CLOUDMAP2(#): Quantitative assessment of outlier detection methods with MISR-stereo height measurements of ground topography
Marie-Ange Denis, Jan-Peter Muller* and Carlos López**
*Professor of Image Understanding and Remote Sensing
MODIS & MISR Science Team Member (NASA EOS Project), HRSC Science Team Member (Mars Express 2003)
**Universitario Autónomo del Sur, Montevideo, URUGUAY
(#)Research Project co-funded by the European Commission Research Directorate-General 5th Framework Programme Environment and Sustainable Development Development of generic Earth observation technologies Contract No. EVG1-CT- 2000-00033
CLOUDMAP2: The project
- Develop the technology, including semi-operational processing systems, to extract cloud
parameters from existing and soon-to-be-launched European EO sensors in near real-time
(=3 hours) and ground-based remote sensing instruments in real-time.
- Create, based on this technology , a distributed database of cloud measurements
- Validate satellite cloud and WV products using ground-based remote sensing instruments
and through satellite-satellite intercomparison
- Assess, qualitatively and quantitatively, how this cloud database can be used to improve the accuracy and/or validation of Numerical Weather Prediction and Climate General
Circulation Models, specifically by providing statistical information
Stereo-Heights:
- Assess whether M4 and PGotcha stereo matching methods using a simple and fast height retrieval technique provide for clear-sky scenes over rugged terrain
- more accurate and reliable results than MISR 2TC.
- results similar to height measurements of ground topography.
- Assess whether blunder detection techniques can improve the stereo-matched heights
matchers.
- Selected cloud-free scenes over mountain ranges with heights up to 9km, with both East-West (Himalayas) and North-South (Rockies) orientation.
- Areas selected were those with good quality NIMA DTED-0 data including maximum heights at 30 derived from 3 DTED-1 data
- MISR 2TC Stereo-Height produced for AnAf camera pair on a 1.1km grid include correction of wind advection effects, irrespective of whether clouds are present.
- Applied M4 & PGotcha to 275m grid MISR 2TC L1B2 ellipsoid radiance for AnAf camera pair. PGotcha results given at sub-pixel level accuracy.
- Use simple equation (no wind correction) to retrieve Stereo-Heights
Results
- Blunder detection techniques improved the Stereo-matched Heights by reducing the number of blunders. However, they did not remove all of them and some good points were
- M4 not affected by the interpolation due to its good coverage and both blunder detection techniques work better
- Less blunders for PGotcha probably due to the sub-pixel
- Felicisimo shows better results for our scenes than Lopez (lower error values & less false alarms)
- Standard deviation & other parameters reduced but at the cost of an increase in bias
References cited
- *Felicisimo, A.M., 1994. Parametric statistical method for
error detection in digital elevation models, ISPRS Journal of
Photogrammetry and Remote Sensing, 49(4), pp. 29-33
- **Lopez, C., 2000, On the improving of elevation accuracy
of Digital Elevation Models: a comparison of some error
detection procedures, Transactions in GIS, 4(1), pp.43-64
- §Muller, J-P., A. Mandanayake, C. Moroney, R. Davies, D.
J. Diner, S. Paradise, 2002. MISR stereoscopic image
matchers: techniques and results, IEEE Transactions on
Geoscience and Remote Sensing, 40(7), pp.1547-1559
- §§Muller, J-P., M-A. Denis, R. D. Dundas, K. L. Mitchell,
H. Mannstein, 2003(submitted). Stereo cloud-top heights
and amount retrieval from ATSR2, International Journal of
Remote Sensing CLOUDMAP special issue.
Presented at: EGS-EGU-AGU joint Assembly, April 2003. Download the poster (143 kb)
Contact address: