Abstract
Meteorological sensors with digital image output have created a need to generate rapid three-dimensional renderings for visualization of atmospheric conditions and processes. This paper describes the use of octree encoding to produce three-dimensional scenes using a digital image as input. The three-dimensional or four-dimensional digital input is composed of pixels with each pixel value represented as a numerical number. The concept is demonstrated using infrared images from the GOES geosynchronous satellite. By relating observed radiance values with atmospheric sounding information, each pixel value is transformed to a height of the cloud in the atmosphere. Subsequent octree encoding and display, with simulated illumination, yields a three-dimensional cloud image. Examples are provided from photographs taken from a DEC VAXStation 8000 image processing system. The octree methodology is suggested as a contribution both to scientific-operational visualization and four-dimensional data assimilation (4DDA).
