An improved technique for the prediction of optimal image resolution(s) for large-scale mapping of savannah ecosystems

Abstract

Past studies to predict optimal image resolution required for generating spatial information for savannah ecosystems have yielded different outcomes, hence providing a knowledge gap that was investigated in the present study. The postulation, for the present study, was that by graphically solving two simultaneous equations of values of image noise index (INI) and degradation level Index (LDI), a robust technique for predicting optimal image resolution for the mapping of savannah ecosystems was developed. The technique involved simulating 0.5 m imagery to different spatial resolutions for two savannah test sites in Masaka district, Southern Uganda. By plotting INI and LDI values against the simulated image resolutions, it enabled the authors to objectively predict that image resolution at 2.25 and 2.5 m were optimal for generating spatial information for savannah ecosystems represented by the two test sites. The new technique will enable vegetation ecologists to objectively determine optimal resolution(s) prior to the choice of imagery, offered by different high-resolution air- and space-borne sensors, for generating spatial information for savannah ecosystems. Future research should focus on using the new technique to determine what ranges of image resolutions are optimal for generating spatial information of different savannah ecosystems in different countries.

Key words: Optimal resolution, savannah ecosystems, image noise index, land cover index, level of degradation index.