An Improved Linear Target Detection Method Based on Probabilistic Hough Transform in a Remote Sensing Image

摘要

Regular linear targets detection in a remote sensing image depends on how to extract the long straight edges of the target from the background efficiently. Hough Transform is a classic method for line detection. But Standard Hough Transform (SHT) may induce too many short lines in the complex background to distinguish the true long linear targets. In this paper, an improved method is proposed, which uses probabilistic Hough Transform (PHT), a faster version of SHT, to detect short lines and then cluster short lines into groups to avoid most of the interfering lines. By this means, the reliability of target detection is improved. Experiments show that our method can detect the linear target efficiently when the background is large and complex.