CAJ | 학술논문

文中提出一种基于离散小波变换的针对数字图像边缘的有效检测算法。主要利用小波变换的多尺度分辨特性，对数字图像进行小波变换后的水平方向、垂直方向以及对角方向的细节信息进行提取，然后对图像中的边界点进行探测，达到提取出边界信息的目的。在处理过程中，首先对除噪以后的图像作一级2-D小波变换分解，得到一个低频和三个高频部分；然后分别计算水平和垂直部分的绝对值均值，再分别计算水平和垂直方向的绝对值标准方差，根据绝对值标准方差构造四个新的二值矩阵，根据构造的矩阵修正水平和垂直方向细节信息，最后利用离散小波逆变换重构，得到检测后的图像。将重构后的图像进行边界提取。实验结果表明，算法效果良好。
문중제출일충기우리산소파변환적침대수자도상변연적유효검측산법。주요이용소파변환적다척도분변특성，대수자도상진행소파변환후적수평방향、수직방향이급대각방향적세절신식진행제취，연후대도상중적변계점진행탐측，체도제취출변계신식적목적。재처리과정중，수선대제조이후적도상작일급2-D소파변환분해，득도일개저빈화삼개고빈부분；연후분별계산수평화수직부분적절대치균치，재분별계산수평화수직방향적절대치표준방차，근거절대치표준방차구조사개신적이치구진，근거구조적구진수정수평화수직방향세절신식，최후이용리산소파역변환중구，득도검측후적도상。장중구후적도상진행변계제취。실험결과표명，산법효과량호。
An effective detection algorithm is proposed based on discrete wavelet transform for digital image edges. In this paper, the wavelet multiscale resolution is used for its characteristics of horizontal,vertical and diagonal direction in order to extract the details,and then detect the boundary points of image,which will get the purpose for extracting boundary information. In the process,firstly,transfor-ming the processed image without less noise into wavelet transform decomposition,the result is that a low frequency and three high fre-quency components are produced. The next work is calculating the mean of the absolute value of the horizontal and vertical portions re-spectively,and calculating the standard deviation of horizontal and vertical direction,according to the standard deviation,four new binary matrix are constructed,the image matrix is corrected according to the horizontal and vertical detail information,the image will be obtained by the way of the inverse wavelet transform. And the boundary extraction is finished on the reconstructed image. The test results show that this method is effective.