Gowri
2001
http://folk.uio.no/bjoernw/BW_extended_abstract_2ndACD_Workshop.doc
stereo image analysis in studying coast line retreat, from aerial images (2300 meters). They got 20-30cm (yes centimeter) accuracy in x,y and height.
They say they push them thro' systems to get DEMs and over time they studied "changes" in the DEMs.
Given, our system may find use on similar lines (over time) we should record location and other relevant details.
A GPS and good digital compass would be required.
Reports testing long range terrestrial range finders would help too. We found one http://www.cosis.net/abstracts/EGU06/04167/EGU06-J-04167.pdf
they use a model with 800meters range.- Riegl LPM-i800HA
This should be a good way to get ground truth.
Considering, we wish to get 10cm depth accuracy, we need something which is better for ground truth (certainly not our eyes! unless we get down to the river and measure with tape).
The above model is reported to give 15mm acc. (http://www.riegl.com/terrestrial_scanners/lpm-i800ha_/i800ha_all_.htm)
And the report confirms it can be packed up for use in low temps.
This would probably help us see (if) vision can help replace the expensive,(time consuming) equipment.
