What’s New in Vision Builder AI 2009

Read about the new features in NI Vision Builder for Automated Inspection (AI) 2009, including algorithm and usability improvements.

Modeling System Variations

The functionality to model system variations provides a way to prototype and test the strength of visual inspections built using Vision Builder AI in a quick, automated manner. You can use this utility to prototype and help predict whether your machine vision application will function as expected when affected by a set of real-world circumstances.

To model system variations, you can start with a set of inspection images, and create a suite of hundreds or thousands of images by modeling changes in overall lighting, camera position, focus, motion blur, and noise. The inspection is then tested on the expanded library of images to determine whether there are any missed defects or false rejects, and the application can then be tweaked until the inspection runs as expected for a desired set of conditions.

The following are examples of real-world phenomena that can be modeled with this tool:

• Window is opened, changing the lighting conditions for the object under inspection
• Lens shaken out of focus by vibrations from nearby industrial equipment
• Camera is bumped, changing its angle, position, or distance relative to the object being inspected
• Conveyer belt speeds up, causing motion blur

[+] Enlarge Image

 

 

Enhanced Geometric Matching Step

The enhanced Geometric Matching step makes major improvements in detecting hard-to-find contours and occluded features, ensuring that your application matches patterns more accurately and misses fewer patterns, even if there are shadows or changes in the object’s texture. Geometric pattern matching is generally used to detect patterns in an image even when the objects under inspection are rotated or scaled.

Using an improved edge-based pattern matching algorithm, the algorithm uses a generalized Hough transform for matching curves found in the template image to the curves found in the target image, as well as edges and contours.

[+] Enlarge Image

The advantages of using the new algorithm include:

• All curves in the template image are directly used. Because no assumption is made about the underlying geometric structure of the object, this method offers a large improvement for matching patterns in objects without well-defined geometric features.
• All curves are used, so there is no need for you to specify what curves to use. This facilitates faster, lower-maintenance implementation.   
• You have the option to remove curves for performance benefits, which makes the new algorithm flexible and customizable.

New Classify Colors Step

With the Classify Colors step in Vision Builder AI 2009, you can create color templates and classify colors in designated regions of interest. The new algorithm makes this type of analysis possible, and uses the Hue-Saturation-Luminance (HSL) color information to learn colors and classify objects.

Color classification is the process of labeling a color based on previously learned colors, and is used as a machine vision tool to inspect products in industries such as automotive, food, textile, wood, and consumer personal products. You can use color classification in conjunction with other inspections to increase reliability of your machine vision application. Examples of color classification in the food and textile industries include:

• Determining if cookies have been baked properly based on their color (process control)
• Analyzing fruits and classifying the quality of the fruit based on color (quality control)
• Establishing the quality of cotton fabrics based on color
• Reading the color codes on electronic components, mechanical parts, or medical devices

Included with this feature is a new stand-alone utility for training colors. With the new Color Classification Training Interface, you can easily create color classifications and import them into Vision Builder AI, so that you can retrain your system for colors even with variations in lighting. Simply drag a region of interest (ROI) over a part of an image and save your newly classified color. 

[+] Enlarge Image

 

Color Image Support for More Image Processing Steps

The expanded native color image support allows you to use color images in more Vision Builder AI steps without forcing you to do conversion to a noncolor supported format. Supported image formats include RGB 32-bit, HSL 32-bit, and RGB 64-bit. 

The following steps now support color images:

Ethernet/IP Unconnected (Explicit) Messaging and New Adapter Class Support

Vision Builder AI 2009 now includes Ethernet/IP adapter class support in addition to explicit messaging support, which enabled Vision Builder AI to set and query tags on a PLC system. This new adapter class support now allows Vision Builder to asynchronously treat requests from a PLC, without affecting the cycle times of the inspection.

In addition to the benefit of being able to integrate more easily in an existing industrial automation system, making Vision Builder AI an Ethernet/IP adapter helps reduce the cycle time of an inspection because Vision Builder serves data in the background more efficiently.

To turn this feature on, the Ethernet/IP server must be started in the Communication Device Manager, and then the Ethernet/IP Adapter step can be used to write data to Vision Builder AI input assembly or read data from output assembly.   

When the Ethernet/IP server is started, two assemblies are allocated, one input and one output, which are used to communicate with a PLC. 

Vision Builder AI can also communicate with Ethernet/IP devices using unconnected messaging (or explicit messaging) for reading or writing tags.

Particle Filtering in Detect Objects Step

The Detect Objects step can now filter objects based on more than 80 shape parameters to produce more reliable and powerful results.

System-Level Spatial Calibration

Among the improvements to calibration in Vision Builder AI 2009 is the fact that calibration files are global and serve the entire Vision Builder AI system.  

This implies that:

• Once created, a calibration can be used in different inspections.
• Within an inspection, a calibration can be applied by multiple steps (improves memory usage).

Calibrations are listed in the System Resource Manager, and are associated with specific cameras. There is also an option to fail the inspection if an incorrect camera is detected by the calibration, and an option to set an expiry date, which will require an eventual recalibration to ensure that the system is precisely configured.

Additional features of this new functionality:

• Calibration can be applied within the acquisition step, streamlining the calibration process. You can define the calibration on the latest image acquired and correct the image from the acquisition step.
• Calibration can be set to expire for quality assurance purposes.
• Calibration can be set to fail the inspection if incorrect camera model is detected.
• Calibration axis and point distances can be updated at each iteration using the Calibrate Image step.

View Measurements Tool

With this new tool for viewing measurement values, you can easily select and monitor specific results, edit result values, and even get important statistical information on numeric results.

[+] Enlarge Image

 

Import States in State Diagram Editor

The state diagram editor now lets you reuse portions of existing code by importing states, including all associated steps, into your current inspection. 

[+] Enlarge Image

Reader Comments | Submit a comment »