Overview

Yu's group working on the cloud/ice discrimination problem using Multiangle Imaging Spectro Radiometor (MISR) data. MISR is a sensor aboard on the Terra satellite, as part of the Earth Observation System (EOS) program. With the multi-angle information provided by MISR, we are trying to classify the cloud and the ice/snow covered land at both poles. ( Here are one MISR image and the classification result from our approach, Linear Correlation Matching Classification.)

The MISR instrument offers a new and promising avenue for cloud detection methods. MISR has nine cameras looking at the Earth from different angles simultaneously. The nine camera angles are 70.5^O, 60^O, 45.6^O, 26.1^O forward, nadir and 26.1^O, 45.6^O, 60^O, 70.5^O afterward  Each angle has four observing bands (red, green, blue, and near-infrared). MISR covers the earth every 16 days. ( Watch an illustration for the MISR instrument here.) The high dimensionality and huge size of the data provide statisticians huge space to play around.

People to Contact

Tao Shi, Bin Yu

Related Publications

Shi, T., Yu, B., and Braverman, A. (2002). MISR cloud/ice classification using linear correlation matching. Technical Report #630, Department of Statistics, U.C. Berkeley.[article]

Shi, T., Yu, B., and Braverman, A. (2002). Discriminating cloud over ice/snow using MISR data. Invited Talk in American Geophysical Union (AGU) 2002 Fall Meeting at San Francisco, CA

Shi, T., Yu, B., and Braverman, A. (2002) Discriminating Cloud over Ice/Snow using MISR data. Poster in MISR science team meeting at Pasadena, CA. [Poster]

more...