KR20170069178A - Multiple optics vision inspection system - Google Patents

Abstract

The present invention relates to a multi-optic module vision inspection system, and more particularly, to a multi-optical module vision inspection system, which includes a dustproof device having an air cylinder structure for vibration reduction and a stage unit transfer module and a mechanical part for sample products having various shapes and sizes in the field of semiconductor materials and display materials In the inspection equipment, there is an effect of detecting surface defects (foreign objects, scratches, pattern errors, etc.) of the inspection target product by using optical illumination and vision inspection system and judging whether or not the product is defective, thereby reducing defective products in the production process.

Description

{Multiple optics vision inspection system}

The present invention relates to a multi-optic module vision inspection system, and more particularly, to a multi-optical module vision inspection system, which is equipped with an air cylinder-type vibration damping device for vibration reduction and a stage unit transfer module In order to detect surface defects (foreign objects, scratches, pattern errors, etc.) on the inspection target products by using the optical illumination and vision inspection system in the inspection equipment including the mechanical part, And an optical module vision inspection system.

The entire market for global machine vision inspection systems and components is divided into three categories: optical lighting (LED or halogen lighting), high-resolution line scan cameras with multi-cameras, area scanners, frame grabbers, The market consists of vision board (hardware) for motion control, vision image processing software for industrial PC, vision inspection system, step motor and step motor driver, and sensor related parts. The five-year CAGR of 10% is expected to lead to rapid growth in the machine vision industry.

Korea's semiconductors, LCD and PDP display industries account for 47% of the global LCD market of Samsung Display and LG Display in LCD and PDP flat panel displays as one of the major industrial strategy export items that support the Korean economy. , Localization of LCD inspection equipment and vision system is required.

LCD and PDP display inspection equipment and vision system for inspection and measurement are used by major corporations in advanced countries such as USA, Canada, and Japan. In Korea, It does not meet the required standards.

In large-scale production of LCD and PDP display panels, large-area display inspection depends on the manual, and micro-precision measurement is required. Therefore, in the factory automation process, Inspection by automation process Due to the difficulty of visual inspection Machine vision combined with robot software and machine software technology and large-area display equipped with computer vision system Detection and pattern inspection equipment and vision inspection system on the surface Inspection equipment and vision Development of inspection system is becoming an important task.

In Korea, the introduction of machine vision is slower than advanced countries such as the United States and Japan. However, in recent years, the application of machine vision has been rapidly increasing in the industrial field and the technology of high-quality machine vision products .

Machine vision (MV) is a fusion of software technology of robot, machine and vision system. It is used for inspection of product surface defects in various fields from factory automation (FA, factory automation) It is used as inspection equipment and vision inspection system to filter PCB inspection, LED chip package, and other product defects. It is used to inspect and measure defects on the surface of mass-produced products to reduce defective products.

Large-area LCD display inspection requires inspection of front panel of LCD panel rather than partial sampling of large-area LCD display screen. Resolution of several tens of micrometers is required, especially in mass production process, A high-speed image processing system that processes a large amount of data in real time is required when image processing is performed. Since a large amount of data of the image itself is required and various operations must be performed in order to extract an image feature to be used, processing and defect detection algorithms to reduce product defects and vision inspection systems.

It is necessary to develop a 20 μm high-precision vision inspection system for a large-area display panel equipped with an inspection equipment for detecting surface defects of products such as semiconductor materials, packages, and displays, which are mass-produced by a manufacturer, and a vision inspection system.

Vision inspection systems are mostly large in scale and costly to modify the inspection environment after completion as well as in the initial production because expensive module is used.

However, the vision inspection system such as the LCD inspection equipment is required to revise the inspection environment inevitably due to many variables such as changes in the requirements of the customer, even if the system is constructed before proceeding to the test.

In order to cope with this case, it is necessary to produce a vision inspection equipment considering the fact that a system capable of adjusting the inspection environment from the beginning can be adjusted to prevent a high cost loss to be required for a subsequent apparatus modification.

Conventional vision inspection system has a disadvantage that it is difficult to inspect materials with different properties and it is possible to make it within a narrow range even if the inspection environment is changed according to the characteristics of one inspection object. And there is a problem that inspection can not be performed if the size of the product to be inspected varies greatly.

It is an object of the present invention to solve the problems of the prior art, and it is an object of the present invention to provide a vibration damping device of an air cylinder structure for vibration reduction and a stage unit transfer module and a mechanical part for a sample product having various shapes and sizes in the field of semiconductor materials and display materials A multi-optic module vision inspection system to detect surface defects (foreign scratches, pattern errors, etc.) on the inspection target products by using optical illumination and vision inspection system in the inspection equipment and reduce defective products in the production process .

In order to achieve the object of the present invention, a multi-optical module vision inspection system is provided with a mechanism part in which an outer case is housed, two computers and a monitor on the front side of the mechanism part, and a keyboard mouse on the lower part, An optical illumination unit disposed at a lower end of the camera unit; an inspection stage positioned at the lower end of the camera unit and the optical illumination unit and on which the subject is placed on the XY-stage; A frame for supporting the left and right sides of the vision inspection equipment, and a stage unit transfer module for controlling the XYZ position movement,

The camera unit is connected to a single PC via a camera interface, and each PC is connected to a LAN via a network hub and a TCP / IP And the main server computer is connected to an encoder / distributor through a LAN and TCP / IP through a network hub, and the encoder / distributor is connected to a multi-camera constituting a line scan camera or an area scan camera Connected,

The camera unit includes multi-cameras constituting a line scan camera or an area scan camera. Each camera has a high-resolution 10 to 100 μm pixel size It uses the TDI camera which can measure,

Wherein the mechanical part further comprises a subject transfer robot for placing an inspection stage (XY stage) by a loader, wherein the mechanical part is an object to be inspected transferred to a conveyor belt of a production line in a factory automation process, And an ID reader that reads the DPM code attached to the product or the ID of the product placed on the inspection stage (XY stage) by the loader or the ID reader that reads the 1D or 2D code and transmits the product ID detected by the computer .

The multi-optical module vision inspection system according to the present invention is applicable to a sample product having various shapes and sizes in the field of semiconductor materials and display materials, an anti-vibration device of an air cylinder structure for vibration reduction, an inspection unit (Such as foreign objects, scratches, pattern errors, etc.) to the object to be inspected by using optical illumination and vision inspection system, and it is judged whether the object is clean or not, thereby reducing the defective product in the production process.

This product is used for LCD, AMOLED, display film and next generation display which are mass produced in large display companies. It is used in large area display material field by using high resolution camera to detect foreign objects, scratches, Precision type vision inspection system for a large-area display panel equipped with an inspection apparatus for detecting surface defects of a product such as a precision measurement and a vision inspection system.

The multi-optic module vision inspection system can be used as a surface defect inspection device on the product line in the factory automation process of large display companies in Korea by commercializing it as a 10 ~ 20㎛ ultra-precision vision inspection system for display panel inspection, , Taiwan AUO, and BOE in China, and has developed the vision inspection equipment. We have developed overseas marketing strategy to secure quality certification, patent registration, overseas export market, , Establishing a cooperative system of marketing, and establishing a practical network. In accordance with marketing 4P and STP strategy, it has the effect of pioneering the domestic and overseas market by marketing by product positioning and service positioning.

It is a vision inspection system that is integrated with a line scan camera or an area scan camera equipped with an optical illumination and a high resolution multi camera. It is equipped with a large area display material Stereo Vision, 3D Vision, which detects product surface defects by comparing and analyzing foreign objects, scratches, pattern errors, defective position precise measurement, normal images and inspection images in patterns of 10 ~ It is a basic technology for establishing a smart factory by creating new industries of embedded vision using smart camera in the future by fostering localization of vision inspection system and image processing software technology, fostering R & D manpower, securing intellectual property rights and securing domestic technology competitiveness. Is used.

FIG. 1 is a block diagram showing an example of a system for setting product information, inspection information and defect information, loading a product by a loader, and displaying images from a high resolution multi-camera constituting a mechanical alignment, a vision alignment, FIG. 5 is a flow chart showing processes of acquiring, image processing, reporting result data, detecting defects, unloading a product by a loader, detecting surface defects, foreign objects, scratches, and pattern errors of a product.
FIG. 2 is a photograph showing a sample or PCB having various shapes and sizes such as a wafer / chip / package semiconductor material, a TSP production in a display material field, and a glass / panel in a cover glass production process.
Fig. 3 is a front view and a side view of the appearance of the multi-optical module vision inspection equipment.
4A is a perspective view (ISO View), a back view, a front view, and a side view of the external structure of the multi-optical module vision inspection device.
4C is a front view, FIG. 4E is a left side view, FIG. 4F is a right side view, FIG. 4G is a rear view, and FIG. 4H is a bottom view.
5A is a perspective view (ISO View), a back view, a front view, and a side view of the internal structure of the apparatus in the detailed design of the multi-optical module vision inspection apparatus according to the present invention.
FIG. 5B is an enlarged view of an internal view of the mechanism of the multi-optical module vision inspection apparatus (ISO View).
6 is a view showing the layout of the internal structure of the mechanism in the detailed design of the multi-optical module vision inspection equipment.
7 is a perspective view of the internal structure of the apparatus.
8 is a diagram showing the hardware specifications of the linear servo motor, the robot X, the robots Z1, Z2 and Z3, and the motion controller.
9 is a diagram showing product specifications of a CMOS image sensor and a lens used as an embodiment of the multi-cameras used in a line scan camera.
Fig. 10 is a diagram showing product specifications of coaxial illumination and oblique illumination and LED light source used as optical illumination of the vision inspection system. Fig.
11 is a diagram showing product specifications of an area scan camera and a lens.
12 is a block diagram of a multi-optical module vision inspection system according to an embodiment of the present invention.
13 is a vision inspection system configuration diagram when the LED optical illumination and the line scan camera of the multi-optical module vision inspection system are used.
14 is a view showing an image processing process of a vision inspection system for acquiring an image of a surface of a product from a line scan camera multi-camera in a multi-optical module vision inspection system and inspecting defects on the product surface such as foreign objects, scratches and pattern errors.
FIG. 15 is a diagram comparing (x, y) points on an image plane photographed by a camera with image arrays.
FIG. 16 is a flowchart illustrating an LCD defect detection auto search algorithm according to an embodiment of the present invention.
17 is a view showing a display panel pattern defect.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an example of a system for setting product information, inspection information and defect information, loading a product by a loader, and displaying images from a high resolution multi-camera constituting a mechanical alignment, a vision alignment, FIG. 5 is a flow chart showing processes of acquiring, image processing, reporting result data, detecting defects, unloading a product by a loader, detecting surface defects, foreign objects, scratches, and pattern errors of a product.

FIG. 2 is a photograph showing a sample or PCB having various shapes and sizes such as a wafer / chip / package semiconductor material, a TSP production in a display material field, and a glass / panel in a cover glass production process.

□ Function

1) Precise shooting of display panel and semiconductor samples / products (vision inspection)

2) Defect detection and judgment

3) Optical camera and lighting control

4) Inspection stage unit feed control

5) Alarm function in case of emergency

□ Features

- Can be tested on various products

- Compatible with line scan and area scan cameras

- Testable for linear motor feed rates up to 400 mm / s

- Linear motor minimizes image errors due to speed error during line scan inspection

- Produced as a system that can cope with coaxial illumination, tilt, and transmission illumination

- Sample size can be up to 400 mm.

Fig. 3 is a front view and a side view of the appearance of the multi-optical module vision inspection equipment.

4A is a perspective view (ISO View), a back view, a front view, and a side view of the external structure of the multi-optical module vision inspection device.

4C is a front view, FIG. 4E is a left side view, FIG. 4F is a right side view, FIG. 4G is a rear view, and FIG. 4H is a bottom view.

1. Exterior design

- Instrument scale of multi-optic module vision inspection equipment (mm): 2663 x 1756 x 2000

Mechanics Configuration: Optical Illumination Module, Feed Module, Server / Client PC (Image Processing and Review)

5A is a perspective view (ISO View), a back view, a front view, and a side view of the internal structure of the apparatus in the detailed design of the multi-optical module vision inspection apparatus. FIG. 5B is an enlarged view of an internal view of the mechanism of the multi-optical module vision inspection apparatus (ISO View).

It has two external computers and a monitor on the front, and a keyboard mouse on the bottom. In one embodiment, the four cameras provided in the cable duct are fixed to a support, the LCD panel of the inspection stage is moved by a linear motor, and the structure of the camera is such that two frames are bent. In another embodiment, If necessary, the camera can be designed to be movable in X, Y, Z so that it can be moved up and down and left and right.

FIG. 6 is a layout of the internal structure of the apparatus in the detailed design of the multi-optical module vision inspection apparatus, and FIG. 7 is a perspective view of the internal structure of the apparatus.

In the multi-optical module vision inspection equipment, the exterior frame is enclosed through the optical illumination module, the conveying module, and the electric part of the retractable door, and the internal module of the PC is controlled through the PC input / .

A multi-optical module vision inspection device includes a mechanism portion in which an outer case is housed, two computers and a monitor on the front side of the mechanism portion, and a keyboard mouse on the bottom portion, and a line scan camera or an area scan camera An optical illumination unit disposed at a lower end of the camera unit; an inspection stage positioned at the lower end of the camera unit and the optical illumination unit, the inspection stage being placed on the XY-stage; a base on which the inspection stage is placed; A frame supporting the left and right sides of the vision inspection equipment, and a stage unit transfer module controlling the XYZ position movement,

The camera unit is connected to a single PC via a camera interface, and each PC is connected to a LAN via a network hub and a TCP / IP And the main server computer is connected to an encoder / distributor through a LAN and TCP / IP through a network hub, and the encoder / distributor is connected to a multi-camera constituting a line scan camera or an area scan camera Respectively.

The line scan camera uses a camera link when a frame grabber is used, and the area scan camera uses a large amount of Gigabit Ethernet (GigE).

The mechanism portion may further include a subject transfer robot for placing the inspection stage (XY stage) by a loader.

The mechanical unit may be configured such that an ID of a product placed on a test stage (XY stage) is detected by a DPM code attached to the product, or a DPM code attached to the product, And an ID reader for reading the 1D or 2D code and transmitting the product ID detected by the computer.

Referring to FIG. 7, the stage unit transfer module uses a linear motor of X, Y, Z1, Z2, and Z3 axes to minimize image errors due to velocity errors during line scan vision inspection. The effective stroke length on the Y axis is 1400 mm, And a linear robot equipped with a loader for transferring the object to be inspected to an inspection unit so that the object can be photographed by the optical camera. The linear motor is inspected at a feed rate of up to 400 mm / s, The transport frame was constructed using stone as a stone rather than as a steel material.

The stage unit transfer module is characterized in that the line scan camera moves in the X-axis direction by the linear motor, the position of the inspection stage moves in the Y-axis direction by the linear motor, and the camera moves in the vertical direction of the Z1, Z2 and Z3 axes .

The linear motor is operated by a motor driving unit interlocked with a control computer, and uses a PID control to control the speed and moving position of the motor.

The vibration damping system is a vibration damping system using an air cylinder structure at the bottom of the grid below the base to reduce vibrations in precision measurement of a vision inspection system using a line scan or area scan camera for precise measurement of 10 to 100 μm pixel size That is, a stone quartz panel and an isolator are provided to prevent external vibration and prevent internal vibration.

2. Detailed design

▷ It shows the optical module structure which can change the inspection conditions flexibly according to the characteristics of the object to be inspected.

- Up to 9 units can be equipped with line scan / area cameras as many as required

- Moving camera up / down / left / right

- The size of test object is designed to be up to 400 mm maximum width.

▷ Stage unit feed module for minimizing image error

- By adopting linear motors of X, Y, Z1, Z2, Z3 axis, it minimizes image error due to speed error during line scan inspection.

- Y axis: a linear robot having an effective stroke length of 1400 mm and having a loader for transferring an inspection object to an inspection section so that the inspection object can be photographed by an optical camera

- Inspection of the linear motor at a feed rate of up to 400 mm / s

- In order to reduce the deformation due to external environmental factors such as temperature, the conveyance frame is constructed using stone rather than steel material.

▷ System structure capable of lighting by type according to characteristics of test object

- Designed as a system that can cope with coaxial / oblique / transmissive lighting

- Jig can be manufactured by optical illumination type.

8 is a diagram showing the hardware specifications of the linear servo motor, the robot X, the robots Z1, Z2 and Z3, and the motion controller.

The line scan camera or the area scan camera is equipped with a multi camera, and one PC connected to each camera through a control client camera interface (frame grabber, GigE) is required. One server PC .

○ Video data synchronization / matching technology

- Vision device inspection synchronization program: The information obtained through the camera in a certain inspection area acquires the image data through the PC connected to each camera, and transmitted to the main server computer through the network hub, The defect of the surface of the product is detected by comparing the image data with the error image data.

- Development of network communication program: In this process, we developed LAN and TCP / IP network communication program (TCP / IP sokcet program) through network hub and developed a database processing program for final image data storage.

In this process, it is important to input accurate information about the location of the image software and the actual position of the image information according to the location of the hardware in the process. Based on this, the error pattern of the product surface image Position, scratch position, spreading position, etc. Finally, the product defect is detected through image processing.

9 is a diagram showing product specifications of a CMOS image sensor and a lens used as an embodiment of the multi-cameras used in a line scan camera.

Fig. 10 is a diagram showing product specifications of coaxial illumination and oblique illumination and LED light source used as optical illumination of the vision inspection system. Fig.

11 is a diagram showing product specifications of an area scan camera and a lens.

Table 1 shows the results of Vision1 (Top Inspection) for inspecting print, crack, chipping, pinhole, foreign body, LOGO, ICON, IR and Camere hole using coaxial illumination at the upper part and glass 1), Vision2 (upper inspection 2) for inspection of scratches, impacts and foreign objects using reflective illumination at the top, Vision3 (back inspection) for scratching, It shows the inspection environment of Vision4 (backside inspection) which checks printing, cracking, and chipping by using illumination and coaxial illumination at the bottom.

Considering the characteristics of the inspection sample, we constructed an environment in which various arrangements of cameras and lighting were applied, and conducted test tests, so that the following conditions were selected and an optimal inspection solution was obtained to obtain more stable and accurate images .

- Adjust various resolutions, magnification and angle of camera and lens

- Adjust the distance, angle, or illumination between camera and sample

- Coaxial, transmission and reflection optical systems

- Line scan camera / Area scan camera selection

 Optical illumination is a process of testing two types of illumination for a sample fixed in a darkroom where one light source is blocked by using one area camera, and checking the image of the surface of the product as a result of the test. do.

Multi-optic module vision inspection system provides flexibility of inspection environment, which is different from the conventional vision inspection system. It is possible to change the inspection environment flexibly, The setting range of lighting conditions is wide and can be easily adjusted.

- Can change the type and number of optical cameras

- Moving up / down / left / right of the optical camera: WD (Working Distance), FOV (Field of View)

- Can change the position and location of lighting

- Irradiation angle of light source, irradiation method (coax / transmission / reflection), illumination adjustment

12 is a block diagram of a multi-optical module vision inspection system according to an embodiment of the present invention. 13 is a vision inspection system configuration diagram when the LED optical illumination and the line scan camera of the multi-optical module vision inspection system are used.

For 20-level for the display panel area high precision vision inspection system 10 to fast display the manufacturing process time and the non-precision measurement by using the inspection system light and vision inspection optical module, fault detection algorithm in the defects in the display panel in the range ¹¹ 100㎛ We developed an integrated control system with an automatic checking function and an algorithm for detecting defects on the images of products obtained by using an optical camera. Finally, a defect detection (foreign object, scratch, pattern error, etc.) And a vision inspection system was developed to reduce the defects of products.

Ultra-high precision vision inspection system for large-area display panels is designed for ultra-precise inspection for large-area display panels with external case having anti-vibration structure for vibration reduction to be applied to other machine vision inspection equipment besides LCD, PDP, Vision Inspection System It is designed and manufactured by the mechanism, and it is manufactured by vibration reduction vibration prevention structure at the lower part of the grid beam, and developed the mechanism part and the motor drive / control part for ultra precision drive and consists of the multi camera inspection part, motor drive part, exterior case, And a controller for controlling the camera. The line scan camera or the area scan camera includes a high-resolution multi-camera capable of precise measurement, and an XY position driver for precise driving of the multi-camera. By linking the information with the hardware Synchronization / matching / video image processing with the processing / fault detection algorithm (image processing) and software with fault detection algorithm has been designed and developed to Lee.

The multi-camera driving unit has a mechanical characteristic that the base panel and the multi-camera part are moved separately by using the linear motor and the servo motor as a whole, so that the XY position transfer mechanism, the reducer, Motor control board.

A series of progressive line scan cameras or area scanners with optical illumination (LED or halogen illumination) installed on the inspection stage and a high resolution multi-camera (TDI camera) that is 100 times more sensitive than conventional CCD cameras in multi- By using a vision inspection system that includes a camera, I / O and motion control of a vision board (control board) mounted on an industrial PC drives the stepper motor driver and motor, By adjusting the speed and position of the motor, machine vision, camera synchronization, image matching, image processing and defect detection algorithms of vision inspection system, Match the images of the camera and extract the features of the normal image and the measured test image (Feature Extraction) By comparing the feature values of the normal image and the inspection image, it is possible to detect defects such as foreign matter, scratches, and pattern errors by precisely measuring foreign objects and defects at a pattern of 10 to 100 μm or less in vision recognition on the surface of the display, Thereby determining whether or not the product is defective.

The PC-based machine vision inspection system is equipped with an optical vision, a camera, a camera interface (Frame Grabber, Gigabit Ethernet (GigE), IEEE1394, Camera Link, USB 3.0), a vision check equipped with a vision board The system detects image, scratch, and pattern errors on the surface of the product through image processing for inspection of acquired images by vision inspection image processing software.

 Vision inspection image processing software provides frame grabbing, imgae processing, feature extraction / defect detection and control functions and interlocks with line scan camera or area scan camera equipped with multi camera The computer vision inspection system is used to check the surface defects of the product and compare the feature values of the normal image and the inspection image by the software with image processing and defect detection algorithm to detect foreign objects, The defect is detected and the defective position is measured to detect the product defect.

Since the image obtained by using the line scan camera is large in the TFT-LCD image obtained from the industrial PC through the camera interface, a fast image processing time is required for the application in the production process. The TFT LCD image has a repetitive pattern. The LCD defect region is easily detected using the image difference from the adjacent pattern.

The Vision Inspection System has been increasing in the market for interfaces such as GigE and CoaXpres for processing high-resolution, high-speed, high-volume data at high speed for micro-precision measurement. Especially, Gigabit Ethernet (GigE) frame grabber ⁴ 4 Video signal obtained from existing video media such as TV, video camera, etc. can be converted into a digital signal that can be processed by a computer, System, which lowers system construction costs and improves performance. Table 2 is a quantitative target item for multi-optic module vision inspection equipment.

- Quantitative goal items Evaluation items
(Key performance
Spec ^{One )} ) unit All items
Occupying
importance ^{2 )}
(%) World-class (owned /
Holding company) Domestic level before R & D Development target evaluation Way ^{3 )} Performance level Performance level Primary
year Secondary
year Third
year 1. Detection Defect Size ㎛ 30 Japan / Hugh Brain
/ 30㎛ 30 30 25 10 Self-assessment 2. Defect detection rate % 20 Israel/
Orbotech /
85% 80 80 90 95 Self-assessment 3. Robot position accuracy ㎛ 10 Israel/
Orbotech / 1㎛ 5 5 3 One Self-assessment 4. System Vibration Rate Hz 5 USA / August
/ 10Hz 15 15 10 10 Self-assessment 5. 2D Optical System Resolution ㎛ 5 Germany / Basler
10 탆 10 10 5 3.5 Self-assessment 6. 2D and check rate % 10 Israel/
Orbotech / 5% 10 10 5 5 Self-assessment 7. Inspection speed (2D) mm / s 5 Israel/
Orbotech / 50mm / s 50 50 60 80 Self-assessment 8. Stage Flatness ㎛ 5 Japan / Technos / 20㎛ 20 30 20 20 Self-assessment 9. 3D Optical System Vertical Resolution ㎛ 10 Japan / Nikon /
0.001 탆 0.1 0.1 0.01 0.005 Self-assessment

o Element technology of 20th grade high-precision vision inspection system for large-area display panel (Example)

- Accurate measurement of foreign objects and defects in patterns of 10 to 100 μm or less in vision recognition

- Establishment of 20 μm high-precision vision inspection system for mechanical parts and large-area display panels using optical illumination and high-resolution optical camera.

For example, according to the characteristics of an object placed on an inspection stage (X-Y stage) on the surface of a large-area display, an arrangement of optical illumination, a high-resolution camera and a light source and a mechanism are designed and fabricated to establish an optimum inspection environment.

- Development of 2D inspection optical module: Multi-optical module vision inspection system for large area display panel is equipped with a series of Progressive Line Scan Camera or Area Scan Camera with high resolution multi camera (TDI camera) for optical illumination and precise measurement. By using the motion control program of the inspection system, I / O and motion control of the vision board (control board) mounted on the industrial PC drives the step motor driver and the motor to feed and control the XY position, By combining camera vision, image synchronization and vision detection technology with SW technology using machine vision and defect detection algorithms, the surface of a display with various shapes and sizes Foreign matter, scratches, and pattern errors of a pattern of 10 to 100 μm or less are detected, Wheat measured to determine product yangbul.

- Development of Geomagnetic Robot Transfer Mechanism for Vibration Reduction: Design and manufacture of vibration reduction vibration mechanism and transfer geomagnetic robot for minimizing vision recognition error of ultra-high resolution camera

- Development of Vision System Mechanism and Motor Drive / Control Unit for Large-Area Ultra Precision Inspection: Design and manufacture of ultra-precision motor drive device and control program for precise driving of multi-camera

- Development of registration / synchronization program for multi-image information processing: Software design with image processing and defect detection algorithm with synchronization / matching / image processing / defect detection algorithm for multi-camera image information interworking with hardware And development

- real-time measurement result, line efficiency analysis (yield analysis) monitoring GUI function implementation: vision Real-time measurement monitoring function during inspection and process capability monitoring function such as yield rate and utilization rate

1) ID reader (DPM code attached to the product, or 1D or 2D code reading): Inspection stage (XY stage) by the inspected object transferred to the conveyor belt of the production line in the factory automation process, or by the loader of the transfer robot ) And transmits the product ID detected by the computer.

2) Optical illumination: In vision inspection systems, optical illumination uses white or red LED illumination, or halogen illumination with fiber optic guides, and LEDs are used in embodiments of the present invention.

When using LED lighting, use a ring LED and lighting controller consisting of a series of white LEDs and a lighting controller, or many LEDs surrounding the camera lens.

To illuminate the precise illumination for machine vision, we use array light type for line or area micro scan, bar light for line scan camera, ring light for area scan camera, Back light type of the periphery - bottom rear illumination of the inspection stage, peripheral illumination of the subject. High-quality LED lighting uses DC power to illuminate the camera's precise illumination, so there is no blinking and the single-color aberration in the lens is much smaller than white light, so accurate measurement is possible.

Optical illumination can also use fiber optic guide halogen illumination instead of LED illumination. For reference, halogen lighting is fabricated from structured light in the form of a line of uniformly bright light from a light source through an optical fiber.

The optical illumination is adjustable so that the brightness of the image captured by the high-resolution multi-camera can be seen clearly by the brightness of the illumination itself.

High-speed scanning for vision inspection High-resolution camera (TDI camera capable of precise measurement of 10 ~ 100μm) when the light is oscillated by optical illumination or the brightness of the LED is too weak in micro-unit precision illumination Obtained image is blurred Vision inspection image Processing software error occurs.

3) Camera: Line scan camera or area scan camera is equipped with multiple cameras, each of which has a high-resolution 10 to 100 μm (pixel size) precision measurement Use a possible TDI camera.

If precise measurement is not required, a general CCD camera or CMOS image sensor (CIS, CMOS Image Sensor) with lower resolution than TDI camera can be used.

In recent years, CMOS image sensors have been available with TDI camera class capable of precise measurement with pixel size of 10 ~ 100㎛ or less, so they can be used according to the requirements of inspection equipment.

The camera performs precise measurement shooting with a pixel size of 10 to 100 μm or less for semiconductor materials, LCD, AMOLED, display film, and LED package, and is connected to each PC through a camera interface and performs image processing Scratches, pattern errors, and defect occurrence positions of patterns of 10 to 100 μm or less (pixel size).

Line scan camera or Area scan camera: Uses a high-resolution, high-resolution, high-speed CCD camera (TDI camera) suitable for a conveyor belt system operated by a server motor in a manufacturer's production line in industrial factory automation processes. The specifications, structure and arrangement of the vision camera are designed to have a multi-camera structure. The multi-speed line scan camera is interlocked with the computer through the camera interface. The precision of the 2D vision inspection is 100 times Precise shooting of high-quality inspection images using high-resolution TDI cameras capable of precise measurement up to 10 ~ 20μm pixel size.

The line scan camera or area scan camera uses a high resolution TDI camera capable of precise measurement up to 10 ~ 20μm unit, which is 100 times more sensitive than general CCD camera.

The camera interface that connects the camera to the PC uses a frame grabber, Gigabit Ethernet (GigE), IEEE 1394a, IEEE 1394b, camera link, and USB 3.0.

Table 3 shows the machine vision camera interface.

division Gigabit Ethernet (GigE)     IEEE 1394a     IEEE 1394b   Camera Link Cable length 100m 4.5m (up to 72m when repeater is used) 4.5m (up to 72m when repeater is used) Up to 15m Bandwith 100 Mbps 32 Mbytes / s 64 Mbytes / s 255Mbytes / s (base configuration)
680 Mbytes / s (full
configuration) Bit rate 1000 Mbps 400 Mbps 800 Mbps > 2000 Mbps Standard GigE Vision Standard IEEE 1394 Trade
Association DCAM
Standard IEEE 1394 Trade
Association DCAM
Standard AIA Camera Link Standard Interface board Gigabit Ethernet board IEEE 1394a board Special frame grabber Maximum number of cameras No limit 16 (DCAM) 16 (DCAM) Per frame grabber
2 units Plug & Play ability Yes Yes Yes No cable Industrial and consumer Industrial and consumer Industrial and consumer Industrial

The line scan camera or the area scan camera provided in the present invention inspects an image of a product moved on a conveyor belt on a fixed support (Example 1). On the other hand, the inspection product is fixed And the line scan camera is moved (Example 2).

4) Frame Grabber: The digital image obtained from each high resolution camera (TDI camera) of line scan camera or area scan camera is transferred to the computer with PCIe card (digital I / O) And detects defects on the display surface through image processing for vision inspection by image processing software of the vision inspection system.

Each camera is connected to a respective frame grabber, and each frame grabber is connected to a respective computer.

One frame grabber can be connected to one computer or two computers depending on the type of product.

For example, if you are using Gigabit Ethernet (GigE) instead of a frame grabber for camera and computer interfaces without using a frame grabber, each high-resolution camera in a line scan camera or area scan camera will capture the captured image Each connected to a computer equipped with a Gigabit Ethernet board via Gigabit Ethernet (GigE), which transmits at high speed. In this case, the separate vision image processing software includes a frame grabber function.

Recently, instead of using a frame grabber as a camera interface, it is a tendency to use most of the device bit Ethernet (GigE) for high-speed transmission.

5) Encoder

In the factory automation process of the industry, the optical lighting and the high resolution multi camera are fixed on the support on the conveyor belt system operated by the step motor driver and the step motor driven by the computer. In a situation where a line scan camera or an area scan camera is installed,

The encoder measures the exact feed amount of the servo motor when operating the conveyor belt in the factory production line.

In the first embodiment, in the factory automation process of driving the production line with the conveyor belt of the industry, the number of pulses per revolution of the encoder is x ppr, the distance that the encoder makes one revolution is the moving distance per actual pulse signal, , The line rate is 18 kHz, and the line rate of the conveyor belt system does not exceed the maximum line rate of the camera.

* Poor inspection of the product of the vision inspection system (when the conveyor belt is used)

Preventing vibration and preventing internal vibration. Line Scan Camera for Precise Measurement of 10 ~ 100μm Pixel Size A vibration control system using an air cylinder structure was used at the lower end of the grating for vibration reduction suitable for a high speed inspection system using a camera.

14 is a view showing an image processing process of a vision inspection system for acquiring an image of a surface of a product from a line scan camera multi-camera in a multi-optical module vision inspection system and inspecting defects on the product surface such as foreign objects, scratches and pattern errors.

Multivision pattern inspection module configuration: Vision camera specifications, structure and layout are designed to have a multi camera structure as shown in the figure. Multi camera is interlocked and one PC is connected to each camera. Designed to perform the task with a minimum of total process time by accepting

Multicamera image measurement method - Matching of actual position and image information: Each camera of the multi-camera is synchronized to obtain image information at the same time, and the obtained image information is transmitted to a computer responsible for each image, After the image is sent back to the computer, the image is collected as one piece of image information per one panel, and then the normal image and the measured image are compared with each other, and the result of the vision check is output and the device moves to the next measurement position.

 Therefore, the positional alignment between the camera image information and the precision motor drive system is an important design factor. Through this process, the scratch, spread, and pattern errors of the product surface are compared with the previously inputted information, .

Machine vision is a fusion technology of vision, image processing software of robot, machine and vision inspection system. It is a technology that automation inspection is performed in online / in-process or online / post-process in industrial factory automation process. The part is identified by ID detector under optical illumination, and image data synchronized in a multi-camera equipped with a line scan camera or an area scan camera is imaged by a vision inspection system of a computer (2) image binarization, (3) image processing, (4) image analysis, and (5) image interpretation, the position of defects such as foreign objects, scratches, And defects such as the positions of the holes on the parts, the number of holes, and the presence or absence of holes are analyzed to detect defective products.

In addition, a robot automation system is applied to the machine vision system, and information about the orientation of the part is mainly used when the end device picks up the part in a particular way. The vision system can provide information and data to properly grip and drive the gripper to pick and place the position and orientation of the product. The position of the part to be inspected is used to locate the object randomly placed on the x, y coordinates. The vision system should measure x and y distances and match the center of the camera with the center of the randomly placed components.

The multi-camera constituting the line scan camera or the area scan camera must be free of vibration when acquiring images of the surface of the display product in the micro-precision measurement. Due to the shaking of the multi-camera, which requires precise micro-scale measurement, which constitutes the line scan camera or area scan camera at the time of vibration, the inspection image of the product surface is distorted and blurred on the micro-unit inspection stage (XY stage) , The pattern error is enlarged due to shaking, an accurate measurement error of the defect position occurs, and the product is judged as an error by the normal state.

Machine vision is a visual processing algorithm for vision inspection of connected computers through a camera interface (Frame Grabber, Gigabit Ethernet (GigE), IEEE 1394a, IEEE 1394b, Camera Link, USB 3.0) Scratches and patterns on the surface of the product through image processing [(1) Image acquisition, (2) Image binarization, (3) Image processing, (4) Image analysis, Error, and defects such as positional measurement of defects as shown in Fig. 15 are analyzed to filter defective products.

FIG. 16 is a flowchart illustrating an LCD defect detection auto search algorithm according to an embodiment of the present invention.

17 is a view showing a display panel pattern defect.

For example, in a TFT-LCD manufacturing process, a TFT-LCD defect detection algorithm (Defect Cell Extraction for TFT-LCD System) that automatically detects a defect position and size on the surface of an LCD display and detects a defect, Rmax and Rmin images are generated using gray-scale morphological processing to correct for sampling errors that may exist in the image.

If the normal image (reference image) and the measured image (test image) are monochrome images, D (x, y) is also a monochrome image. If the reference image and the test image are color images, D (x, y) is also a color image. In this D (x, y) image, the defective portion has a relatively large value, and the defective portion has a value close to zero.

- Find reference point using image feature pattern

A block matching algorithm is a method that uses a feature template image (a block pattern including a feature point to be searched in an image to be inspected), which is registered in advance, when the image to be inspected is given, Similar image patterns can be found. Among the block matching reference functions, typical block matching reference value functions include Mean Absolute Difference (MAD) and Mean Square Error (MSE). Robust shape matching criterion (AND operation) can be used for binary images without using MAD and MSE. MAD and MSE are comparative reference values applied to gray images, but in order to apply the AND reference value, all images are binarized to detect defects.

Experimental results for Laplacian edge detector and Sobel edge detector for edge detection, respectively, show that the Sobel operator has less noise than the Laplacian edge detector, regardless of the edge strength.

In the vision inspection system, image processing algorithm basically processes with gray image. Recently, various algorithms applying HSV model to fine image discrimination such as banknote discrimination or glass piece color line star are studied. In one embodiment, each Plane (H, S, V, HS) of an HSV image as well as each Plane (R, G, B) of an RGB image can be selected. In order to increase the accuracy of the vision check, Gray Scale Auto, which applies Otsu's threshold determination method [(gray conversion, filtering, histogram, histogram smoothing, threshold applied clustering / binarization, morphology operation, labeling, automatic sorting, feature output, Selection function detects display surface defects in gray image during image processing.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, It will be understood that various modifications and changes may be made without departing from the scope of the present invention.

Claims (7)

A camera unit having a high-resolution multi-camera having a line scan camera or an area scan camera on a fixed support base, a camera unit having a display unit, An optical illumination unit provided at the lower end of the camera unit; a test stage positioned at the lower end of the camera unit and the optical illumination unit and placed on the XY-stage; a base supporting the inspection stage; A stage unit transfer module for controlling the stage unit transfer module,
The camera unit is connected to a single PC via a camera interface, and each PC is connected to a LAN via a network hub and a TCP / IP And the main server computer is connected to an encoder / distributor through a LAN and TCP / IP through a network hub, and the encoder / distributor is connected to a multi-camera constituting a line scan camera or an area scan camera Connected,
The camera unit includes multi-cameras constituting a line scan camera or an area scan camera. Each camera has a high-resolution 10 to 100 μm pixel size It uses the TDI camera which can measure,
Wherein the mechanical part further comprises a subject transfer robot for placing an inspection stage (XY stage) by a loader, wherein the mechanical part is an object to be inspected transferred to a conveyor belt of a production line in a factory automation process, And an ID reader that reads the DPM code attached to the product or the ID of the product placed on the inspection stage (XY stage) by the loader or the ID reader that reads the 1D or 2D code and transmits the product ID detected by the computer Multi-optic module vision inspection system. The method according to claim 1,
Wherein the camera interface for connecting the line scan camera and the area scan camera to the PC uses Gigabit Ethernet (GigE), IEEE 1394a, IEEE 1394b, Camera Link, and USB 3.0. The method according to claim 1,
The camera unit may include a multi-camera constituting a line scan camera or an area scan camera, and each camera may be a general CCD camera with a lower resolution than the TDI camera or a CMOS image sensor (CIS, CMOS An image sensor) is used for the vision inspection system. The method according to claim 1,
The stage unit transfer module
It minimizes image errors due to velocity error during line scan vision inspection using linear motors of X, Y, Z1, Z2 and Z3 axes. Effective stroke length of Y axis is 1400mm. And a loader. The linear motor is inspected at a conveying speed of up to 400 mm / s. In order to reduce deformation due to external environmental factors such as temperature, the conveyance frame is made of a non-steel stone Wherein the multi-optical module vision inspection system comprises: 5. The method of claim 4,
The stage unit transfer module
The line scan camera is moved in the X-axis direction by the linear motor, the position of the inspection stage is moved in the Y-axis direction by the linear motor, and the camera section is moved in the vertical direction of the Z1, Z2 and Z3 axes. Inspection system. 6. The method of claim 5,
Wherein the linear motor is operated by a motor driver interlocked with a control computer and uses PID control to control the speed and movement position of the motor. The method according to claim 1,
The mechanical unit is provided with an optical module capable of flexibly changing the inspection conditions according to the characteristics of the object to be inspected. It can mount a line scan camera or an area scan camera as many as necessary up to a maximum of nine units, and can move the camera up and down and left and right And the size of the object to be inspected is designed to be up to a maximum width of 400 mm.

Images