Title :
Research on road centerline extraction from aerial image based on Sorting
Author :
Guo, Bingxuan ; Li, Qi ; Shao, Yuanzheng
Author_Institution :
State Lab. of Inf. Eng. for Surveying, Mapping & Remote Sensing, Wuhan Univ., Wuhan, China
Abstract :
According to the road extraction from moderate and low resolution aerial image, this paper presents a fast method which is based on sorting. In the initial stage of the method, an algorithm for detecting ridge or ribbon like linear features based on sorting and simple judgment scheme is adopted; after thinning the centerline, using the initial image road features to check the correctness of the result, then adjusting the centerline to ensure the roads´ validity; at last, fitting the centerline based on the least square method and Newton method to get a proper road expression for user. Experiment showed, this method is simple to use, of high accuracy, after improvement can be used in the semi-automatic or automatic road extraction.
Keywords :
Newton method; feature extraction; geophysical signal processing; least squares approximations; object detection; road traffic; Newton method; least square method; low resolution aerial image; ribbon like linear feature extraction; ridge detection; road centerline extraction; sorting; Computer vision; Data mining; Image resolution; Least squares methods; Newton method; Object detection; Remote sensing; Roads; Solid modeling; Sorting; curve fitting; least square method; newton method; raod centerline detection; road centerline adjusting; self-check;
Conference_Titel :
Geoinformatics, 2009 17th International Conference on
Conference_Location :
Fairfax, VA
Print_ISBN :
978-1-4244-4562-2
Electronic_ISBN :
978-1-4244-4563-9
DOI :
10.1109/GEOINFORMATICS.2009.5293471
