Presentation PDF (569.5 kB)
The camera images were used to determine three-dimensional structure of the developing convection. Using the intrinsic (focal length and dimensions of the imaging chip) and extrinsic (location and orientation of the cameras) camera parameters allows the definition of projection matrices that relates points in 3-d space to the projection onto the image planes. Rather than used measured values of the extrinsic parameters, an iterative scheme is used to refine the measured values. Tests of the scheme using known landmark points show that scheme is accurate to 50 meters at a distance of 30 km.
This talk will describe the method for three-dimensional stereo reconstructions and show an application to a case of deep convection that occurred during the CuPIDO experiment. On August 10, 2006 (Intensive Observing Period #10) there was a transition from a roughly 2 hours period of cumulus congestus followed cumulonimbus development. Half hourly soundings taken from the vicinity of the mountain peak showed an inversion present at about 500 mb that capped the convection. The combined effect of warming below and cooling at the top of the inversion layer erodes layer. The onset of the deep convection corresponds to the breakdown of the inversion.