欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
提出了一种基于Contourlet变换域统计模型的图像去噪算法,用于复杂背景下的路面病害图像,可以获得较为清晰的裂纹病害信息。利用χ2统计假设检验方法得出Contourlet变换系数近似服从拉普拉斯分布,进而应用最大后验贝叶斯估计推导出对含噪图像Contourlet系数的萎缩公式。实验结果表明,该算法能较好地实现噪声抑制和信号保留间的平衡,与小波去噪算法比较,其去噪后图像的信噪比提高4 dB。
英文摘要：
A novel image denoising algorithm based on the satistical model in Contourlet transform domain is proposed,which can obtain comparatively clear pavement crack information under the complicated background.The conclusion that the Contourlet coefficients obey the Laplacian distribution is derived on the statistical theory of χ2 hypothesis testing.Moreover,the shrink formula for the Contourlet coefficients of noisy images is deduced based on Bayesian maximum posterior estimation.The experiment results show that...
同期刊论文项目
同项目期刊论文
The spontaneous emission rate of an atom in a two-layer dielectric slab between two parallel mirrors
The recurrence spectra of the Rydberg hydrogen atom interacting with a metallic surface and an elect
期刊信息
位置：