小波变换在图像边缘检测中的应用

The Application of Wavelet Transform to Edge Detection Image

研究一种新的提取图像边缘特征的方法．方法依据图像的多分辨率小波分解及其所表现出的多尺度特性、局部化分析特性及方向选择性等特点，通过二进尺度下小波变换局部极大值的检测提取图像边缘特征．结果利用此方法在计算机上对２５６×２５６×８ｂｉｔｓ／像素的Ｌｅｎａ图像做了边缘提取，得到不同尺度下的连续、光滑、单像素宽的边缘链图像．结论边缘提取准确，表明这种特征提取方法是有效、可行的．

Aim To study a new method of extracting features of the image. Methods According to the properties of its multi- scale, direction and local characteristic, the image edge features were provided by determining the local maxima of wavelet coefficients at dyadic scale of the image . Results The computer simulation was carried out on the 256× 256× 8 bits/ pixel Lena image, and the continuous and smooth edge image under multi-scale can be got. Conclusion The features of image are extracted accurately . The experiment result reveals that the method is efficient and practicable.