论文部分内容阅读
为解决传统的整数阶图像增强方法在路面裂缝图像处理效果中的不足,研究了一种新的基于分数阶偏微分的路面裂缝图像增强新模型。首先,针对现有的路面裂缝图像增强方法存在的问题进行分析,提出了采用分数阶微分理论解决路面裂缝边缘信息难以较好保留的问题;其次,根据分数阶微分差分表达式,以Prewitt算子的水平和竖直方向3×3模板为基础,推导出了一种新的5×5梯度掩模;最后,用该模型和传统算子分别对路面裂缝图像进行增强处理。试验结果表明该模型不仅具有大幅提升信号高频成分,增强信号中频成分、非线性保留信号的低频等特性,而且在边缘处达到极值检测边缘,去掉部分伪边缘,对噪声具有平滑作用。
In order to solve the shortcomings of traditional integer image enhancement methods in pavement crack image processing, a new model of pavement crack image enhancement based on fractional partial differential is studied. First of all, according to the existing problems of the pavement crack image enhancement method, the problem of using the fractional differential theory to solve the problem that the pavement crack edge information is difficult to be preserved well is proposed. Secondly, according to the fractional differential differential expression, the Prewitt operator Based on the horizontal and vertical 3 × 3 templates, a new 5 × 5 gradient mask was deduced. Finally, the model and the traditional operator were used to enhance the image of the cracked surface. The experimental results show that the model not only has the characteristics of greatly improving the signal high frequency components, enhancing the intermediate frequency components of the signal, and preserving the low frequency of the nonlinear signal, but also reaching the edge of extreme detection at the edge, removing some pseudo edges and smoothing the noise.