基于Haar小波的图像变换方法的研究

系统地介绍了基于快速滤波器组卷积算法的第一代小波和基于提升方案的第二代小波。通过对图像先行后列分离,结合惰性滤波器进行奇偶分割,由欧几里得算法求出预测算子和更新算子,实现了由此提升方案构造的Haar小波1级图像分解。实验结果表明提升方案构造的Haar小波实现起来相当容易,简化了很多复杂的计算过程,其本身的诸多特点更便于在实际中应用,且在变换后的效果上,与第一代基本无差异。

Study of Image Transform Method Based on Haar Wavelet

Systematically introducing first generation wavelet based on the fast filter banks algorithm of convolution and second genera- tion wavelet based on the lifting scheme. Separating the image from row to column, combining lazy filter carries on the odd and even split and extracting the prediction operator and the update operator by the Euclid algorithm, which has accomplished 1 level of image decomposition by this the lifting scheme structuring Haar wavelet. The experimental result indicates that realization of the lifting scheme structuring Haar wavelet is quite easy, simplifies the very many complex computations process, and its itself many character- istics were advantageous in actual application. In the transformation effect, it basically does not have the difference with the first generation wavelet' s.