Browse Bar: Browse by Author | Browse by Category | Browse by Citation | Advanced Search
Author(s): Carpenter, G.A. | Gopal, S. | Shock, B.M. | Woodcock, C.E. |
Year: 2001
Citation: Proceedings of the World Congress on Computers in Agriculture and Natural Resources, Igua?a Falls, Brazil, September, 2001.
Abstract: The ability to detect and monitor changes in land use is essential for assessment of the sustainability of development. In the next decade, NASA will gather high-resolution multi-spectral and multi-temporal data, which could be used for detecting and monitoring long-term changes. Existing methods are insufficient for detecting subtle long-term changes from high-dimensional data. This project employs neural network architectures as alternatives to conventional systems for classifying changes in the status of agricultural lands from a sequence of satellite images. Landsat TM imagery of the Nile River delta provides a testbed for these land use change classification methods. A sequence of ten images was taken, at various times of year, from 1984 to 1993. Field data were collected during the summer of 1993 at 88 sites in the Nile Delta and surrounding desert areas. Ground truth data for 231 additional sites were determined by expert site assessment at the Boston University Center for Remote Sensing. The field observations are grouped into classes including urban, reduced productivity agriculture, agriculture in delta, desert/coast reclamation, wetland reclamation, and agriculture in desert/coast. Reclamation classes represent land use changes. A particular challenge posed by this database is the unequal representation of various land use categories: urban and agriculture in delta pixels comprise the vast majority of the ground truth data available in the database. A new, two-step training data selection method was introduced to enable unbiased training of neural network systems on sites with unequal numbers of pixels. Data were successfully classified by using multidate feature vectors containing data from all of the available satellite images as inputs to the neural network system.
Topics:
Machine Learning,
Applications:
Remote Sensing,
Models:
ARTMAP,
