CAJ | 학술논문

具有非线性、非平稳性特征的肾脏断面图像信号分为高频和低频两部分，传统的轮廓线提取方法没有针对不同部分采用不同方法进行提取，提取的轮廓线图象存在大量冗余数据且边缘并不清晰。针对肾脏断面图象的特点，提出一种EMD和小波分析及快速模糊算法的虚拟肾脏断面解剖信息轮廓线提取方法，避免了传统方法因为没有划分而造成提取不精的弊端，利用EMD算法将原始虚拟肾脏断面信号加噪后的图像每行（列）分解成个模态函数，根据其对图像质量的权重选取滤波后的高频中适当的阈值与低频叠加，得到去噪后的肾脏断面图像，利用小波分析算法提取图像高频部分轮廓线，利用快速模糊算法得到完整肾脏断面图像轮廓线。通过实验结果表明，提出的算法能有效的去除图像随机噪声，提取出的虚拟肾脏断面解剖信息轮廓线较其他方法更为准确和清晰。
구유비선성、비평은성특정적신장단면도상신호분위고빈화저빈량부분，전통적륜곽선제취방법몰유침대불동부분채용불동방법진행제취，제취적륜곽선도상존재대량용여수거차변연병불청석。침대신장단면도상적특점，제출일충EMD화소파분석급쾌속모호산법적허의신장단면해부신식륜곽선제취방법，피면료전통방법인위몰유화분이조성제취불정적폐단，이용EMD산법장원시허의신장단면신호가조후적도상매행（렬）분해성개모태함수，근거기대도상질량적권중선취려파후적고빈중괄당적역치여저빈첩가，득도거조후적신장단면도상，이용소파분석산법제취도상고빈부분륜곽선，이용쾌속모호산법득도완정신장단면도상륜곽선。통과실험결과표명，제출적산법능유효적거제도상수궤조성，제취출적허의신장단면해부신식륜곽선교기타방법경위준학화청석。
With nonlinear non-stationary characteristics of renal section image is divided into low frequency and high fre?quency part of signal, the traditional contour extraction method is no different parts according to the different methods for extraction, extraction of contour image exist a large number of redundant data and the edge is not clear. According to the characteristics of the kidney section image, based on the contour line extraction is more clear and accurate information, and put forward a kind of EMD and wavelet analysis and fast fuzzy algorithm of contour extraction method, avoid to extract the traditional method because there is no division and the disadvantages of not jing, original virtual kidney section after the sig?nal and noise image in each row (column) decomposed by the EMD algorithm was the modal function, according to their weight after selecting filter on image quality of a proper threshold in the high frequency and low frequency superposition, get the kidney section of after denoising, image and high frequency part by using wavelet analysis algorithm to extract im?age contour line, by using fast fuzzy algorithm to extract image low frequency part contour line, using the direct integration method to incorporate two parts contour line, get complete kidney cross-section image contour line. Through the experimen?tal results show that the proposed algorithm can effectively remove the random noise of the image, to extract the virtual kid?ney section anatomy contour information than other method is more accurate and clear.