KR101271620B1

KR101271620B1 - Method of detecting painted surface and painting method using the same

Info

Publication number: KR101271620B1
Application number: KR1020100140352A
Authority: KR
Inventors: 박영준; 조기용; 최두진; 이재상; 최승준; 배성준; 최윤규
Original assignee: 삼성중공업 주식회사
Priority date: 2010-12-31
Filing date: 2010-12-31
Publication date: 2013-06-11
Also published as: KR20120078143A

Abstract

A paint surface detection method is presented. According to an exemplary embodiment, a method for detecting a painting surface may include: reinforcing an image of a painted surface, distinguishing colors of the enhanced image, extracting boundaries of paint regions divided according to the divided colors, and Setting up the node.

Description

Coating surface detection method and coating method using the same {METHOD OF DETECTING PAINTED SURFACE AND PAINTING METHOD USING THE SAME}

The present invention relates to a coating surface detection method and a coating method using the same.

When painting in hull blocks or other unfavorable working environment, it is common to work in unmanned environment using a painting robot. When working in such an unmanned environment, it is necessary to move the painting robot to the correct position in order to paint the desired color in the desired area.

When a person works, the painter visually checks the painting surface and paints it at a desired position, but in the case of a painting robot, the painting robot predicts the painting environment and programs a route accordingly to move the painting robot. However, the painting robot may not be able to move to the desired location if the predicted environment and the actual physical environment are different or an error occurs.

One embodiment of the present invention is to enable the painting robot to move to the correct position on the painting surface to perform the painting operation.

In the painting surface detection method according to an embodiment of the present invention, the method may further include reinforcing an image photographing the painting surface, distinguishing colors of the enhanced image, and extracting a boundary of a painting area divided according to the divided colors. And setting up a node at said boundary.

The image enhancement step may include performing histogram specification.

The histogram specification step may include: converting a color coordinate of each pixel of the enhanced image into a YCrCb color coordinate system; Generating an ideal histogram and an ideal cumulative distribution function for the illuminance of the enhanced image, and matching the ideal histogram and the ideal cumulative distribution function with the actual histogram and the cumulative distribution function. The method may include calculating a histogram and a final cumulative distribution function, and converting color coordinates of each pixel based on the final histogram.

The calculating of the final histogram and the final cumulative distribution function may include finding an ideal cumulative distribution function value (hereinafter referred to as an “ideal function value”) closest to the actual cumulative distribution function value of the gray level for each gray level. Finding a gray level having the abnormal function value in the ideal cumulative distribution function, finding a value of the gray level having the abnormal function value in the ideal histogram (hereinafter, referred to as an “abnormal histogram value”), and each gray level. The method may include generating the final histogram by assigning the abnormal histogram value to, and obtaining the final cumulative distribution function from the final histogram.

The coordinate conversion step may include converting the Y coordinate of each pixel to a gray level having the abnormal function value.

The color classification step may include converting the coordinate converted color coordinates for each pixel into an RGB color coordinate system, normalizing the RGB color coordinates of each pixel, and normalizing the standardized RGB color coordinates and the reference color of each pixel. Calculating a distance between the RGB color coordinates, and replacing the normalized RGB color coordinates of each pixel with the standardized RGB color coordinates of the reference color at the closest distance.

The reference color may be a color painted on the painted surface.

The boundary extracting step may use a Laplace boundary.

The node setting step may include setting a start point and an end point of the boundary.

The painting surface detection method according to an embodiment of the present invention comprises the steps of photographing the painting surface, reinforcing the photographed image of the painting surface, distinguishing the color of the enhanced image, divided according to the divided color Extracting a boundary of a painting area, setting a node at the boundary, moving a painting module according to the node, and painting the painting surface using the painting module.

The image enhancement step may include converting a color coordinate of each pixel of the enhanced image into a YCrCb color coordinate system, and calculating an actual histogram and an actual cumulative distribution function for illuminance which is Y coordinate of the YCrCb color coordinate system for the enhanced image. Generating an ideal histogram and an ideal cumulative distribution function for the illuminance of the enhanced image, and matching the ideal histogram and the ideal cumulative distribution function with the actual histogram and the cumulative distribution function. The method may include calculating a histogram and a final cumulative distribution function, and converting color coordinates of each pixel based on the final histogram.

The color classification step may include converting the coordinate converted color coordinates for each pixel into an RGB color coordinate system, standardizing the RGB color coordinates of each pixel, and normalizing one or more reference colors painted on the painting surface. Calculating a distance between an RGB color coordinate and a normalized RGB color coordinate of each pixel, and replacing the standardized RGB color coordinate of each pixel with a standardized RGB color coordinate of a reference color at a closest distance. .

In the painting surface detection method according to an embodiment of the present invention, the coating surface may be photographed and processed, and then the painting module may be moved to a more accurate position because the target position of the painting module is found.

1 is a schematic diagram of a painting apparatus according to an embodiment of the present invention.
2 is a schematic diagram of a painting module according to an embodiment of the present invention.
3 is a flowchart of a painting operation using the painting apparatus according to an embodiment of the present invention.
4 is a view showing the painted surface of the coating target according to an embodiment of the present invention.
5 is a flowchart illustrating a process of processing a photographed image of a painted surface in a painting method according to an embodiment of the present invention.
6 is a view for explaining an image reinforcing step in the painting method according to an embodiment of the present invention.
7 is a graph showing an actual histogram and a cumulative distribution function in an image reinforcing step of a painting method according to an embodiment of the present invention.
8 is a graph showing an ideal histogram and cumulative distribution function in an image reinforcing step of a painting method according to an embodiment of the present invention.
9 is a graph showing a final histogram and a cumulative distribution function in an image reinforcing step of a painting method according to an embodiment of the present invention.
FIG. 10 is a diagram for describing a process of obtaining a final histogram in an image reinforcing step of a painting method according to an embodiment of the present invention.
11 is a view for explaining the color separation step in the coating method according to an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The above and other features and advantages of the present invention will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings in which: FIG. The present invention may be embodied in many different forms and is not limited to the embodiments described herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and the same or similar components are denoted by the same reference numerals throughout the specification.

First, the structure of the coating apparatus according to an embodiment of the present invention will be described in detail with reference to FIGS. 1 and 2.

1 is a schematic diagram of a painting apparatus according to an embodiment of the present invention, Figure 2 is a schematic diagram of a painting module according to an embodiment of the present invention.

Referring to FIG. 1, the painting apparatus 100 according to the present embodiment includes a frame 10, a painting module 20, a plurality of wires 30, an autonomous movement control module 40, and a controller 50. .

The painting module 20 paints the surface of the workpiece 60 in the frame 10.

The painting module 20 includes a platform 21 movable within the frame 10, a work robot 22 and a camera 25 mounted on the platform 21, and a sprayer mounted on the work robot 22 ( 23).

The movement of the platform 21 is performed by a plurality of wires 30 connected between the frame 10 and the platform 21. One end of each wire 30 is fixed to the frame 10 and the other end is fixed to the platform 21. The number of wires 30 is for example four pairs, where the position of the platform 21 can be determined by adjusting the length of each wire 30. The length adjustment of the wire 30 is made by the autonomous movement control module 40.

The working robot 22 may include a multi-joint manipulator that can change the position on the platform 21 and, in some cases, adjust the position and angle.

The sprayer 23 is provided with a plurality of spray nozzles 24 to spray the paint 29 onto the surface of the work object 60.

The camera 25 photographs the painted surface of the work object 60.

The controller 50 is connected to the work robot 22, the camera 25, and the autonomous movement control module 40 of the painting module 20 through a cable 55 to control them. Specifically, the control unit 50 separates the coating area of the coating surface from the image of the coating surface photographed by the camera 25, extracts the boundary between the coating regions, and moves the coating module 20 based on the painting to move the coating. The operation is performed, and this will be described in detail with reference to FIGS. 3 and 4 together with FIGS. 1 and 2.

3 is a flowchart of a painting operation using a painting apparatus according to an embodiment of the present invention, Figure 4 is a view showing a painted surface of the coating target according to an embodiment of the present invention.

Referring to FIG. 3, first, a work robot 22, a sprayer 23, a camera 25, and the like are installed on a platform 21, and equipment installation is completed, such as filling a paint of a color to be painted on the sprayer 23. And initialize (S10).

After photographing the painting surface of the work object 60 through the camera 25, image processing is performed, and based on this, the target position to be moved by the painting module 20 is determined and compared with the current position to move the path. Perform the measurement operation to determine the (S20).

For example, as shown in the photograph of the painting surface of FIG. 4, the painting of the upper three painting areas 61, 63, and 65 has been completed in the previous work, and the lowermost painting area 67 has only the base color painted and the main color not yet painted. Let's say the state. In this case, it is assumed that the upper two coating areas 61 and 63 are coated with paints of different colors, and the coating areas 65 thereunder are coated with paints currently filled in the painting module 20. The controller 50 obtains a clean image by preprocessing the photographed image, and then checks the color painted on each of the painting areas 61, 63, 65, and 67 with a reference color, and checks the painting areas 61, 63,. The boundary between 65 and 67 is extracted to determine that the point below the boundary between the third painting area 65 and the lowest painting area 67 is the target position. Next, the target position is compared with the current position to determine the path to move. This image processing process will be described in more detail later.

When the movement path is determined, the painting module 20 is moved to the target position along the corresponding path (S30).

Taking the paint surface of the work object 60 once again and grasping the target position through the substantially same process as described above (S40), and comparing the current position with the current position of the painting module 20 and the current position and the target position It is determined whether the same (S50). If the current position is the same as the target position, the filled paint is sprayed onto the work object 60 to perform a painting operation (S60), otherwise, return to step S30 to correct the position of the painting module 20.

Painting operations are usually performed by spraying the paint while the painting module 20 moves from the right end to the left end of the paint surface or in the opposite direction. In FIG. 4, the boundaries between the coating areas 61, 63, 65, and 67 are curved, but may be formed in a straight line and parallel to each other. In this case, the painting module 20 may move along a straight line without changing the height. .

When the painting module 20 starting from one end reaches the opposite end, it is determined whether the painting work is completed (S70). If the painting work is completed, the equipment is dismantled (S80), otherwise, the process returns to step S30.

Next, an image processing process of processing an image of the painted surface photographed with reference to FIGS. 5 to 11 will be described in detail.

5 is a flowchart illustrating a process of processing a photographed image of a painted surface in a painting method according to an embodiment of the present invention, and FIG. 6 is a view for explaining an image reinforcing step in a painting method according to an embodiment of the present invention. 7 is a graph showing the actual histogram and the cumulative distribution function in the image enhancement step of the coating method according to an embodiment of the present invention, and FIG. 8 is an ideal image enhancement step of the coating method according to an embodiment of the present invention. 9 is a graph showing a histogram and a cumulative distribution function, FIG. 9 is a graph showing a final histogram and a cumulative distribution function in an image reinforcing step of a painting method according to an embodiment of the present invention, and FIG. 10 is a graph showing a cumulative distribution function according to an embodiment of the present invention. FIG. 11 is a diagram illustrating a process of obtaining a final histogram in an image reinforcing step of a painting method, and FIG. 11 is a view of an embodiment of the present invention. In the coating method according to an embodiment, a view for explaining a color classification step is performed.

First, the image photographed by the camera 25 is digitized so that the position on the painted surface is a pixel, and the color of the position is stored in a storage device (not shown) or the like as a color coordinate of the pixel.

Referring to FIG. 5, first, the controller 50 reads an image stored in a memory device (S21).

Next, an image enhancement operation using a histogram specification method is performed on the captured image to reduce distortion of the image due to reflection or shadow by light (S22). For example, the input image shown in (a) of FIG. 6 may have an unclear color due to a shadow on a corner of the painted surface, and an unclear color due to light reflection in the lower right middle part. If the input image is preprocessed using the histogram specification method, an image as shown in FIG. 6 (b) can be obtained. Although the overall image is darkened, the color is clearly displayed in all areas.

To describe the histogram specification method, first, the color coordinate of each pixel is converted from the RGB color coordinate system to the YCrCb color coordinate system. Then, the color coordinate of each pixel is represented by the Y coordinate corresponding to the illuminance and the other two coordinates. The illuminance may be displayed in 256 gray levels, for example.

Then, each illuminance value, that is, the number of pixels having gray levels is calculated, and a histogram representing the illuminance and a cumulative distribution function corresponding thereto are generated. In FIG. 7, (a) is a real histogram and (b) is a real cumulative distribution function, and the actual cumulative distribution function P _o (j) for gray level j is

. Here, p _o (i) is the number of pixels having the i-th gradation.

Next, an average histogram, a maximum deviation, a minimum value, and the like of grays of each pixel are obtained, and an ideal histogram representing a normal distribution and a cumulative distribution function corresponding thereto are generated based on the average and standard deviation of the gray levels. 8 (a) is an ideal histogram and (b) is a corresponding cumulative distribution function, the ideal cumulative distribution function P _id (j) for gray level j is

. Here, p _id (i) is the number of pixels having the i th gray scale in the ideal histogram.

The final histogram and the corresponding cumulative distribution function shown in FIGS. 9A and 9B can be obtained by matching the ideal histogram and cumulative distribution function and the actual histogram and cumulative distribution function thus obtained. In FIG. 9, dotted lines represent actual histograms and actual cumulative distribution functions.

To explain this process in detail, first, for each of the gray scales, the value of the ideal cumulative distribution function closest to the actual cumulative distribution function of the gray scale is found. As shown in Fig. 10, assume that the graph on the left is the actual cumulative distribution function and the graph on the right is the ideal cumulative distribution function. In the graph of the actual cumulative distribution function on the left, the actual cumulative distribution function value of the gray level k is P _o (k). In the ideal cumulative distribution function graph on the right, there are two values, P _id (m) and P _id (n), of which the value close to P _o (k) is P _id (m).

Next, find the grayscale with the ideal cumulative distribution function value found. M corresponds to this case. Next, find the value for gray m, i.e., the number of pixels p _id (m) in the ideal histogram, and set the value p _f (k) for gray k in the final histogram as p _id (m).

The final cumulative distribution function is calculated in the same way as the actual cumulative distribution function is obtained from the actual histogram.

Thus, since the gray level k in the actual histogram is changed to the gray level m in the ideal histogram, it is the last step to change the gray level of the pixel whose gray level is k to m in the input image.

5 again, when the image enhancement operation is completed, color classification is performed (S23). To do this, first convert the color coordinates of each pixel represented by the YCrCb color coordinate system to the RGB coordinate system and normalize them. The distance between the normalized RGB color coordinates of each pixel and the standardized RGB color coordinates of the various reference colors already painted (for example, four reference colors painted on the four paint regions 61, 63, 65, and 67 in FIG. 4). The color classification step is completed by calculating and replacing the color coordinates of each pixel with the color coordinates of the nearest reference color. For example, the image shown in (a) of FIG. 11 is converted to the image shown in (b) after the color separation step is completed, and each color and its boundary appear clearly in the converted image.

Subsequently, the boundary between the painting areas 61, 63, 65, and 67 is extracted using a Laplacian edge or the like (S24). In FIG. 4, the borders 62, 64, 66 of the coating areas 61, 63, 65, and 67 extracted using the Laplacian edge are indicated by white lines.

Finally, referring to FIG. 4, nodes such as start points a, c and e and end points b, d and f are set at the extracted boundaries 62, 64 and 66 (S25). The starting points a, c and e and the ending points b, d and f serve as reference for setting the movement position of the painting module 20.

As described above, in the painting surface detection method according to the present embodiment, the coating surface may be photographed and image processed, and then the painting module may be moved to a more accurate position because the target position of the painting module is found.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, Of the right.

10: frame
20: painting module
21: platform
22: working robot
23: sprayer
24: spray nozzle
25: camera
29: paint
30: wire
40: autonomous movement control module
50: control unit
55: cable
60: work object
61, 63, 65, 67: paint area
62, 64, 66: boundary of paint zone

Claims

delete

Reinforcing the image of the painted surface,
Distinguishing colors of the enhanced image;
Extracting a boundary of the painting area divided according to the divided colors; and
Establishing a node at the boundary
Lt; / RTI >
The image enhancement step includes performing histogram specification.
The histogram specification step,
Converting a color coordinate of each pixel of the enhanced image into a YCrCb color coordinate system;
Generating an actual histogram and an actual cumulative distribution function for illuminance which is a Y coordinate of the YCrCb color coordinate system for the enhanced image;
Generating an ideal histogram and an ideal cumulative distribution function for the illumination of the enhanced image,
Matching the ideal histogram and the ideal cumulative distribution function with the actual histogram and the cumulative distribution function to obtain a final histogram and a final cumulative distribution function, and
And converting a color coordinate of each pixel based on the final histogram.

4. The method of claim 3,
Obtaining the final histogram and the final cumulative distribution function,
Finding a value of an ideal cumulative distribution function (hereinafter, referred to as an “ideal function value”) closest to the actual cumulative distribution function of the gray level for each gray level,
Finding a gray level having the ideal function value in the ideal cumulative distribution function,
Finding a value of the grayscale having the abnormal function value in the ideal histogram (hereinafter, referred to as an “ideal histogram value”);
Generating the final histogram by assigning the abnormal histogram value to each of the gray levels, and
Obtaining the final cumulative distribution function from the final histogram.

5. The method of claim 4,
The coordinate conversion step includes converting the Y coordinate of each pixel to a gray level having the abnormal function value.

6. The method according to any one of claims 3 to 5,
The color separation step,
Converting the coordinate converted color coordinates for each pixel into an RGB color coordinate system;
Normalizing the RGB color coordinates of each pixel;
Calculating a distance between the standardized RGB color coordinates of each pixel and the standardized RGB color coordinates of the reference color; and
And replacing the normalized RGB color coordinates of each pixel with the standardized RGB color coordinates of the reference color at the closest distance.

The method of claim 6,
The reference color is a coating surface detection method, which is a color painted on the coating surface.

In claim 7,
The boundary extraction step uses a Laplace boundary, painting surface detection method.

9. The method of claim 8,
The node setting step includes setting a start point and an end point of the boundary.

delete

Photographing the painted surface,
Reinforcing the photographed image of the painted surface;
Distinguishing colors of the enhanced image;
Extracting a boundary of the painting area divided according to the divided colors;
Establishing a node at the boundary,
Moving the painting module according to the node, and
Painting the painted surface using the painted module
Including;
The image enhancement step,
Converting a color coordinate of each pixel of the enhanced image into a YCrCb color coordinate system;
Generating an actual histogram and an actual cumulative distribution function for illuminance which is a Y coordinate of the YCrCb color coordinate system for the enhanced image;
Generating an ideal histogram and an ideal cumulative distribution function for the illumination of the enhanced image,
Matching the ideal histogram and the ideal cumulative distribution function with the actual histogram and the cumulative distribution function to obtain a final histogram and a final cumulative distribution function, and
And converting the color coordinates of each pixel based on the final histogram.

12. The method of claim 11,
The color separation step,
Converting the coordinate converted color coordinates for each pixel into an RGB color coordinate system;
Normalizing the RGB color coordinates of each pixel;
Calculating a distance between a standardized RGB color coordinate of at least one reference color painted on the painted surface and a standardized RGB color coordinate of each pixel, and
And replacing the normalized RGB color coordinates of each pixel with the standardized RGB color coordinates of the reference color at the closest distance.