JRobot

Hi,
I'm interested in using the AVRCam to track multiple colored objects, but am confused about one thing. What happens when I try to track a colored object and two objects of the same color are in a captured frame at the same time? Does it track both objects and return two bounding boxes? Does it track the first blob and return one bounding box. Does it create a bounding box to encompass both objects? Or am I missing something all together?

I would appreciate any help.

Thanks,
Adam Q

The AVRcam would return two different tracked objects (two different bounding boxes), with one bounding box around each object.

How does the AVRcam find that the two objects are seperated from eachother? Doesn't the CMUcam record all objects of a colour as the same object?

The AVRcam tracks connected regions (i.e., two or more adjacent pixels of the same color) of colorful pixels. Up to 8 connected regions can be tracked simultaneously, starting from the top of the image and moving downward. If more than 8 connected regions are encountered, the remaining regions are ignored. Parameters about each connected region (x,y coords of upper left corner of region bounding box, x,y coords of lower right corner of region boudning box, and color) are then sent out the serial port.

In other words, as long as two objects of the same color aren't adjacent to each other (at least one pixel of some other color separates them), the AVRcam will consider them two different objects.

Wow that sounds pretty cool!