Detection of Linear Structures in Remote-Sensed Images

Over the past decades, considerable progress had been made in developing automatic image interpretation tools for remote sensing. There is, however, still a gap between the requirements of applications and system capabilities. Interpretation of noisy aerial images, especially in low resolution, is still difficult. We present a system aimed at detecting faint linear structures, such as pipelines and access roads, in aerial images. We introduce an orientation-weighted Hough transform for the detection of line segments and a Markov Random Field model for combining line segments into linear structures. Empirical results show that the proposed method yields good detection performance.

• Type: Conference
• Year: 2009
• Where: ICIAR
• Authors: Rui Gao, Walter F. Bischof
• Keywords: Aerial Images, ICIAR 2009, Image, Line Segments, Linear Structures