Abstract: A processing chain, from coarseness to refinement, for the spatial analysis of the data recorded by ground penetrating radar (GPR) was presented according to the property of a buried cylinder giving rise to a hyperbolic signature in GPR images. Firstly fractal theory was applied to pick up the fuzzy object area from complex background of GPR images, and then the appropriate directional operators were adopted to locate edges accurately. Finally an improved Hough transform was used to fit curves and extract target hyperbolas. Result shows that hyperbolas in GPR images can be discovered and reconstructed successfully.

Key words: ground penetrating radar, GPR images, fractal, directional operator, Hough transform

摘要： 针对管道物体在探地雷达回波信号剖面图像(GPR图像)中表现为双曲线的特点,提出了一套由粗而精的图像处理方法。该方法首先运用分形理论从雷达图像的复杂背景中提取出模糊的目标区域,然后选用适当的方向算子对模糊区域进行边缘的精确定位,最后用改进的Hough变换进行曲线拟合,提取出目标双曲线。实验结果表明该方法能检测出图像中的目标双曲线并将其成功地恢复出来。

关键词: 探地雷达, GPR图像, 分形, 方向算子, Hough变换