Quantitative Analysis on Pipelines from GPR Images Based on Improved Hough Transform

摘要

In this paper, we propose a quantitative analysis method to identify pipelines from GPR image. According to the wave form characteristics reflected from buried utilities, Hough Transform algorithm had been improved by adjusting transformation parameters. Combined with extraction of region of interest, the attribute of underground utilities, including position, depth and diameter, could be automatically identified. To illustrate the performances of the proposal, we conducted a thorough experimental study based on synthetic GPR images and real GPR data for pipelines detection. The results show that the improved HT algorithm exhibits robust and rapid recognition performance.