Work in progress

1
Work in progress

• Adapting SPSF (Small patch sphere fit) to fit to
  higher order shapes (as bellows window)
• Investigating harmonics
• Via surface model
• Via test of thick window
• Writing two papers as potential NIM publications
  Mucool notes)
• Development of photogrammetric methods for
  quality control of non standard windows
• Development of photogrammetric methods of stress
  analysis

2
Photogrammetry measurements
John with the digital camera
Twelve photos are taken at various pressure
stages
3
Preparation of the test at NIU
View of the window, video camera, digital camera
and tripod and projector
View of the window, scale bars and targeting sys.
4
New presentation tool from GSI

• 3D Viewer
• Rays
• Camera stations
• Scale bar
• Autobar (Crucifix or Southern Cross?)
• Shapes
• Object targets

5
Run first bundle (refine camera parameters and
coded target positons)
Scan image
Detect Autobar and coded targets
Resect for initial cameral parameters
Automatch
Make a list of coded targets in resected images
Triangulate object coordinates
If a coded target is in gt2 images
Run final bundle
Triangulate coded target coordinates
Does image contain gt 4 coded targets?
Resect camera parameters
6
Photogrammetric Record, 16(93) 503517 (April
1999) AN APPLICATION OF CLOSE RANGE PHOTOGRAMMETRY
IN DOLPHIN STUDIES By S. BRA GER and A. K.
CHONG University of Otago Abstract Dolphins are
agile animals and are very difficult to measure
at sea. However, for the conservation of
threatened or endangered species, measurement may
be vital because it allows a demographic analysis
of the population. Hectors dolphin
(Cephalorhynchus hectori) is a rare
species living only in coastal waters around New
Zealand where it is studied from small boats.
This paper describes a stereophotogrammetric
technique developed to measure body length
accurately at sea without having to capture the
individual. Constant calibration with the use of
a control frame allowed accurate body length
measurements of dolphins to be made with this low
cost system, with a measurement error of 4 per
cent to 6 per cent of actual length.