Real World Objects in GEOBIA through the Exploitation of Existing Digital Cartography and Image Segmentation

摘要

Descriptions of Geographic Object-Based Image Analysis (GEOBIA) often identify image segmentation as the initial step. This may be reasonable in some cases, but segmentation might also be considered a "black art," due to its image dependence and the limited amount control available to users. The resulting segments reflect the spectral structure of the image rather than the physical structure of the landscape with no one-to-one relationship between real world objects and segments. Geographic analysisoften begins in the context of existing mapping. In regions with high quality large scale cartography, an obvious question is why is this information not used in the GEOBIA process? It is therefore proposed that GEOBIA be redefined to use the best existing real world feature datasets as the starting point before segmentation is considered. Such an approach would increase opportunities for integration, improve map update initiatives, and widen uptake by end user communities.