Case Studies
« Back To Vision & Imaging Case Studies
Manufacturer of machine vision systems. We are a machine vision supplier from Japan with over ten years' experience, and more than 30,000 installations outside Japan. Our goal is to provide the best solutions from installation to support our customers. We have 6 types of main machine unit based on inspection goal. The same software runs into all models. Offering hardware compatibility for cameras for long term use. There are over 60 inspection tools including general purpose and specific application tools as a default in all VTV-9000 models with high processing speed. No programming required and easy to set up for beginners. ViSCO can handle ranging from VGA to 29M pixel camera.
Content Filed Under:
Industry:
Medical Devices
Medical Devices
Application:
Visual Inspection & Testing
Visual Inspection & Testing
Automated Optical Inspection of X-Ray Ceramic Matrices Using NI Image Processing Tools and LabVIEW
POSTED 07/08/2011
by:
"We completed and validated all of the image processing work for various models of ceramic matrices in just 16 weeks, which would have been impossible without the excellent vision development tools from National Instruments and LabVIEW to very quickly build a production-ready application."
– Anand Krishnan, Soliton Technologies Private Limited
The Challenge:
Developing an automated optical inspection (AOI) system for detecting, quantifying, and classifying surface defects on X-ray ceramic matrices (scintillator elements).
The Solution:
Using NI image processing tools to build a highly reliable AOI system that adapts to naturally varying shades of the ceramic matrices and lets the user change the defect detection and classification criteria through software.
Identifying visual defects in X-ray scintillator matrices is critical to the image quality produced by medical imaging devices that employ these elements. Previously, a group of trained operators manually inspected the surface by viewing the unit under inspection (UUI) under a microscope. This made defect identification and classification subjective, necessitating multiple inspections. Our customer wanted to automate this process with a system meeting the following requirements:
Flexibility – Capability to inspect various models of scintillator ceramic matrix elements
Reliability – Inspection had to meet strict requirements for reliability and repeatability
Networking – Permit configuration changes and statistical reviews on the inspection results from any PC within the customer’s network
Compact Form and Ergonomics – Efficient use of valuable manufacturing floor space with ease of use
We selected a 1,280 x 960 pixel Sony IEEE 1394 digital camera that came with many of the programmable features needed for this application. More than 12 parameters, including shutter speed selection and filter, could be configured from the application software.
We wrote the application software using the NI LabVIEW graphical development environment to give the customer the flexibility to configure the defect classification criteria, set up automatic e-mail alarms, and transfer data through FTP, among other features.
System Hardware
The inspection station consists of an enclosed imaging chamber and a fixture for the UUI, which moves in and out of the chamber on a pneumatic slide. We built the fixture in a modular fashion so the user can easily change the ‘holder’ for different matrix types in less than a minute.
Digital lines control the pneumatics and lighting and the UUI is manually loaded. The pneumatic slide takes the UUI into the imaging chamber where it is imaged. The image is processed and the results are displayed on the screen. The operator unloads the UUI after inspection and puts it into the appropriate bin.
System Software
The Sony IEEE 1394 camera provides features the customer can use to capture the optimal image for processing. The NI-IMAQdx driver provided easy access to all of the camera parameters and simplified the setup processes. By obtaining optimal image settings, we minimized the image processing challenges.
The application needed to identify and classify different defect types on a variety of surfaces. In the prototype phase of the project, we used the extensive libraries in the NI Vision Assistant. It provided a nice prototyping environment to identify the key image processing parameters that best differentiated each kind of defect. For instance, we realized that a combination of the extent of ‘dispersion’ in a discolored area of the UUI face (parameter 1) and the straightness of edges detected in the area of the discoloration (parameter 2) could help classify a particular kind of defect from another. The design phase also involved studying the experience of the trained inspection personnel and applying that knowledge to develop algorithms that could mimic the skilled inspector.
The advantage of using NI technology was the unifying nature of LabVIEW as the software platform. We easily translated the image processing algorithms developed during the prototype stage into production-ready application software. We also used LabVIEW to develop a user-friendly human machine interface (HMI) that interacts with the digital controls for sequencing the inspection process, generates reports, provides statistical process control (SPC) analysis tools, generates e-mails, and uses FTP. We used the e-mail and FTP capabilities of LabVIEW to build the remote debugging and updating feature into the design.
System Benefits
Table 1 describes the main benefits of the system.
Conclusion
We built an error-proof inspection system using the latest in virtual instrumentation and machine vision technology. Custom-built, adaptable algorithms ensured reliable inspection irrespective of the variations in the components. The rapid prototyping and batch processing features of the Vision Assistant proved invaluable in efficiently translating the visual inspector’s experience into quantitative parameters, resulting in an objective, repeatable, and error-proof inspection system. The system helped the customer reliably control quality at early stages of production. The wealth of data generated by the system provided feedback to improve the manufacturing process and quality. We are presently looking at upgrading the system to run on the NI Compact Vision System to provide further reliability and offer a more compact system.
The power of virtual instrumentation was evident in this application. Using NI drivers, the various controls of the pneumatic slide, lighting, and image acquisition from the camera were seamlessly integrated with the image processing routines. We completed and validated all of the image processing work for various models of ceramic matrices in just 16 weeks, which would have been impossible without the excellent vision development tools from National Instruments and LabVIEW to very quickly build a production-ready application.
Author Information:
Anand Krishnan
Soliton Technologies Private Limited
Soliton Technologies Pvt Ltd,#683, 15th Cross Road J.P. Nagar 2nd Phase
Bangalore
India
Tel: +91 (80) 4120-8600
"We completed and validated all of the image processing work for various models of ceramic matrices in just 16 weeks, which would have been impossible without the excellent vision development tools from National Instruments and LabVIEW to very quickly build a production-ready application."
– Anand Krishnan, Soliton Technologies Private Limited
The Challenge:
Developing an automated optical inspection (AOI) system for detecting, quantifying, and classifying surface defects on X-ray ceramic matrices (scintillator elements).
The Solution:
Using NI image processing tools to build a highly reliable AOI system that adapts to naturally varying shades of the ceramic matrices and lets the user change the defect detection and classification criteria through software.
Identifying visual defects in X-ray scintillator matrices is critical to the image quality produced by medical imaging devices that employ these elements. Previously, a group of trained operators manually inspected the surface by viewing the unit under inspection (UUI) under a microscope. This made defect identification and classification subjective, necessitating multiple inspections. Our customer wanted to automate this process with a system meeting the following requirements:
- Identify and classify various surface defects such as spots, cracks, voids, grind marks, dark bars, and chip offs
- Devise and record a quantitative measure for the severity of these defects
- Provide the capability to tune the image processing parameters to modify the defect classification thresholds
- Incorporate a software architecture for the vision algorithm developer to tweak algorithms, run a regression test, and update software from a remote site when new types of defects are found
Flexibility – Capability to inspect various models of scintillator ceramic matrix elements
Reliability – Inspection had to meet strict requirements for reliability and repeatability
Networking – Permit configuration changes and statistical reviews on the inspection results from any PC within the customer’s network
Compact Form and Ergonomics – Efficient use of valuable manufacturing floor space with ease of use
We selected a 1,280 x 960 pixel Sony IEEE 1394 digital camera that came with many of the programmable features needed for this application. More than 12 parameters, including shutter speed selection and filter, could be configured from the application software.
We wrote the application software using the NI LabVIEW graphical development environment to give the customer the flexibility to configure the defect classification criteria, set up automatic e-mail alarms, and transfer data through FTP, among other features.
System Hardware
The inspection station consists of an enclosed imaging chamber and a fixture for the UUI, which moves in and out of the chamber on a pneumatic slide. We built the fixture in a modular fashion so the user can easily change the ‘holder’ for different matrix types in less than a minute.
Digital lines control the pneumatics and lighting and the UUI is manually loaded. The pneumatic slide takes the UUI into the imaging chamber where it is imaged. The image is processed and the results are displayed on the screen. The operator unloads the UUI after inspection and puts it into the appropriate bin.
System Software
The Sony IEEE 1394 camera provides features the customer can use to capture the optimal image for processing. The NI-IMAQdx driver provided easy access to all of the camera parameters and simplified the setup processes. By obtaining optimal image settings, we minimized the image processing challenges.
The application needed to identify and classify different defect types on a variety of surfaces. In the prototype phase of the project, we used the extensive libraries in the NI Vision Assistant. It provided a nice prototyping environment to identify the key image processing parameters that best differentiated each kind of defect. For instance, we realized that a combination of the extent of ‘dispersion’ in a discolored area of the UUI face (parameter 1) and the straightness of edges detected in the area of the discoloration (parameter 2) could help classify a particular kind of defect from another. The design phase also involved studying the experience of the trained inspection personnel and applying that knowledge to develop algorithms that could mimic the skilled inspector.
The advantage of using NI technology was the unifying nature of LabVIEW as the software platform. We easily translated the image processing algorithms developed during the prototype stage into production-ready application software. We also used LabVIEW to develop a user-friendly human machine interface (HMI) that interacts with the digital controls for sequencing the inspection process, generates reports, provides statistical process control (SPC) analysis tools, generates e-mails, and uses FTP. We used the e-mail and FTP capabilities of LabVIEW to build the remote debugging and updating feature into the design.
System Benefits
Table 1 describes the main benefits of the system.
Table 1: Comparison Between Manual Inspection Process and Inspection System Based on NI ProductsConclusion
Conclusion
We built an error-proof inspection system using the latest in virtual instrumentation and machine vision technology. Custom-built, adaptable algorithms ensured reliable inspection irrespective of the variations in the components. The rapid prototyping and batch processing features of the Vision Assistant proved invaluable in efficiently translating the visual inspector’s experience into quantitative parameters, resulting in an objective, repeatable, and error-proof inspection system. The system helped the customer reliably control quality at early stages of production. The wealth of data generated by the system provided feedback to improve the manufacturing process and quality. We are presently looking at upgrading the system to run on the NI Compact Vision System to provide further reliability and offer a more compact system.
The power of virtual instrumentation was evident in this application. Using NI drivers, the various controls of the pneumatic slide, lighting, and image acquisition from the camera were seamlessly integrated with the image processing routines. We completed and validated all of the image processing work for various models of ceramic matrices in just 16 weeks, which would have been impossible without the excellent vision development tools from National Instruments and LabVIEW to very quickly build a production-ready application.
Author Information:
Anand Krishnan
Soliton Technologies Private Limited
Soliton Technologies Pvt Ltd,#683, 15th Cross Road J.P. Nagar 2nd Phase
Bangalore
India
Tel: +91 (80) 4120-8600
Vision in Life Sciences
This content is part of the Vision in Life Sciences curated collection. To learn more about Vision in Life Sciences, click here.