Multiple image analysis

[whitelancer]

To anyone that can help,

I am a student at the University of Delaware and I want to use LabView to analyze a set of 10 or so images. These images are of a robot and mesurements such as speed, accleration, and turning radius need to be taken. I can acquire the images and clean them up, but I was hoping somone had a VI that could then anaylze the set of images. Any help would be greatly appreciated.

Sincerely,
Jon

[Jeffrey_P]

Good morning Jon,

Are all 10 images taken in the same Frame of Reference?  That is, can you use some arbitrary reference in all 10 images and the robot's relative position to that reference to determine the speed and acceleration of the robot?  This would also require that each image has some sort of inherent time stamp.  I don't think that National Instruments would have any shipping examples that exclusively would perform these types of calculations; however, I can point you to the Vision Concepts Manual to help you become familiar with the theory behind the VI's that will help you to accomplish this task:

I hope that this information is helpful.  Please post back if you have any further questions.

Respectfully,

Message Edited by Jeffrey P on 08-21-2006 10:26 AM

[whitelancer]

Hey Jeffrey,

Thanks for getting back to me. All of the pictures are taken in the same frame of reference and using an arbitrary reference would certainly work. My question is now how do you implement a time stamp on each image and then process them through a VI one after another. The images are taken using the NI Vision Acquistion and then are cleaned up to be processed through a VI, but how do compare several images? The Vision Manual has been very helpful in teaching the basics of what I am trying to do; however, making the ideas work has been harder. Thanks again for any help that you can provide.

Sincerely,

Jonathan Hinkle

[Jeffrey_P]

Jonathan,

I found a similar question answered in a previous discussion forum.  The forum discusses a particle, but could easily be adapted for a robot.  Here is a link to this discussion forum.  If you have any questions regarding how to implement any of these ideas, please post back.  Most of the ideas in this post can be discovered through the LabVIEW Shipping Examples.

Respectfully,

[whitelancer]

Thanks  Jeffrey,

The particle form can be used for speed measurements for the robot. Using Vision Acquisition I can calibrate each image from pixels to real world units and since I am using the same frame of reference that calibartion can be applied to the following images. A formula is easy enough to write to find time elapased between the first and last image. The Compare example demonstrates a method to import two images (the first and last), however, I am having problems determining how to take the real world calibration in one image and using it to find the distance travled and then outputting a velocity. I was hoping that you had a suggestion to help out. Thanks again,

Jon

[Jeffrey_P]

Jon,

You said that you can calibrate each image to real world units.  You should be able to take the difference, either in number of pixels or real-world units, to calculate the "distance" traveled.  Then, use the difference in image number, and based on the frame rate, determine the amount of time that passed.  The combination of those two values should give you your velocity!

Cheers!,

[whitelancer]

Yeah, Im just not sure how I can take the calibartion for a single image found that I do in the Vision Assistant and then use it to find the distance between two images in a VI in labview. Once i have that distance the calculation is simple. Thanks for your help.

Sincerely,

Jon

[Jeffrey_P]

Jon,

I apologize.  I misunderstood the question.  On the Image palette in Vision Assistant, the very last option is called Run LabVIEW VI.  You can browse to any LabVIEW VI that you have created.  You can then associate results from previous steps within Vision Assistant with the controls for this VI.  For example, you can write your VI to have all of the input data that you need as controls.  Then,  you can perform all of the calculations within the VI.  When you call this VI from Vision Assistant, you can associate all of the parameters from previous steps to be those control values.  Then, you can take the result from this VI, and pass it back into the VA script!

I hope that helps!

[whitelancer]

Thank you!

Thats very helpful. I guess the only thing that i am a little confused on is how the VI will process two images even though Vision Assistant will only work with one image at a time. Thanks agian. I greatly appreicate all of the help.

Sincerely,

Jon

[Jeffrey_P]

Jon,

The actual VI won't be handling any of the images.  It will handle the numeric results from the images!  This data is what should be passed into the VI.