Effects of aggregation methods on image classification

A major concern in scale- and resolution-related issues is to develop methods for determining the most appropriate scale and resolution of study and assessing their effects (Cao and Lam 1997). The choice of an appropriate scale, or spatial resolution, for a particular application depends on several factors. These include the desired information about the ground scene, the analysis methods to be used to extract the information, and the spatial structure of the scene itself (Woodcock and Strahler 1987). When an appropriate scale or resolution is determined, the next step is to get the corresponding images. Unfortunately, the resolutions of existing remote sensing satellite images are discrete and one may not be able to obtain an image with desired resolution (e.g. 7m). In this case, resampling techniques are often used to interpolate an image into desired resolution and aggregation is a particular resampling technique widely practiced for up-scaling image data from high resolution to low resolution (Bian and Butler 1999). It can be visualized that different aggregation methods may introduce different kinds of noise, create different kinds of mixed pixels and thus lead to different results. Therefore, inferring spatial data across scales is an important challenge faced by scientists (Wang et al. 2004).