Are the two objects going to move around different regions of the image? Are they ever going to touch or overlap?
If the objects are different sizes, then you could use the particle analysis to extract only the object of interest based on size. If they are different color, then you could possibly use the color threshold to pull one object.
You can also keep track of the position and velocity of your object in each frame and use them to make an estimation of the position in the next frame. Use this estimated position to set a search ROI, which can both exclude other objects (if they don't overlap the desired object) and greaty improve computational speed. Some combination of trajectory prediction and particle property filtering (ex. size, height, etc.) mi
ght work well. Try to use any information you have about the contraints of motion for the objects to choose a good algorithm.
Regards,
Brent R.
Applications Engineer
National Instruments
