how do i put the results from a calculation as the offset value

[AnalogKid2DigitalMan]

Have you experimented with using colored filters in front of the camera lens to increase the contrast of the yellow (or nearly block out anything but yellow)?

[Torba]

Hello kupikupi,

In order to see what you have written so far, could you attach your current program along with a sample image for processing?

Regards,

Mike T
National Instruments

[kupikupi]

hi .. thanx for all the replies..

i've attached my codes for masking the image to remove most of the background, i need help in extracting the colored finger tips..thanx

[Torba]

Hello kupikupi,

I now have a much better idea on what you are trying to do.  I have a simple example attached that demonstrates how to extract a specific color out of an image, and locate all the instances where that color is found in the image.  I produced this example using Vision Assistant.  If you have the Vision Development Module installed on your machine, you should also have this program.  You can find it by going to Start >> Programs >> National Instruments >> Vision Assistant.  I would highly recommend that you try this program since it offers a simpler interface to perform tasks, and it generates LabVIEW code once a sample script is complete.

In the event that you do not have this program available, you can build your program off of the example that I have attached.  The program will prompt for two files:  One is the yellow.png file that I have attached that is an extraction of the yellow color that you have in the image that you sent me, and the other is the image that you wish to search through for that color.  In this case I selected the actual image of the glove that you sent me.  The results of the program will locate all of the objects in the image that contain the specific yellow color that is defined in the template .png file.  It will then return the coordinates of those instances.  Use this example to identify the colors in your image.

Also notice how the code is arranged on the block diagram of the example.  A very good tip that I would suggest would be to try and arrange your code in a similar fashion as what is demonstrated in the example to improve visual clarity and ease of debugging.  The appearance of the code can go a long way as you further develop your application.

Regards,

Mike T
National Instruments

[allen_h]

Message Edited by allen_h on 07-11-2006 05:45 PM

[kupikupi]

hi .. i tried it out with vision assistant ,.. i got the coordinates ... but when i tried the VI file .. it doesn't work .. the read image and vision info had an error.. it says.. it's not the latest version of IMAQ.. i'm using v7.1 .. and how do i convert programs from vision assistant to LV ? from the program, it shows the coordinates of the matched sections, and the matched section is only a small part of the whole yellow region, how do i extract the whole region based on the coordinates? and how do i actually  extract the desired section of the images.. now i understand the matching part and getting the coordinates, but i dont have a clear idea on how to use that info to extract the images?? thanx very much for ur help .. Torba..   

[Torba]

Hello kupikupi,

Since it sounds like you have Vision Assistant 7.1, lets continue working with it.  I attached a new script that was provided graciously by our Vision development team.  This code will extract the entire regions of each color on the glove.  Check it out.

The errors that you were getting when you attempted to run the previous example raise a question.  What version of IMAQ do you have installed?  You can find out by opening Measurement and Automation from your desktop, and expanding the Software tree on the left side once it is opened.  You should see a listing for NI-IMAQ under the Software category.  Click on the NI-IMAQ listing once, and the right-hand window should display the version of the software.  If you have a version of IMAQ that is older than 3.5.1, you can obtain the latest version of this driver by going to the following website to download it:

NI-IMAQ Version 3.5.1 for Windows 2000/XP
http://digital.ni.com/softlib.nsf/websearch/A7B53F9DE1D288ED862570D500623EC4?opendocument&node=132060_US

When this page opens, click on the "Vision Acquisition Software" link to navigate to the installer that contains the IMAQ driver.  You will have to be logged into our website to download it.  Once you have the VisionAcquisitionSoftware800.zip downloaded, uninstall the current version of the IMAQ driver that is on your PC by going to the Control Panel >> Add/Remove Programs >> National Instruments Software.  Click "Change" and a list of all your NI software will appear.  Locate the NI-IMAQ driver here, and select uninstall.  Restart your machine, and then extract and install the Vision Acquisition Software package that you have downloaded.

To answer your other question about converting a script to code, you can convert a script to LabVIEW code from the Vision Assistant by going to Tools >> Create LabVIEW VI in the top bar of the Vision Assistant.  Follow the steps, and a VI should be created that you can open in LabVIEW.

I hope all of this information helps.  Please let me know if you have any questions about the script.

Regards,

Mike T
National Instruments

[kupikupi]

hi Torba, thanx alot for ur help, i've been stuck for quite some time, but since u started helping me, my progress has been quite good,

i've got problem with thresholding, from the program u made, it seems like u can easily set the threshold range accurately, my understanding is quite poor tho i've gone thru a few books, i dont know when to use RGB, HSL, HSV or HSI. can u please explain to me? rite now i'm still working on the thresholding part based on ur program.

my IMAQ is quite old, im downloading the latest one.

from the program u wrote, the masking only works for that particular image and can't be used for any other random images, but from ur idea , i managed to creat one for ramdom images. can u please help me check ? i've attached my program
does vision assistant have all the functions i need? i find it very useful, but there are quite a lot of settings which  i dont know how to use,

also from the program, which i converted to .vi, i only get the results for the last procedure, i can't view the other thresholded images, the thresholding was done for different colrs, how do i combine them to be processed? how do i do color matching , since the images are in gray scale?

i think at the moment my biggest problem is the thresholding part. i hope u wont mind helping me out,  ur explanation has very very detail, makes it easy for me to learn LV, i'm new , but i've learn alot frm u .. thanks very much.

[Torba]

Hi kupi kupi,

I have a few good resources to go to that describe what all of these color modes mean.  Here is the link:

Color Image Representations
http://zone.ni.com/devzone/conceptd.nsf/webmain/53327b1c1e37d8938625685e008152c2

You will also find a few more links at the bottom of that page that discuss how thresholding works that should help. 

In general, if you are curious to know more about how a function works, you can usually find some help on that item by opening the context help in LabVIEW (ctrl-H) and putting your cursor over the VI in question.  The context help will display a summary of what the function does, and what all the inputs and outputs are.  You can obtain detailed help by clicking on the "Detailed Help" link in the Context Help window.  Occasionally, this detailed help window will also contain references to examples that demonstrate how to use that particular function.  I do know that the Color Thresholding VI has a link to an example in its detailed help.  The location for the example was claimed to be:  C:\Program Files\National Instruments\LabVIEW 7.1\examples\Vision\2. Functions\Color\ColorThreshold Example.vi.

Another awesome resource for help on Vision theory is already at your fingertips.  If you navigate to www.ni.com/vision, it will take you to a page that focuses on everything about vision.  To find specific help on anything (i.e. color thresholding) simply type the keywords in the searchbar in the upper right corner of the page, and perform the search.  The search should return with material in different categories throughout our website.  You can focus in on example material, issues that previous customers have had specific to your keywords, or even theory documents on our Developer Zone site.  I actually found the Color Image Representations page by searching for "RGB HSL" in the searchbar, and selecting Developer Zone for the Search Within dropdown list.  Try this.  You would be surprised of the results.

Your one statement about the script only working with a specific image is correct mainly because I specified an image to load at the beginning of the script.  If you want the script to process whatever image you tell it, simply convert the script to code with no specific image loaded.  You will notice in the block diagram that there will be a File Dialog VI placed in the beginning to prompt the user to load any image for processing.

To summarize, don't hesistate to give our website a shot for searching information on all of the functionality and terms that are being used for Vision.  We try to provide our users with as much supporting information as possible to make them successful as efficiently as possible.  We typically use the discussion forums as a means for users to ask very specific questions that they haven't been able to find answers to elsewhere.  This utility in conjunction with all the documentation on our website and the user manuals hopefully covers all the bases for support.

Thanks kupikupi!

Mike T
National Instruments

[kupikupi]

hi .. i've done thresholding for the different colors, so i get three results, but how do i combine all the results to be further processed? and how do i do color matching since the images are grayscale??