Panchromatic satellite data currently have spatial resolutions of 2.0 (SPIN), 5.8 m
(Indian Remote Sensing Satellite, IRS) and 10 m (Satellite Pour' Observation de la Terre,
SPOT), and these data are valuable to deduce the location of roads, canals, ditches and
boundaries of individual fields, and small command areas. Black and white panchromatic
images are useful for describing cartographical changes due to construction and
encroachment of built-up areas or deserts. Crop identification at small farm plots
(<0.5 ha) should be done with multi-spectral images. Ground truth is imperative for
upgrading the information from spectral radiance to land cover and specific crop types.
The growth of GIS has greatly enhanced the opportunity to integrate conventional and
remote sensing data to form the basis for the development of digital expert systems (e.g.,
Kontoes et al., 1993). Computation by powerful systems allow the application of the
newest mathematical object classification techniques (neural networks, fuzzy logic,
probabilistic segmentation) to discern object classes. The overall transferability of
classification methods is very limited and a general guideline to classify land cover and
crop types does not exist.
การแปล กรุณารอสักครู่..