value of brightness difference of neighboring pixels smaller than 10 in the smallest rectangle surrounding the needle candidate region into the same region Wherein the step of detecting the image of the foot is performed by the user. 15. The method of claim 14, wherein removing the false candidate region using the size and brightness of the acupuncture candidate region comprises:
The coupling component labeling is performed on the binarized image to obtain the area as the number of pixels constituting each acupuncture candidate region, the long axis length and the short axis length of the ellipse surrounding the acupuncture candidate region,

Are determined as a false needle candidate region and removed,
Using the average and variance of the pixel values constituting the acupuncture candidate region,

Candidate regions satisfying all the conditions of < RTI ID = 0.0 >
Area ₁ is the upper limit reference value of the area, Area _2, the lower limit reference value of the area, l ₁ is the lower limit reference value of the long axis length reference value, l ₂ are shorter in length, l ₃ is an upper limit reference value of the shortening, Avg is the reference value of average brightness value, Var Is a variance reference value. 13. The method of claim 12,
Performing a radon transformation on an edge image of the acupuncture candidate region and finding a straight line component having a longest length,
The largest linear length l _max of L (l _max) is

Judging that there is no needle in the candidate region,
If the largest straight line is found, the length L (l _i ) of the remaining straight line l _i is used

If there is no corresponding straight line, it is determined that there is no needle in the candidate region;
And determining the largest straight line l _max and the angle difference between two straight lines in order to find a straight line parallel to the straight line looks for the less than 5 degrees, in the case that there is no straight line corresponding to a candidate region has no needle,
l _max , l _{i Using} the distance D (l _max , l _i ) between two straight lines

Video monitoring method for balchim comprises a linear l _i is determined to be present, determining needle and, if there is no straight line corresponding candidate region if there is no needle satisfying.

Images