BUILDING ROOF DETECTION FROM A SINGLE HIGH-RESOLUTION SATELLITE IMAGE IN DENSE URBAN AREAS

摘要

This work aims at extracting 2D buildings from a single high-resolution satellite image in densely built-up urban areas. The whole algorithm follows the hypothesis-verification strategy. The key contribution of this paper is the edge verification method of the hypothesis verification process, which can greatly improve the accuracy and precision of the building extraction results from complex scenes. To extract more accurate and precise edges, we construct a probabilistic model and an optimization frame. First, a probability being the optimal edge is given to each possible edge, then the constraints of any two possible edges are estimated based on a machine learning method and the image evidences, finally, these constraints and other prior knowledge are integrated into an optimization problem, by solving which these probabilities can be computed and the optimal edge can be selected. At last, we provide some experimental results on large and complex scenes that demonstrate the robustness and accuracy of our algorithm.