The estimation of hurricane-top structure and motion using meteorological satellite images is an important application area for computational methods developed in computer vision, especially in nonrigid motion analysis. Accurate hurricane heights and winds are important for a number of meteorological and climate application, such as cloud model verification, physically-based numerical weather prediction and data assimilation, cloud-wind height assignment, convective intensity estimation, and radiation balance estimation for Mission to Planet Earth type climate baseline studies. However, it is a difficult task to develop automatic computational motion analysis algorithms capable of handling cloud motion.

[Data Acquisition]

The current generation of Geostationary Operational Environmental Satellite (NOAA GOES 8, 9, 10) has an Imager instrument with five multispectral channels of high spatial resolution, and very high dynamic range radiance measurements with 10-bit precision. The Imager instrument can image with high spatial, temporal, radiometric and spectral resolution \cite{hpkbuc98}, which makes it possible for automatic cloud-tracking algorithms to track mesoscale atmospheric phenomena such as hurricane and severe convective storms. Also, GOES Imager super rapid scan sequences at 1-minute intervals of mature hurricane give new capabilities to observe hurricane dynamics. Cloud details can now be seen due to the good contrast, in spite of very bright clouds. The 1-minute interval between images makes it possible to track features with high accuracy and reliability. These techniques are essential because they provide information which is independent of other meteorological measurements.

[Algorithm]

Assumption: Cloud images are segmented into small square areas. Each small cloud region is undergoing nonrigid motion according to the same given nonrigid motion model.

[Experiments]

Extensive experiments on the GOES image sequences of hurricane Luis have been performed. The data include 490 frames of hurricane Luis starting from 09-06-95 at 1023 UTC to 09-06-95 at 2226 UTC, provided by NASA-Goddard. Although five spectral bands are available for each frame, only visible channel having 10-bits per pixel is being used for our experiments.

Input GOES image sequences of Hurricane Luis (115K).

Recovered nonrigid motion (Note: This is the 2D projection of the recovered 3D motion) (141K).

Recovered cloud-top heights of Hurricane Luis (142K).

Animation and movie (.mpg, 2342K).

New results with dynamic fluid constraints

movie clip 1(.mpg, 1798K).

movie clip 2(.mpg, 2351K).

movie clip 3(.mpg, 2428K).

movie clip 4(.mpg, 2229).

CVPR'2000 movie(.mpg, 20.26M).

Motion-compensated cloud image interpolation results

Original visible movie(.mpg, 227K).

Interpolated visible movie(.mpg, 287K).

Original IR movie(.mpg, 223K).

Interpolated IR movie(.mpg, 282K).

[Validations]

Almost all the possible ways of validations are employed because it is almost impossible to get the ground truth of cloud top-heights and motion.

• Comparing SMAS estimated structure against the automatic stereo analysis results and manual tracked results: the estimated results are very close to the corresponding structures (within 1 pixel) for most areas of the hurricane.
• Comparing the recovered cloud-top heights and the stereo analysis results against the IR (infrared) cloud-top heights, as IR is believed to be the closest to the ground-truth for cloud-top heights (at the areas where the clouds are dense): the left is the IR image and the right is the error distribution.

[Back]