Posted by Nancy Kays on Tue, Apr 24, 2012 @ 02:04 PM
Gerald Budd, president of Phoenix Imaging, defined this Custom-Tailored Solution for the Automotive Industry.
Phoenix Imaging, a Detroit-based machine vision integrator, has teamed with PPT to define a custom-tailored solution, with the flexibility to integrate vision hardware and software to meet the specific requirements of the automotive industry. These applications are often written in C++ or C# and output data in special formats, using a large tool set for rapid machine vision algorithm development.
During the development phase of this project, Phoenix Imaging performed an integration project for one of largest automotive companies in the world. Like most automotive companies that have used machine vision, this customer was looking for a “global” approach, with a reliable solution that could be implemented in their factories worldwide. For this application, a custom, small-format industrial computer, coupled with PPT Vision software, was developed.
PPT Vision and Phoenix Imaging specialize in surface inspection applications that isolate some type of abnormality on a machined or stamped metal surface or on other materials, including glass, plastic and die-castings
The image above shows the presence of porosity on a face of a cylinder head. The porosity is above the specified limit, but it is located outside of the sealing region. A human inspection would need to lay down a template to determine the location of porosity and then determine if it is above the limit using another see-through gauge. The machine vision system can easily find the component datum and overlay the inspection mask.
The image below illustrates the critical seal regions and the non-sealing region (shown in red) which has less stringent defect size requirements. An added benefit of the system is that it can measure the actual core sizes and their locations with respect to the datum.
The machine vision system can perform this inspection is a fraction of second and on 100% of the components before they are released to the next machining operation. If a defective component is found at this point in the manufacturing process, it can be removed from the product line, saving the cost of finishing a defective product. 
The image to the right indicates that the defect cluster was outside of the sealing region and would not cause any performance issues. This application represents one of the many uses of the surface Inspection algorithms to isolate typical surface abnormalities in automotive components. It is very effective in isolating many types of surface imperfections and will ensure that only the highest quality components pass through to the customer.
This custom-tailored approach provides seamless, cost-effective integration with automotive assembly processes and operating systems to conform to corporate specifications. Added benefits include improved productivity and quality, as well exceptional system performance – guaranteed.
Posted by Brian Robertson on Wed, Apr 18, 2012 @ 09:30 AM
When you are looking at the technical specifications of any machine vision camera, the CCD or CMOS imager is always referred to by its “inch-based sensor format” or “optical format.” You will see references like a 640 x 480 CCD in a 1/3” format chip or a high resolution camera that uses a 2/3” imager.
You may wonder where this inch-based reference originates because, if you look closely at the physical chip size dimensions, there is no apparent correlation between the nomenclature and the physical size.
For example, let’s look at a 1/2” imager in this image.
If you look at the diagonal measurement (7.99 mm = 0.314 inches), that’s nowhere near 0.500” or 1/2”. Where does this reference originate?
According to the internet sources noted below, this “antiquated” naming convention originates in the 1950s with Vidicon camera tubes.
The approximate optical format is defined as the diagonal length of the sensor (in mm) divided by 16. The result is expressed in inches (not converted to inches!). For instance, a 6.4 x 4.8 mm sensor has a diagonal of 8.0 mm and therefore an optical format of 8.0/16 = 1/2". The reason why it is expressed in inches is historical: a standard size one inch vidicon tube has only 16 mm of useful imaging area. Therefore 1 inch optics became 16 mm, 1/2 inch optics became 8 mm, etc.
Luckily, the camera and optics manufacturers are on the “same page” when it comes to this designation, even though it’s a bit confusing. In addition, if the detailed specs are available, one can always check the imager’s diagonal size versus the lenses’ Maximum Imager Diagonal allowed.
If you’d like to learn more about the camera imager sizes and optics that PPT Vision has to offer, and M-Series cameras, please visit pptvision.com.
Internet resources…
http://en.wikipedia.org/wiki/Optical_format
http://www.dpreview.com/news/0210/02100402sensorsizes.asp
Posted by Steve Maves on Fri, Dec 30, 2011 @ 04:46 PM
Fastener manufacturing companies and their customers face a common challenge, how to insure that a process is fed with all the same fasteners. If the wrong fastener is mixed into a shipment and fed into an automated assembly machine, it can cause more than just damage and rework on one part, it can also damage the tooling of the assembly machine, causing a longer period of downtime, and extensive repairs. Even a handful of these incidents can justify the cost of an inspection system, not to mention the intangible costs like loss of customer goodwill.
PPT Vision has extensive experience in the fastener inspection industry. PPT Vision smart cameras and software are used to measure various important parameters, including overall width and length, thread pitch, and head diameter. These insure high quality, rejecting any fasteners of the wrong type for 100% of the fasteners delivered, and run at the high rates typical of the fastener manufacturing industry.
The software is set up to inspect parts on an indexing dial machine for part sorting. The machine feeds and seperates individual parts. The machine includes a PC based operator interface so that users can see the products under inspection. It includes operator screens that show all measured parameters for each part, and engineering screens to allow changes in tolerance, and calibration of the system in either English or Metric units. Inspections can be setup and modified in the same operator interface that monitors the production. The hardware consists of a minimum of a single Smart Camera with a back lighted image of the part showing in silhouette as shown in the image below. Additional views, such as a top view provide information like head diameter and head type.
If you have fastener inspection challenges, please contact us, we can help.
Posted by Steve Maves on Tue, Dec 13, 2011 @ 04:36 PM
PPT Vision has had a Technical Support Self-Help Forum on the web, here...
http://forums.pptvision.com/
The PPT Vision forum is a place where our end users and partners can interact with one another, and PPT Vision staff about problems that they face, and find answers to common questions. This forum is moderated by PPT Staff, but contains contributions from everyone.
The forum has been going strong for more than 5 years, and has hundreds of useful articles for common topics. I'll mention a couple of interesting posts here.
Using the X-Bar charting capability of our CPM User Interface software...
http://forums.pptvision.com/index.php?topic=336.0
Changing Recipies on the Vision System...
http://forums.pptvision.com/index.php?topic=268.0
You can also search the forums for topics including; tips and tricks for using various vison tools, communication with 3rd party devices, and How-Tos covering a wide variety of topics.
Note, you have to join the forum in order to contribute articles, or download the sample programs that are attached to many of the posts. Please visit the forums, click on "Register", and join our online community today!
Posted by Nancy Kays on Fri, Dec 02, 2011 @ 01:31 PM
Datalogic S.p.A. has agreed to acquire PPT Vision. Bob Heller, president and CEO of PPT Vision, shares the news in his post below:
I am writing to share our news that Datalogic S.p.A., a globally recognized leader in automatic data capture and industrial automation and provider of bar code readers, data collection mobile computers, RFID and vision systems, has agreed to acquire PPT Vision. This marks a giant leap forward in realizing PPT’s goal to deliver world-class machine vision solutions and extraordinary customer service.
We are excited that Datalogic has recognized PPT Vision’s leadership position in the machine vision industry for nearly 30 years. Under the terms of the acquisition agreement, PPT will continue to operate as a separate company and as a center of excellence for machine vision within Datalogic Automation, the Industrial Automation division of Datalogic Group. Our headquarters will remain in Minneapolis, and our PPT team will continue to grow to develop new business opportunities with our customers around the world.
The partnership with Datalogic also signals a strategic focus to define and develop cutting-edge technologies and applications that respond to the growing needs among our customers for our products, from sophisticated sensors and smart cameras to embedded vision systems and custom-tailored solutions. We have already begun to team with Datalogic on innovative new products that will debut early in 2012 to evolve into an industry-leading product roadmap.
Most importantly, we are committed to provide our customers throughout the world with the highest level of technical support and customer service, delivered at the local level. Our global network of strategic business partners, distributors and integrators will expand in the months ahead to support the growing needs of organizations like yours and to deliver world-class machine vision solutions – guaranteed.
Next year marks the 30th anniversary of PPT Vision, and there is much to celebrate. I am pleased to continue to lead PPT Vision as we move forward, together with our loyal customers, to a bright future and a bold new brand: PPT Vision, a Datalogic Company.
See full annoucement...
http://pptvisionnews.blogspot.com/2011/12/ppt-visions-pending-acquisition-by.html
Posted by Brad Quist on Fri, Sep 02, 2011 @ 04:08 PM
The Circle Gauge Tool in PPT VISION's VPM software lets you find the center of a circle very accurately. Because it combines a large number of sub-pixel results, it averages out the errors associated with individual measurements.
In the image below, the circle Region of Interest (ROI) is searching for the transition of dark pixels to bright pixels from the outside of the circle to the inside. Once the edge points are determined, a best-fit line is drawn through those points. Then, based on other tolerances defined in the tool, the center point of the circle is calculated as a best-fit circle and drawn on the image.
The circle ROI can be adjusted over just a fraction of the circle and the tool will still calculate the best-fit center of the circle.
The image below shows only a part of the circle within the Field of View (FOV), with the circle ROI drawn on the image. The calculated center point X-Y coordinate is actually outside the FOV, but this is not a problem for the software. Even though this center point is not displayed on the image, it can still be used for measurements and tolerances can be placed on these measurements.
Posted by Brian Robertson on Fri, Aug 19, 2011 @ 03:46 PM
One of the most unique and powerful features of IMPACT software is the ability to work with lists of data. When performing image analysis, inspection locations can sometimes be numerous and data sets can be lengthy. In this case, the IMPACT list handling tools can be very useful and efficient.
The primary list handling tools are the List Sorter, List Loop, Data Instance and Data Set tools.
List Sorter
The List Sorter, as the name implies, sorts the incoming data by the selected Sort Feature. Incoming data can be a Blob, Point, Integer, or Real list. The corresponding Sort Features are Blob Features (area, height, width, etc), X or Y point value, Integer value, or Real Value. The output of this tool is a Sorted List.
List Loop
The List Loop tool can “step through” a list while doing an operation at each step. For example, you might want to step through a Centroid List, and, at each centroid point, execute another vision tool to inspect a feature. The List Loop tool makes this implementation much simpler, since you can use one tool and move it to many locations to do the same inspection, rather than adding many duplicate tools.
Data Instance and Data Set
The Data Instance tool is used to store values in a list, which can then be added using the Data Set tool.
The example below demonstrates all of these list handling tools. The goal of the inspection is to determine the internal radius for each pair of tabs in each hole.
Determine the Radius
To determine the radius of a pair of tabs, a Circle Gauge tool is used. But, rather than adding forty circle gauge tools, we will use one tool, which will be moved to each of the holes, executed, and the radius value will be stored.
The complete toolset is described below (the number in the graphic below corresponds to the detail in the bulleted list that follows.)
- Blob tool - finds the centroid of each of the forty holes and creates a centroid (point) list. These centroids will be used to position a Circle Gauge tool at each hole.
- List Sort - puts the centroids in Y-order direction.
- Data Set tool - clears the previous result (from tool 8), the list of Radii.
- List Loop tool - steps through the centroid list. The tools inside the loop execute for every element in the list.
- Data Instance - uses the current centroid list element (as the List Loop tool steps through them) to create an Origin (x, y, and angle).
- Circle Gauge - finds the internal radius of the tabs in each hole.
- Data Set tool - adds the current radius value to the final list of radii.
- Data Instance - stores the list of radii for all the holes.
The final result is the list of 40 radius values as shown in the Data Instance tool property tab below.
Summary
In summary, IMPACT software is very efficient at working with list data from vision algorithms. The List Sorter, List Loop, Data Instance, and Data Set tools can easily manipulate and step through list data to facilitate repeated operations.
To learn more about the PPT VISION IMPACT software, you can download a free version of the software here.
Posted by David Wick on Fri, Jul 22, 2011 @ 11:26 AM
Minneapolis, MN—July 11, 2011—PPT VISION, here, today announced the addition of two high-performance JAI cameras to its M-Series embedded machine vision system line, including an ultra-violet (UV) camera for inspections that benefit from this shorter wavelength of light, as well as a remote-head “industrial endoscopy” model for special applications. Additionally, the release of the IMPACT™ 10.2 software supports the growing M-Series vision system family with streamlined functionality and expanded image management capabilities.
Read the rest of the Press Release Here...
Posted by Brad Quist on Mon, Jun 27, 2011 @ 04:47 PM
Here’s an overview of today’s embedded machine vision systems, which appeared in Control Engineering magazine. With inspection speeds that can exceed 200 images per second, these systems are a perfect fit for pharmaceutical, food and beverage, electronics and other discrete manufacturing applications.
Posted by Chris Walkowicz on Fri, Jun 03, 2011 @ 03:33 PM
As I promised in an earlier blog posting, here is a tip for increasing the robustness and decreasing the process time of the Pinpoint Pattern Find® tool. The goal here is to search for the white box which encapsulates the data matrix code in the sample images.
Please note that I am using images from the IMPACT Software Suite sample programs. These images have names starting with “Sample 1D 2D code.”
There are a number of ways to find the white box surrounding the data matrix code. In this example I will use a pattern find tool to locate the white square and ignore the actual data matrix code. Since the data matrix code changes from part to part, it is helpful to ignore it during the locate. Also, this method of finding is less susceptible to noise and is quite fast in comparison with most other methods.
Below, I explain the steps needed to set up this example using IMPACT’s Vision Program Manager.
1. Load the designated images into the file camera
2. Open a new vision program.
3. Open the Locating tool box and drag and drop the Pinpoint Pattern Find tool into the program.
4. Trigger the camera to load an image into the program.
5. At the top of the Image Window, find the radio button labeled “2. ROI” and click it. Note that Step 1 was completed automatically
6. On the image, find the yellow ROI (region of interest). Click the upper left corner and drag it to the position on the image shown below. To resize the ROI, click one of the edges and drag it to the desired size. Try to position it to exclude the dark letters below the white square.
7. With the ROI still selected, click the rectangle ROI button in the menu on the left side of the Image Window and select Create Rectangle ROI.
8. Drop the new ROI inside of the existing ROI. This will subtract the small ROI from the larger ROI, creating a mask as shown in the image below.
9. Reposition and resize the interior ROI as shown in the image below. The idea is to include the white edge of the white box and exclude the data matrix code.
10. In the Search Limits box, click the drop down menu Possible Rotations and select + - A Few Degrees. This tells the tool to search for that shape even it rotates a few degrees. You can define “a few degrees” in the Start and End boxes. See the image below.
11. Click the Train button. As you can see in the image below, only the perimeter of the white box is trained.
12. At the top of the Image Window, find the radio button labeled “3. Search ROI and Pass/Fail” and click it. Position the green ROI on the image as seen below. This defines the area where the tool will search for the pattern you just trained.
And that’s it! Here are some images showing the Pinpoint Pattern Find® tool in action.
Please let us know how we can assist you with your Pattern Find, or other Machine Vision Applications.