A one-class classifier for identifying urban areas in remotely-sensed data
- New Mexico Univ., Albuquerque, NM (United States)
- Los Alamos National Lab., NM (United States)
For many remote sensing applications, land cover can be determined by using spectral information alone. Identifying urban areas, however, requires the use of texture information since these areas are not generally characterized by a unique spectral signature. We have designed a one-class classifier to discriminate between urban and non-urban data. The advantage to using our classification technique is that principles of both statistical and adaptive pattern recognition are used simultaneously. This prevents new data that is completely dissimilar from the training data from being incorrectly classified. At the same time it allows decision boundary adaptation to reduce classification error in overlap areas of the feature space. Results will be illustrated using a LANDSAT scene of the city of Albuquerque.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE; USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 5222192
- Report Number(s):
- LA-UR-92-1262; CONF-9205141-4; ON: DE92013539
- Resource Relation:
- Conference: Ideas in science and electronics (ISE) symposium, Albuquerque, NM (United States), 14 May 1992
- Country of Publication:
- United States
- Language:
- English
Similar Records
A detailed comparison of backpropagation neural network and maximum-likelihood classifiers for urban land use classification
Large-Scale Classification of Urban Structural Units From Remote Sensing Imagery