基于curvelet变换的地震数据去噪研究-地球探测与信息技术专业论文.docxVIP

英文摘要Subject： 英文摘要 Subject： Curvelet—based seismic data denoise Speciality： Geodetection and Name： Cui Shitao (s Instructor：Xiao Zhongxiang(signature) ABSTRACT The Wavelet transformation has a favorable analysis ability of partial time-ffequency，and could well express one-dimensional partition smooth signal，but doesn’t suit the expression of two—dimensional signal．Although Wavelet could filter the noise，but actually lost the picture edge detail informations，and easily caused picture fog．Therefore a method that can both to be able to effectively reduced noise and Carl well reserve picture edge informations，is the goal that people always attempted to achieve． Ridgelet transformation inherited the good partial time—frequency analysis ability of wavelet transformation，and has a comparative strong selectivity of direction and discern ability．The ridgelet transformation can effectively express the irregularity characteristics of straight line，and have a good approach performance to the multivariable function of straight line irregularity,but regarding to the image curve edge’S description，it approach performance is only equal to wavelet transformation，and didn’t have the optimal misalignment approximate error of weaken step．the calve irregularity Was still a CHI Ve rather than a spot， and the irregularity of Wavelet expression will not be sparsed，SO as to its ridgelet expression． Curvelet Transform is one kind of new multi-criterion transformation which develops from wavelet transformation，and belongs to the category of sparse function expression theory． It’S structural element not only includes scale parameter and location parameter but also includes the position parameter,which makes Curvelet transform has a very good position characteristics．Therefore，the Curve l et transform Can better expression than Wavelet and Ridgelet transformation in image’S edge，such as geometry characteristics of curve，straight and SO on．This characteristic caused Curvele