Feature reduction using a singular value decomposition for the iterative guided spectral class rejection hybrid classifier |
| |
| Authors: | Rhonda D Phillips Layne T Watson Randolph H Wynne Christine E Blinn |
| |
| Institution: | 1. Department of Computer Science, 2050 Torgersen Hall, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, United States;2. Department of Mathematics, 2050 Torgersen Hall, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, United States;3. Department of Forestry, 313 Cheatham Hall, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, United States;1. School of Civil Engineering, School of Geosciences and Info-Physics, Central South University, 22 Shaoshan South Road, Changsha 410004, PR China;2. Department of Land Surveying & Geo-Informatics, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong;1. Electrical Engineering and Computer Science Department, South Dakota State University, SDEH 213, Box 2222, Brookings, SD 57007, United States;2. Microsoft Corporation, Bing Imagery Technologies R&D, Boulder, CO 80302, United States |
| |
| Abstract: | Feature reduction in a remote sensing dataset is often desirable to decrease the processing time required to perform a classification and improve overall classification accuracy. This paper introduces a feature reduction method based on the singular value decomposition (SVD). This SVD-based feature reduction method reduces the storage and processing requirements of the SVD by utilizing a training dataset. This feature reduction technique was applied to training data from two multitemporal datasets of Landsat TM/ETM+ imagery acquired over a forested area in Virginia, USA and Rondônia, Brazil. Subsequent parallel iterative guided spectral class rejection (pIGSCR) forest/non-forest classifications were performed to determine the quality of the feature reduction. The classifications of the Virginia data were five times faster using SVD-based feature reduction without affecting the classification accuracy. Feature reduction using the SVD was also compared to feature reduction using principal components analysis (PCA). The highest average accuracies for the Virginia dataset (88.34%) and for the Rondônia dataset (93.31%) were achieved using the SVD. The results presented here indicate that SVD-based feature reduction can produce statistically significantly better classifications than PCA. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
