Fourier Filtering of Landsat Data for Information Extraction in Surveying and Mapping

Various spatial filtering techniques have been developed to process digital Landsat data. This research aimed to filter the same digital data within the frequency domain, and involved the use of the Butterworth lowpass and highpass filters. The lowpass filter is used primarily for the reduction of noise whereas the highpass filter is used primarily for edge enhancement. The analysis of these two filters was based on a comparative approach, utilizing what had already been accomplished in the spatial domain. It was found that the Butterworth lowpass filter did an inferior job of reducing noise and keeping data distortions to a minimum than did the spatial filters. Also, the spatial filters clearly defined edges to greater detail than the highpass Butterworth filter. However these results only apply to the frequency filters used in this research. It cannot be inferred that all frequency techniques would be inferior to spatial filters.