Management of multi-modal data using the Fuzzy Hough Transform: Application to archaeological simulation

Management of multi-modal data is a classical problem in data mining and exploration. Geographic Information Systems usually store multi-modal data and employ them in spatial and image analysis. This paper proposes a fusion of multi-modal data in the context of line detection with Fuzzy Hough Transform (FHT). The goal of this work is to simulate the Roman streets according to periods from partial and uncertain excavation information. These simulations are based on three modalities : coordinates, orientations and dates. Data are stored into a geodatabase. As Roman streets form usually segments of lines, they are detected using three FHTs. The final visualization is obtained after merging the three FHTs. The proposed method gives a new tool helping archaeologists into their expertise and diagnostic processes. We validate it with Reims Roman streets excavations (France).