摘要
提出一种基于ridgelet变换的线性特征检测方法,用于提取在复杂背景下高速路面的线性裂纹,克服了小波变换只能处理点奇异性的局限性。利用Radon变换与一维小波变换相结合的算法实现ridge-let变换,提出了Radon变换重建原图像的基本条件。应用于实际的路面数据检测,结合默认阈值为27.939 5的方法去除噪声,结果表明,ridgelet变换在提取直线性裂纹及抑制点噪声方面优于经典小波变换的处理效果,其分辨率达到2 mm精度。
The method based on the ridgelet transform is proposed to detect the linear feature. The method is used to pick up the linear splits from the complex environment on the freeway surface,which is against the limitation of the wavelet. The ridgelet transform algorithm is implemented by combining the Radon and wavelet transform. The basic condition to reconstruct the original image is proposed. We applied the method to detect the actual surface data and initialize the default threshold (27. 939 5) to reduce the noise and obtained the better results than traditional wavelet at the respects of linear split picking-up and point-noise restraining. The effect shows that the resolution factor is within 2 mm.
出处
《光电子.激光》
EI
CAS
CSCD
北大核心
2006年第6期724-727,共4页
Journal of Optoelectronics·Laser
基金
南京理工大学青年学者基金(200401)的资助