Semi-automatic road detection from satellite imagery

In this work, we present a semi-automatic road extraction scheme to detect roads from low resolution satellite imagery. In our approach, we used the a trous algorithm to locate and identify roads. Principle component analysis (PCA) is used to combine the detail coefficients. After detecting possible road pixels, we used a graph searching algorithm to draw a final complete road network. We found that our approach leads to an effective method to form the basis of a road detection approach.