Machine Vision

[nesslersreagent]

Hi

I've succesfully built and tested a fully functional Motion detector, using the "Image Subtraction" procedure.

But in doing so, I'm getting multiple particles(blobs) for a single moving object, possibly due to Image subtraction.

By using Dilate, Erode, and other Morphological options, I've managed to combine multiple "closely spaced blobs" into one as desired.

However, it's not enough. For example, a human appears as 3 distinct(closely spaced) blobs, instead of one.

No problems with large objects(cars) though.

Could anyone suggest any method to track a collection of blobs as one blob?

Note:

I've tried collecting the center of masses of the particles, and performing some basic calculations to group them into groups of closely spaced blobs, and then calculate an average value of center of masses per group. However that didn't work.

Included images show the results.

[nesslersreagent]

Any ideas, anyone?

[nesslersreagent · Accepted solution]

Problem solved.

Constructed a VI that handles the task well.

Attached image shows the idea.

Works quite well now with great results.