Dear Forum
I'm using the pattern matching vi's (LV 7) to find a reference point (shift only) on several images. I have a rather complicated vi to train and find the right settings (based on one of the labview example). Basicaly it works but I'm not quite satisfied with the results, sometimes it misses a match while it seems obvious...
So here are a few questions I'm hopping someone could give me a help with:
- when defining the template I'm using a poligonal ROI, however I have the impression the template is in fact defined by the ROI bounding rectangle, is it true?
- I'm looking an image which is 640x480 pixels, my ROI is about 100x60 pixels. From the labview examples I have the feeling that the general vi's (learn pattern 2, find pattern 2, etc) are optimized for smaller ROI sizes, So I'm trying now to use the advanced pattern matching vi's. from the literature I identified two parameters which could be of interest: Initial match list and sample size. I have however no idea where to start; a hint on what are the default parameters (such used is the non avanced vi's) would be usefull
- I've tried several filters mostly gradiant for edge enhancement, smoothing or Nth order for getting rid of noise and math look-up for contrasting the image, maybe there is a better solution?
I have a few sample image I'm unfortunatly not allowed to show on a forum, should a LV Guru be interested by my problems I would be glad to e-mail
Best wishes
PiF
