KR20140000559A - Vision recognition method and system thereof - Google Patents

Abstract

A method and a system for image recognition are disclosed. The image recognition method according to an embodiment of the present invention comprises the following steps of: photographing an object; collecting brightness values of pixels from the photographed object area; and determining the actual size of the object for the pixels based on the brightness values of the pixels. [Reference numerals] (AA) Step of determining the actual size of each pixel corresponding to the brightness value of each pixel which the object occupies using a least square method; (BB) Object size (pixel) = sum (size calculated using a brightness and size graph) x pixel resolution; (S100) Step of photographing an object; (S200) Step of collecting brightness values of pixels from the photographed object area; (S210) Step of determining the brightness value of each pixel based on a predetermined range of brightness values; (S300) Step of determining the actual size of the object

Description

Image Recognition Method and System {VISION RECOGNITION METHOD AND SYSTEM THEREOF}

The present invention relates to an image recognition method and system, and more particularly, to an image recognition method and system for determining the size of an object using a camera and lighting.

In general, an inspection system used for semiconductor and display processing equipment or various inspection equipment for inspecting the state of a micro area is based on the number of pixels (pixels) of a camera using a camera and lighting to detect fine or abnormal objects. Use a system to determine size.

As the mass production of recent products is gradually changed to high quality, the inspection equipment is required to have a function of detecting a defect of a smaller object, since the camera has a limitation of physical size (ie, cell size of CCD). The size measurement method based on the number of pixels has a limit on the accuracy of the inspection object.

That is, when the size of the object is recognized by the camera in an area of several tens of micrometers, pixels of an area having a brightness lower than a brightness boundary value which is set as a reference and a brightness boundary value set based on the determined brightness boundary value are determined. It is not treated as the area of the object, only pixels with high brightness are recognized as the area of the object.

In practice, the object is not delimited along the boundaries of the pixels having a certain pattern, but may span some area of the pixels. In this case, all internal regions of each pixel may be recognized or not all may be recognized according to a predetermined brightness boundary value.

Therefore, in the related art, the size of an object is recognized by determining the size of the object as an area of pixels having a brightness higher than the set brightness boundary value, and thus, the size of the object may vary depending on which brightness boundary value is used. In comparison, errors could occur.

That is, there is a problem that the actual size of the object or object recognized by the camera and the size of the object determined according to the brightness boundary value, which is the reference based on the image, are different from each other.

[Patent Document 1] Republic of Korea Patent Application No. 10-2009-0108726

Accordingly, an aspect of the present invention is to provide an image recognition method and a system capable of reducing fine errors of a substantial region of an object recognized by a camera and a recognition region of an object determined based on the image. will be.

According to an aspect of the invention, the step of photographing the object; Collecting brightness values of pixels from an area of the photographed object; And determining an actual size of the object with respect to the pixels based on the brightness values of the pixels.

Collecting brightness values of pixels from an area of the object may include determining a brightness value corresponding to each of the pixels from a range of predetermined brightness values.

The determining of the actual size of the object may include determining an actual size of each pixel corresponding to a brightness value of each pixel occupied by the object by using a least squares equation.

, x: 크기, y: 밝기)을 구하는 단계; 및 a, b에 관한 최소화 함수(

)를 구하는 단계를 더 포함할 수 있다.Determining the actual size of each pixel, the equation of brightness and size through the sample (

, x: magnitude, y: brightness); And the minimization functions for a and b (

It may further comprise the step of obtaining.

The determining of the actual size of the object may further include adding the actual size of each pixel occupied by the object.

The photographing of the object may include photographing the object using alternating transmission light and reflection light.

The photographing of the object may include photographing the object by using one of the transmission light and the reflection light; And photographing the object by using the transmission light and the reflection light at the same time.

According to an aspect of the invention, the photographing apparatus for photographing the object; A reflection illumination device installed at one side of the object to provide reflection illumination to the object; A transmission lighting device installed on the other side of the object to provide transmission light to the object; An illumination control device for operating the reflection lighting device and the transmission lighting device; And an image recognition device configured to determine an actual size of the object with respect to the pixels based on a ratio of brightness values of the pixels of the object and a reference brightness value photographed by the photographing apparatus. .

The lighting control device may include a cross lighting controller for selectively performing on / off for the reflective lighting device and the transmission lighting device; And an illumination signal encoder for providing on / off signals of the reflective lighting device and the transparent lighting device to the cross lighting controller.

The image recognizing apparatus may determine a brightness value corresponding to each pixel from a predetermined range of brightness values, and use the least-squares equation to actually apply each pixel corresponding to the brightness value of each pixel occupied by the object. And a computer having an image recognition program for determining the size and summing the actual size of each pixel occupied by the object.

According to the present invention, the number of pixels of an area occupied by an object according to brightness values of pixels collected and measured from an area of an object recognized by the camera is determined based on a ratio of brightness values of pixels to a reference brightness value. As a result, it is possible to reduce the fine error of the actual region of the object recognized by the image and the recognition region of the object determined based on the image.

1 is a flow chart of an image recognition method according to an embodiment of the present invention.
2 is an original image and a recognition image of an object for explaining an image recognition method according to an embodiment of the present invention.
3 is a graph illustrating a relationship between brightness and size of an object for explaining an image recognition method according to an exemplary embodiment.
4 is an exemplary image of an object for explaining an image recognition method and a determination image for determining the size of the object according to an embodiment of the present invention.
5 is a block diagram of an image recognition system according to an exemplary embodiment.
FIG. 6 is a signal transmitted from an illumination encoder to a cross lighting controller and reflected from a cross lighting controller when the photographing apparatus of the image recognition system according to an exemplary embodiment of the present invention and an object of the reflective lighting apparatus and the transmissive lighting apparatus are photographed. This is a timing diagram of the signal transmitted to the lighting device and the transmission lighting device.
FIG. 7 is a diagram illustrating a photographed image and an image of a photographed object of the photographing apparatus of the image recognition system according to an embodiment of the present disclosure, separately according to an illumination method.

In order to fully understand the present invention, operational advantages of the present invention, and objects achieved by the practice of the present invention, reference should be made to the accompanying drawings and the accompanying drawings which illustrate preferred embodiments of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Like reference symbols in the drawings denote like elements.

1 is a flowchart of an image recognition method according to an embodiment of the present invention, FIG. 2 is an original image and a recognition image of an object for explaining an image recognition method according to an embodiment of the present invention, and FIG. 4 is a graph illustrating a relationship between brightness and size of an object for explaining an image recognition method according to an embodiment of the present disclosure, and FIG. 4 is an example image of an object for explaining an image recognition method according to an embodiment of the present invention. The discrimination image determines the size.

Referring to FIG. 1, an image recognition method according to an embodiment of the present disclosure may include photographing an object (S100), collecting brightness values of pixels from an area of the photographed object (S200), and And determining an actual size of the object with respect to the pixels based on the brightness value (S300).

 Referring to FIG. 2, an object gradually decreasing in area along a diagonal direction of the pixels 10 arranged in a grid is illustrated, and an image obtained by photographing the object 20 with a camera is illustrated.

Comparing the object 20 and the photographed image 30 of FIG. 2, the real object 20 is gradually narrowed in an area along the diagonal direction of the lattice arrangement, while the photographed image 30 of the object is The area is shown as it is, while the brightness gradually darkens.

That is, the left image of FIG. 2 is an original image showing objects smaller than the size of the camera CCD pixel. If the size of the pixel is 10um x 10um, the objects are smaller than that. When the image is acquired by the camera as shown in the left image of FIG. 2, the larger the size of the object, the closer to 255, and the smaller the size of the object, the closer to zero. Therefore, if each pixel is in a category smaller than the size of one pixel, the brightness can be measured.

In this case, the range of the brightness value is 0 to 255, and the reference brightness value as a reference may be 255. Typically, the range used for contrast values from white to black is 0 to 255.

Based on this, the brightness of the object with respect to the size of the object in each pixel 10 arranged in a grid may be represented by a graph of increasing relationship, as shown in FIG. 3. Derivation of the graph shown in Figure 3 will be described later in detail in the step of determining the actual size of the object. Referring to FIG. 3, since the brightness and the size of the object are mutually increased, the specific gravity of the real region occupied by each object in the pixel may be obtained from the brightness of each pixel of the captured image.

4 illustrates an object photographed in a pixel area and a brightness value of each pixel of the photographed object.

2 and 4, the collecting of brightness values of the pixels 10 from the region of the object 20 (S200) may include brightness values corresponding to the respective pixels 10 from a range of predetermined brightness values. Determining step (S210) may further include.

At this time, the step of determining the actual size of the object 20 (S300) includes the step of determining the actual size of each pixel corresponding to the brightness value of each pixel occupied by the object using the least-squares equation.

The graph shown in FIG. 3 is a graph showing the brightness and the size of an object in one pixel. By using such a graph, the size of an object whose object outline is outlined (outline) is smaller than the size of one pixel or pixel. It can be measured accurately.

In order to obtain such a graph, it is obtained by using the Least Square Equation using N pieces of sample data obtained through samples having respective inherent sizes. In this case, assuming that the least-squares equation is a quadratic equation passing through the origin, the graph shown in FIG. 3 may be represented by Equation 1.

(x: size, y: brightness)

Equation 1. Equation of brightness and magnitude obtained through the sample

Here, the least-squares method provides a method of obtaining a polynomial by minimizing measurement error in order to obtain a relationship between values obtained based on experience.

Equation 2 below shows a process for obtaining a and b.

Equation 2. Minimization function for a, b

Equation 3. Minimize function on a

Equation 4. minimize function for b

 Equation 3 and Equation 4 can be used to derive a and b of Equation 2, through which the relationship between brightness and magnitude can be represented by a graph as shown in FIG. For example, the actual size of the object may be determined using the brightness value of the pixel as shown in FIG. 4.

That is, the object recognized by the camera is illustrated in the leftmost grid region shown in FIG. 4, and the brightness value of the pixels occupied by the object is described in the central grid region. The brightness values of the pixels in the grid region may be substituted into Equation 1 to determine the size of each pixel, and the size of the object may be determined by adding them all together. In this case, the number 1 written in each pixel indicates that the object occupies all the areas of the pixel, and 0.8 indicates 80% of the pixel area. Accordingly, the area of the object according to the brightness value of each pixel may add the number described in the pixel, that is, measure the actual size of the object from the number of pixels.

In this case, the pixel 10 filled with white in the area of the object 20 captured corresponds to the brightness value 255, and the pixel filled with black outside the area of the object 20 is the brightness value 0. According to the region filled by the object in the region of the photographed object 20, the pixel has a brightness value of 0 to 255.

According to the relationship between the brightness of the object derived from the image recognition method according to an embodiment of the present invention and the number of pixels, the specific gravity of the object in the pixel area can be calculated from the following equation.

Actual size of the object = SUM (calculated size using graph of brightness and size) ⅹ area for one pixel (Pixel Resolution)

In the above equation, SUM (calculated size using a graph of brightness and size) corresponds to the sum of the pixel sizes described in the area occupied by the object, that is, the number of pixels occupied by the object.

In addition, as described above, the area of one pixel may have a size of 100 micro square meters when the micrometer unit, for example, one pixel is 10 μm × 10 μm.

At this time, the size of the actual object and the size of the pixels constituting the recognition image of the object may be determined based on the ratio of the brightness value of the pixels and the reference brightness value.

In addition, since the pixel has a lattice size in micro units, the area of the object can be calculated by the number of pixels to calculate the area of the photographing object that is proportional to the area of the real object. With distance, the size of the object can be calculated accurately.

The image recognition method according to the present invention can be used not only to detect the minute error generated in the process of performing various processes for the display panel or the semiconductor wafer in the range allowed by the unit pixel of the imaging device, but also other biotechnology It can also be used to test objects in chemical, chemical or material fields.

Shooting of the object is possible by providing an illumination to the object, by photographing the illuminated object with a camera arranged in a number suitable for the area of the object, the step of photographing the object for the image recognition method according to the invention (S100), It may include the step of photographing the object using the transmission light and the reflection light alternately.

In the past, a technique of providing a single reflection light or a transmissive light to an object and storing an image by a camera on a computer has been used. The contrast of the object due to the reflected light and the transmitted light is recognized to the contrary, and even in the specific gravity of the area occupied by the object in the pixel, minute differences may occur depending on the characteristics of the object.

Therefore, the step (S100) of photographing the object according to the present invention adopts a photographing method that selectively uses two methods of reflecting light and transmission light, rather than adopting any one method to provide illumination according to the characteristics of the object. By providing a method, the variation in brightness with respect to the object can be reduced, and the actual size of the object in the area occupied by the object can be accurately calculated.

To this end, the step (S100) of photographing the object according to an embodiment of the present invention, photographing the object using any one of the transmission light and the reflection light, and using the transmission light and the reflection light at the same time to take the object It may include the step of photographing.

In such a case, the edge of the object can be saved by illumination of two methods or a mixture of both methods, so that an accurate brightness value with less variation can be obtained, and the size of the area occupied by the object can be accurately calculated based on this. .

5 is a block diagram of an image recognition system according to an exemplary embodiment.

Referring to FIG. 5, an image recognition system according to an exemplary embodiment of the present disclosure includes a photographing apparatus 100 for photographing an object, a reflection illumination device 120 installed at one side of the object, and providing reflection illumination to the object; And a light control device 130 installed at the other side of the object to operate the transmission lighting device 122 to provide the transmission light to the object, the reflection lighting device 120 and the transmission lighting device 122, and the photographing device 100. The image recognition device 140 determines an actual size of the object with respect to the pixels based on a ratio between the brightness value of the pixels of the object and the reference brightness value.

The lighting control device 130 includes a cross lighting controller 131 for selectively performing on / off of the reflective lighting device 120 and the transparent lighting device 122, and the reflective lighting device 120 and the transparent lighting device ( An illumination signal encoder 132 for providing the on / off signal of the 122 to the cross lighting controller 131 may be included.

The image recognition device 140 may determine a brightness value corresponding to each pixel from a predetermined range of brightness values, and determine a brightness value of each pixel corresponding to the brightness value of each pixel occupied by the object by using a least square method. The computer 142 may include an image recognition program installed to determine an actual size and to add an actual size of each pixel occupied by an object.

At this time, the range of brightness values is 0 to 255, and the reference brightness value as a reference is 255.

Image recognition of the object using the image recognition system according to the present invention is connected to the plurality of cameras 115, the computer 142 is installed with an image recognition program for automatically performing the above-described image recognition method to determine the captured image Is performed by.

Referring to FIG. 5, in FIG. 6, when photographing an object by the photographing apparatus 100, the reflection lighting apparatus 120, and the transmission lighting apparatus 122 of the image recognition system according to an embodiment of the present disclosure, A timing diagram of the signal transmitted from the illumination signal encoder 132 to the cross lighting controller 131 and the signal transmitted from the cross lighting controller 131 to the reflective lighting device 120 and the transmission lighting device 122 are shown. The photographing of the camera 115 proceeds along the timing diagram of the cross lighting controller 131.

Referring to FIG. 6, when an encoder pulse having a predetermined amplitude and period is applied from the illumination signal encoder 132 to the cross lighting controller 131, the reflection lighting apparatus 120 may be set by the highest and lowest values of the pulses. Or periodic power pulses (pulse 1, pulse 2) for operating the transmission lighting device 122 are generated from the cross lighting controller 131 and transmitted to the reflection lighting device 120 and the transmission lighting device 122 side. According to the pulse, the reflection illumination device 120 and the transmission illumination device 122 are operated to provide the corresponding illumination to the object, and photographing of the object according to the illumination method is performed.

FIG. 7 is a diagram illustrating a cross-section of a photographed image and an image of a photographed object according to an illumination method of the image recognition system according to an embodiment of the present invention, separately according to an illumination method.

Referring to the photographed image of the O-type object illustrated in FIG. 7, a photographed image of the O-type object in which black and white brightness is crossed using alternating reflection and transmission lights is shown on the left side. The black part of the captured image of the O-type object is an image using reflection light, and the part shown in white is an image using transmission light.

In addition, the upper right of FIG. 7 illustrates an O-type object separated by an image of reflected light, and the lower right of FIG. 7 is an O-type object separated by an image of transmitted light.

When photographing using pulse 1 generated from the cross lighting controller 131 of FIG. 6, a transmission image in which the O-type object is shown in white is generated, and when photographing using the pulse 2, the O-type object is black. A reflection image shown as is generated.

As shown in FIGS. 5 to 7, the image recognition system according to the present exemplary embodiment may acquire a mixed image in which brightness is intersected with respect to the same object by using reflected light and transmitted light alternately. It is possible to separate the objects corresponding to and obtain two brightness values with less deviation from the pixels for each object. Based on this, the number of pixels of the area occupied by the object can be derived by the above-described calculation method of the image recognition method. have.

That is, in the image recognition system according to the present embodiment, the transmission lighting device 122 and the reflection lighting device 120 are alternately photographed, but the transmission lighting device 122 and the reflection lighting device 120 operate simultaneously. Since the photographing may be made, the scope of the present invention is not limited to the photographing method using the transmissive lighting device 122 and the reflective lighting device 120 alternately, and the transmissive lighting device 122 and the reflective lighting device 120. ) May be included in any case.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. Accordingly, such modifications or variations are intended to fall within the scope of the appended claims.

100: recording device
115: camera
120: reflective lighting device
122: transmission lighting apparatus
130: lighting control device
131: cross lighting controller
132: illumination signal encoder
140: video recognition device
142: computer

Claims (10)

Photographing the object;
Collecting brightness values of pixels from an area of the photographed object; And
And determining the actual size of the object with respect to the pixels based on the brightness values of the pixels. The method of claim 1,
The step of collecting the brightness value of the pixels from the region of the object, determining the brightness value corresponding to each of the pixels from a range of a predetermined brightness value. 3. The method of claim 2,
The determining of the actual size of the object may include determining the actual size of each pixel corresponding to the brightness value of each pixel occupied by the object using a least squares equation. The method of claim 3,
Determining the actual size of each pixel,
Through the sample, the equation of brightness and magnitude (

, x: magnitude, y: brightness); And
Minimization function for a, b

The image recognition method further comprises the step of obtaining. The method of claim 3,
The determining of the actual size of the object further comprises adding the actual size of each pixel occupied by the object. The method of claim 3,
The photographing of the object may include photographing the object by using transmission light and reflection light alternately. The method of claim 3,
Photographing the object,
Photographing the object using any one of the transmitted light and the reflected light; And
And photographing the object by using the transmission light and the reflection light at the same time. A photographing apparatus for photographing an object;
A reflection illumination device installed at one side of the object to provide reflection illumination to the object;
A transmission lighting device installed on the other side of the object to provide transmission light to the object;
An illumination control device for operating the reflection lighting device and the transmission lighting device; And
And an image recognition device configured to determine an actual size of the object with respect to the pixels based on a ratio of brightness values of the pixels of the object and a reference brightness value photographed by the photographing apparatus. 9. The method of claim 8,
The illumination control device includes:
A cross lighting controller for selectively performing on / off for the reflection lighting device and the transmission lighting device; And
And an illumination signal encoder for providing on / off signals of the reflection lighting device and the transmission lighting device to the cross lighting controller. 9. The method of claim 8,
The image recognizing apparatus may determine a brightness value corresponding to each pixel from a predetermined range of brightness values, and use the least-squares equation to actually apply each pixel corresponding to the brightness value of each pixel occupied by the object. And a computer provided with an image recognition program for determining the size and summing the actual size of each pixel occupied by the object.

Images