multiple particle coordinate recognition

[systems_eng]

Right now I have an image where I have put through to isolate various regions of highlighted pixels (hot spots).  The final threshold filters creates a binary image with multiple dots which are to be recognized.  I would like to identify each one of these dots and apply cooresponding (x,y) coordinates to each one.  At this point, I have been able to apply the (Detect Objects tool) to find all dots, and then used the (Calculator function) to apply coordinates to the first object.

The calculator allows me to manually select which objects I want to calculate a position for.  Is there a way to make this tool a little more dynamic... making it calculate/display the coordinates of each dot that the detect-objects-tool finds?  Ex. for 5 dots, the program displays 5 different coordinates to the screen.

[systems_eng]

an additional question that is closely related to this topic, is how to acquire and export the data that the particle analysis calculates.  being able to calculate some of the variables of each individual object, which the particle anlalysis offers would more than satifsfy my needs.

[David S.]

Hi systems_eng -

It looks like I already replied to this topic here

In general, I encourage you to post only once in the forums on each topic.  We like to keep all responses and ideas on a topic grouped into a single thread so that future searches lead a person to all the information available.  It will also be easier for people helping you to keep track of the replies already posted.

[svif]

For finding the center of highlighted spots:

I have spent a lot of time trying to make a fast and reliable function to detect highlighted spots - also identifying when a spot is really two connected spots. If you are not extremely time limited (it takes several ms to run them) then using two convolutions and centroiding is very effecitve in finding centers of these spots. The imaq centroid finds the center of mass. It is somewhat faster to use Imaq particle analysis for finding the center of mass, it however uses a binary image - which is slightly less acquate (depending on the shape of your spots).

Here is a pair of routines i have used.

/simon