基于提升方案的小波边缘检测算法研究

Research Wavelet Edge Direction Based on Lifting Project

通过构建一种提升小波来提取图像的边缘,根据小波的提升框架理论提出了一种基于线性提升小波的边缘检测。该方法首先构建一组小波预测函数,并推导了一个提升小波更新函数。通过对原始图像进行分裂、预测、更新运算,实现图像的小波分解。利用分解后的小波系数构成图像的小波梯度图和方向图,在小波方向图和梯度图的引导下实现了图像的边缘跟踪与准确定位,实验表明利用提升小波对图像进行边缘检测的方法是行之有效的。

By building up a set of lifting wavelet to direct image edge, put forward edge direction based on linear lifting wavelet according to wavelet＇ s lifting frame theory. At first this method built up a set of wavelet predictive function, and then deduced a linear update function. After a series of operations such as splitting, predicting and updating. Then the gradient map and direction map were obtained with the decomposed wavelet coefficient. According to the principle of direction profile detection, the images edges were extracted based on the gradient map and direction map. Finally, tracking and accurate localization for edges were realized. The experiments show that the method was effective to make image edge direction by using lifting wavelet.