| 沈瑜,党建武,王阳萍,冯鑫.基于NSCT域压缩感知模型的路面病害图像滤波算法[J].光电子激光,2014,(8):1620~1626 |
| 基于NSCT域压缩感知模型的路面病害图像滤波算法 |
| A compressed sensing pavement distress image filtering algorithm based on NSCT domain |
| 投稿时间：2013-07-29 |
| DOI： |
| 中文关键词: 滤波 压缩感知(CS) 非下采样Contourlet变换(NSCT) |
| 英文关键词:filtering compressed sensing (CS) nonsubsampled Contourlet transform (NSCT) |
| 基金项目:国家自然科学基金(60962004,61162016,61202314)、甘肃省科技攻关计划基金(0708GKCA 047)、甘肃省青年科技基金(145RJYA249)、甘肃省科技计划(1310RJZA057)和兰州交通大学青年基金(2012003)资助项目 (1.兰州交通大学 电子与信息工程学院,甘肃 兰州 730070； 2.重庆工商大学 机械工程学院,重庆 400067) |
|
| 摘要点击次数: 1378 |
| 全文下载次数: 122 |
| 中文摘要: |
| 针对目前路面图像滤波算法复杂度高且难以耦合噪 声抑制和信号平衡的缺点,提出一种基于非下采样Contourlet变换(NSCT)域压缩采样的滤波 算法。首先,使用NSCT对含噪路面病害图像进行分 解,得到变换后的低频子带系数和高频子带系数；然后,对高频子带系数建立压缩感知(CS) 去噪模型,并采用伪随机傅里叶矩阵 对系数进行观测,之后使用分裂Bregman迭代方法对系数进行重构,得到去噪模型重建后的 高频子带系数；最后,采用逆NSCT对低频子带系数和高频子带系数进行重构,得到滤波后的 图像。实验结果分析表明了本文算法的有效性。 |
| 英文摘要: |
| Pavement distress image filtering algo rithm is very important for the pavement distress automatic detection technology . According to the imbalance between signals and noises,this study proposes a nov el compressed sensing filtering algorithm in nonsubsampled contourlet transform (NSCT) domain.Firstly the NSCT is adopted to conduct the multi-scale and multi -direction decomposition on the noisy pavement distress image.The decomposition results can give the low-frequency coefficients and the high-frequency coefficients.Secondly,the high-frequency coefficients are denoi sed by the compressed sensing algorithm.In the compressed sensing filtering algorithm,the high-frequency coefficients are ob served by the pseudo-random Fourier matrix and reconstructed by the split Bregman iteration method to get the optimum high-fre quency coefficients.Finally,the optimum high-frequency coefficients and the low-frequency coefficients are reconstruc ted by the inverse NSCT to get the denoised image. The experiment results show that the pavement distress image filtering algorith m in this paper improves the effectiveness of the filtering. |
| 查看全文 下载PDF阅读器 |
| 关闭 |
|
|
|
|
|