next up previous contents
Next: Object Recognition Up: Stereo Vision Previous: Stereo Vision   Contents

Practical

The software you should use for this exercise is in 9.STEREO.

The goal of this practical is to use the TINA vision system to perform edge based stereo and wireframe reconstruction on the sets of images provided in the images directory, (shaft_assembly, saucer, big_crystals). The algorithm used is a variant on the PMF algorithm (as covered in the lectures) and produces data which is typical of edge based approaches. More details are given in the Tina Users Guide, which can be found on the tina web site.

Start the tool using tinaTool -r to get the stereo setup. Load and view the first pair of images. Ensure that the .cam file have been loaded too. Compute the edge maps for these images using the canny button. You need to rectify the edge data using the rectify button before you compute the depth with stereo. Lines and curves can be fitted to the disparity data using init geom3 and geom 3. You can control all of these algorithms using the parameters in the 3 Params dialogs at the bottom of the edge tool (the help system will give you assistance in using these).

This vision system is intended to provide data suiatble for location of objects in a robot control task. Perform experiments which enable you to discuss the following in your write up:

Describe the following in your report:


next up previous contents
Next: Object Recognition Up: Stereo Vision Previous: Stereo Vision   Contents
root 2017-11-20