| 基于自适应堆滤波的边界检测 |
| |
| 引用本文： | 黎明,汪德宗.基于自适应堆滤波的边界检测[J].中国图象图形学报,1996(Z1). |
| |
| 作者姓名： | 黎明 汪德宗 |
| |
| 作者单位： | 南京航空航天大学测试中心 |
| |
| 摘 要： | 从被噪声干扰的图象中提取边界是图象测试与分析的关键之一。通常需要先滤除图象中的噪声，再用边界检测算子求出边界。本文介绍了一种边界直接检测法，即将边界检测与噪声滤波相结合，它是基于自适应堆滤波的边界检测法。首先非线性堆滤波器用于求出图象某象素点邻域内的灰度最大值与最小值的最优估计，然后以此两估计值之差代替原象素点灰度值。最后对之二值化求出边界。本文根据最小平均绝对误差准则，采用自适应方法求解堆滤波器。这种方法类似于线性自适应滤波器的ＬＭＳ方法，先任设一初始堆滤波器，利用期望图象与合噪声图象对堆滤波器进行迭代训练，最后求出最优化的自适应堆滤波器。文章最后给出了采用自适应堆滤波法求取图象边界的试验结果，表明这种方法可以有效地抑制各种分布的噪声干扰。
|
| 关 键 词： | 图象处理，边界检测，非线性滤波，自适应滤波，堆滤波器 |
Edge Detection Based on Adaptive Stack Filter |
| |
| Li Ming,Wang Dezong.Edge Detection Based on Adaptive Stack Filter[J].Journal of Image and Graphics,1996(Z1). |
| |
| Authors: | Li Ming Wang Dezong |
| |
| Abstract: | Detecting edge from noisy images is a key step for image test and analysis. Generally, filters are used to filter noises before edge detectors are applied. In this paper, a direct edge detection method which combines filtering and detecting is introduced. The edge detection is based on adaptive stack filter.First, nonlinear stack filter is applied to estimate the maximum and minimum of noise-free image's gray scale value in a local region. Then the difference of the maximum and minimum is thresholded to obtain edge map. An adaptive algorithm is used to obtain the stack filter under the mean absolute error criterion, it is similar to the LMS algorithm used to solve linear adaptive filter. An initial stack filter is randomly selectod, then noise-free images and noisy images are used to train the stack filter is randomly selected, then noise-free images and noisy images are used to train the stack filter until it is close enough to an optimal stack filter. In the end of the paper, a satisfactor result is given. It shows that. This method can be widely used in image test and analysis. |
| |
| Keywords: | Image processing Edge detection Nonlinear filtering Adaptive filtering Stack filter |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《中国图象图形学报》浏览原始摘要信息 |
|
点击此处可从《中国图象图形学报》下载免费的PDF全文 |
