Result variation and inconsistency Ni Vision Builder AI

I am currently using OR Vision Builder HAVE (2012) to analyze a series of images, and the preparation of appropriate inspections, I noticed that the results are not reproducible.  I am currently using various measures of analysis such as the color of the measure, number of pixels and the strength of edge and using the data logging feature.  The problem is that if I double-click on a stage as if to change, click on transfer to the next image, without changing anything, click OK, and then run all the tests again, I get different results.

For things like the number of pixels that is relatively small - anywhere.0001 about 3 when it comes to figures around 200, but for the strength of edge especially these range from 10-20 at a time and sometimes he could find even an advantage that is then seen in over 40 years (and the edge resistance threshold is well below this value).  If this is inevitable then I can run many tests and average results, but it will be a lot of time because it also involves the steps wrong edition.

Mainly, I would like to know if it is common and expected and if there is a known percentage error, or maybe I can change the configuration of the program in an attempt to overcome the problem?

Thank you!

The good news is that we have found what is the problem. When you change a step that uses a coordinate system and the current image is not the same as that used during the stage was originally set, the return on investment is transformed to the new position and step updates the settings used for this new position, when you press OK. Rounded the errors that occur when the transformed KING is caluclated, the new system of basic contact information you may not in the same positions of the KING. Kings use integers and therefore round us to the nearest integer, but it does not exactly the same positions as the coordinate system original base product. This may cause the new coordinate system of the step being out of up to 1 pixel and causes the return on investment deal with a slightly different part of the image

For a given inspection/images, the results will not change unless you change the step that's important.

If you change a step, but do not want to make changes, press Cancel instead of OK, and all settings will be the same.

We will study a real fix for this problem that involves only update the coordinate system basis when the user redraws the return on investment, but the bad news is that this correction will not be in the version of Vision Builder AI we are about to go out.

I'm sorry for this gap, and we plan to address in the future.

Thank you

Brad

Tags: NI Hardware

Similar Questions

  • I can't create an installer, after install the Vision Builder demo and then uninstall it. I can't create a Labview Installer more, he asks me an installer of generator of Vision, even if Labview my program does not have nothing to do with it...

    I can't create an installer on LabView, after install the Vision Builder demo and then uninstall it. I can't create a Labview Installer more, he asks me an installer of generator of Vision, even if Labview my program does not have nothing to do with it...

    Hi, Shada, this can happen if you include additional installers to your installer for example DAQmx. When you do this the installation program must include other files like MAX (Measuremente and Automation Explorer). Probably when you installed Vision Builder he updated MAX, so when your Installer try to understand MAX, his request the source where it was installed in your computer in this case Vision Builder.

    Here you can cache information so it is not asked once again, or you can download a version more recent fo DAQmx, then it will update MAX and change the dependecy.

    You can find more details about it in this KB:

    3S88PJMY knowledge base: why do I get an error message 'Missing Distribution' when you try to build an install...

    And to keep a copy, here's more information:

    How to stop missing delivery Notification whenever I build an installer?

    Best regards

  • vision builder to labview migration

    Hello

    I have a project of vision. Pluggable VISION BUILDER pour image processing and the LabVIEW pour interfacing.

    My question is: how to call an inspection done by VISION BUILDER in the LabVIEW software.

    NB: I found in the Tools menu of VISION BUILDER a feature titled "Migrating to LabVIEW. but she could not be executed because I don't own

    the DEVOLEPMENT VISION MODULE. Have to buy another software OR (VISION DEVOLEPMENT MODULE) pour fix my problem? : smileysurprised:

    Thanks in advance

    Hello

    Vision Builder is provided with an ActiveX control that you can call from LabVIEW. You will find enclosed a few Directory promotion the in examples

    \

    The ActiveX component allows to control Vision Builder as you would you interface greengrocers. Once you have created your inspection in Vision Builder, you can use the ActiveX component to connect a Vision Builder of LabVIEW, open an inspection, execute, recover and display the image (you need not purchase the Vision Development Module pour it), and get the results of inspection and each step of the inspection.

    Hoping that this answer has your expectation

    Kind regards.

    Christophe

    OR R & D Vision

  • Migrating to labview vision Builder... Many VI

    Hello

    I use the latest version of labview and OR vision builder. Using Vision builder get the coordinates X and Y of the object in this area. I need to export this to labview. I tried going in "Migrating to labview. I created a new folder then I gave the name. When I click OK. It creates many VI. I need this in VI only. When I opened the "main.VI", I see that "make, start." When I check the block diagram. I couldn't find the programe I needed for it.

    Hope you understood my question.

    Can you please help me.

    Thanks in advance

    Another option, which could be used in conjunction with the API mentioned Christophe solution is to your X / Y values the variables that you update your VBAI inspection and then read in your code LV. Here's a tutorial on the use of shared variables to communicate the results to the LV of the VBAI:

    http://digital.NI.com/public.nsf/allkb/D492139A746F91A8862576E80051BBCE

    Here is a video showing how to use the API that Christophe has mentioned:

    http://zone.NI.com/WV/app/doc/p/ID/WV-2424

    Here's a video of the use of shared variables (demo shows using targets of the RT, but the same idea might work locally):

    http://zone.NI.com/WV/app/doc/p/ID/WV-2423

  • Control of a system of engine not to not 2 - axes of Vision Builder

    Hello

    is there a way to control 2 Motors step by step via Vision Builder?

    What are the components do I need?

    BR,

    Klaus

    Hello knuemm,

    There is no action, which allows to directly control the player. It depends on your hardware configuration, some engines step by step has a serial interface and also the device the vbai is running on other devices of movement has a digital input and some vbai devices have outputs digital that may be used for this.

    Fleeing a LabVIEW VBAI Inspection

    I followed the smartcam NOR control pilot Motion

    You can also create steps for the VBAI to run LabVIEW code

    NEITHER Vision Builder HAVE Development Toolkit

  • Vision Builder 2013 DURATION: Selection of Inspection

    VBAI2013 I would like to insert a step of selection of inspection in the selection state of inspection which applies to the system on which the Vision Builder is installed and not to individual inspections. Therefore, it is not possible to how in the Version of the Runtime-license-where to add the steps?

    Thanks in advance.

    The steps for product status select are saved in the "Vision Builder system Config.ini' in the folder VBAI. You should be able to copy this file to the machine with the runtime version and if it has the same version of the VBAI, it should be able to apply all the steps and the INI settings.

    Hope that helps,

    Brad

  • Integration of third party Vision Builder C++?

    Hello
    A customer of ours uses NI Vision Builder in an application and questioned on the possibility of integrating our algorithms with her exclusive image processing.  Our code is a C/C++ library.

    I did some reading and downloaded the eval package, but it is not clear to me if this is allowed with the product.  The defined C API allow access to built-in functions, but I've not seen an obvious way for third party algorithms to access images or add new features.

    Vision Builder only supports it the algorithms or can developers add custom product image (in C/C++) processing algorithms?  Thank you.

    Hi eepdx,

    VBAI is not a contribution to the third-party code integrated directly in the script. And you are right about the C API.

    You can use the range of Vision Builder AI communications to communicate with the C++ application. Or you could translate your final VBAI script by LabVIEW code, then, LabVIEW can call a dll, so you could create a C++ application dll and call it from the LabVIEW code (it would be more complex).

    I hope this helps.

    Daniel C.

  • NEITHER Vision Builder built exe does not have logging.

    Hello!
    I did have an application with NI Vision Builder, and the executable version does not save the log file. During the test it or vision, she recorded when I migrate to labview it records, but when I build the executable file, which does not have logging. What could be wrong?

    Thanks for your replies, the solution was easier than I thought. There is no problem with the path or something like that. The problem is that I checked the box for "Log in the background", and (I think) it is not a program, which could run this step. So I unchecked that and now it creates the log file that I wanted to.
    Once again, thank you for your quick response!

  • vehicle followed by using vision builder

    Can someone help me to follow a vehicle using vision builder?

    The object tracking feature to track multiple objects.

    To do exactly what is shown in the video, where the number of cars is variable, you must use the LabVIEW module and the development of the Vision to have more control over the creation of objects. (The limitation with Vision Builder AI, is that each object to track is associated with a follow-up session, so you should know in advance how many object you must follow, but VBAI to track multiple objects).

    The principle I described should be the same in LabVIEW: detect somehow the new car which comes within the field of vision and learn at this time where the object so that it can be followed. The prerequesite should be that you need to know where to detect them the car that comes in the field of vision. The difficult part of the application must define the area of interest of the new object that you want to track.

    The LabVIEW example for tracking of objects can be found here:

    C:\Program Files (x 86) \National Instruments\LabVIEW 2014\examples\Vision\Tracking\Object Tracking.vi

    Christophe

  • NEITHER Vision Builder to tiara

    Hello NOR the community.

    I built a system to measure through NI Vision Builder. Now I can record the measured data (only!) in a *.csv or *.txt file.

    I want to import the data measured in DIAdem automatically.

    Is there a better way to do it? Is there a direct Interface between the 2 software systems?

    Thanks in advance!

    Hi Steffen,

    When you install DIAdem you are also automatically installing a separate program, called "My DataFinder", which is a database and indexing service.  The tiara and "My DataFinder" use the DataPlugins even index of load/data files.  When you complete a session of text use wizard, it creates and registers a use in addition to load the data file into the data portal in DIAdem.  The idea is that eventually DIAdem both the DataFinder can use the use that you created for the load index / all similar data files.  In practice, the use you get from the Assistant is often not ideal for use with a variety of similar data files.

    My initial inquiries indicate that ASCII files, you are already creating the best leads that we have between Vision Builder and tiara.  The marketing director of DIAdem is now advocating for options in future versions of Vision Builder who could provide a superior connection.  For now, however, I think our best bet is to outfit you with strong DataPlugins which will provide drag & drop of loading in the NAVIGATOR tiara as well as automatic indexing to the DataFinder database.

    So, could you post some of the ASCII files you created with Vision Builder?  I would be willing to create a use or a set of DataPlugins that should give you the basic data functionality and connectivity you are after with DIAdem.  Or you could send to [email protected]

    Brad Turpin

    Tiara Product Support Engineer

    National Instruments

  • Version of LabVIEW for Vision Builder AI 2012 plugins

    Hi all

    II was a little challenging to find out what version of LabVIEW has been necessary to develop customized Vision Builder AI 2012 plugins, I found it by starting the installer, LabVIEW 2011 (but I wish I could find this information in the Readme of Vision Builder AI 2012).

    Now my question is: is it possible to use LabVIEW 2011 SP1 to develop plugins for Vision Builder AI 2012?

    Right... I tried, I installed all versions VBAI and LabVIEW, I could.

    I turns out LV 2011 sp1 is just an update of 2011 and so LV2011 SP1 can be used to develop customized Vision Builder AI 2012 plugins.

  • NEITHER Vision Builder, acquire Images with fake delivery

    Hello world

    I am a student and I work as a trainee in a company, which produce helmets.

    To control the quality of helmets, there is a Webcam, wich take pictures of each helmet product.

    These Images are saved on the server. With the tool 'Simulate delivery', I can analyze the images, which are now in the folder.

    But when a new Image is the drop in the folder, I actually start the inspection again.

    How can I realize that NI Vision Builder always takes the current Images in the folder?

    Thank you very much for your help!

    Best regards

    Romain Menz

    Here is a simple example that illustrates how to dynamically read the images from a folder. Once the image has been retruned to VBAI, the image file is moved to a subfolder called Complete, so he has not read again the next time step is called. Step returns an error if there is not more images of the file type in the folder.

    Hope this helps,

    Brad

  • Error - Vision Builder Inspection to LabVIEW Migration


    You can verify that you have read/write access to your temporary folder (in general C:\Documents and Settings\Settings\Temp).

    This error is due to the fact that a temporary file (VBAI.bin) is probably not written in this file or cannot be read during code generation. Is it present in this folder after you try to generate the code?

    What privileges you have when you run Vision Builder AI? If you are logged on as administrator? You can try to run Vision Builder with administrator privileges and see if you still get the same error?

    Also, what operating system are you using?

    Thank you

    Christophe

  • run a VI in labview with Vision Builder (VBAI)

    HELO,

    I try to use the 'run VI' tool of the vbai (2010). When I lunch the inspection on my PC (with labview 2010) it's ok, but on another PC (without labview), I got an error message (dll is missing).

    I have to copy the data in a specific folder to use this tool?

    You have solutions to solve my problem?

    Thank you

    Ulysses

    If you read the context-sensitive help for the LabVIEW execution step, there will be clear instructions to operate the LLB for Vision Builder AI. I'm not including instructions here because they change slightly depending on the version of the VBAI you are using. However, here are some useful links:

    http://digital.NI.com/public.nsf/allkb/231D87B9D8C3B71A862572D4007EAAF8

    http://forums.NI.com/T5/machine-vision/quot-make-sure-all-of-the-VI-dependencies-including-VI-lib-VI...

    Access the LabVIEW execution step in VBAI, change it, then go to the help > show contextual help (Ctrl-H). In the FAQ tab, there will be a link to describe in detail how to build your LLB for distribution... read carefully and follow the directions.

    Hope this helps,

    Brad

  • several webcams in for usb or imaq vision Builder

    Hello

    I have a question about imaq nor for usb or vision builder AI 3.6 (I did a vi with imaq and I run in the constructor of the vision) and I want to know how can I connect multiple webcams to the computer, and how many of them can be connected. can they all be of a make and model and can they be executed with a driver installed on the computer?

    Thank you

    Hello

    In general NOR-IMAQ for USB can support only one camera at a time. You can acquire multiple cameras (not simultaneously) using USB IMAQ list VI cameras to identify each device plugged into your system and access it via the other IMAQ live to USB.

    I would also like to point out that Vision Builder AI does not support the NO-IMAQ for cameras USB driver. It is designed to solve machine vision applications, but USB cameras provide in general or advanced necessary for machine vision triggering modes.  Because the machine vision and USB cameras will usually together, we do support USB cameras in Vision Builder AI.

Maybe you are looking for