Vision ROI consistency

[DarkLordOfG]

It would be nice if there was some consistency in the definition of regions in the vision development kit.  For example, the extract takes a 4 point array as a rectange.  BUT the ROI to rectangle tool outputs a cluster with the rotation:

It would be nice if there was a generic IMAQ extract ROI and if there was one "rectangle" definition not two.

[muks]

AS mentioned by you the rotation is included and that difference will be helpful (Or am i used to this):smileyvery-happy:. Still contemplating...

[usrfrnly]

The inconsistancies are astounding in a lot of the vision tools - the ROI stuff especially. Note the cooercion dot in the above image. ROI outputs often get that dot from ROI inputs so ROI's arent even consistent to themselves. The discrepancy appears to be in the "ID" and "Type" controls in the "Contours" array elements. 

You mention that the root of the problem for rectangles is the 5th item in the cluster above - rotation - which is a double and negates the ability to create an array directly because of dissimilar data types but that isnt the only way rectangles are called out. IMAQ Detect Rectangles returns an array of 5 clusters of XY points as the rectangle descriptor. There are some other variations on the theme. Maybe they should standardize the ROI descriptor and use that for all IMAQ and vision shape call outs.