We developed algorithms to detect roads from aerial images taken from an Unmanned Aerial Vehicle. A histogram-based adaptive threshold algorithm is used to detect possible road regions in an image. A probabilistic hough transform based line segment detection combined with a clustering method is implemented to further extract the road.
Figure 1: Results of every step from possible road region selection to road separation. (b)
is the result after pre-procession. (c) is the result after histogram based thresholding.
(d) is the result of line segment detection. (e) is the result of line selection. (f) is the
result from points clustering.
We varied the line segments selection methods to detect “curved” roads in desert
Figure 2: Varying the line segments selection methods to detect curved roads in deserts.
For urban roads, we use a Naive-Bayes classifier instead of histogram-based thresholding to select possible road regions.
Figure 3: Some results for urban roads.