激光共焦显微图像多尺度边缘提取算法研究  

Multiresolution Edge Detection of Confocal Microscopic Image

作　　者：黄向东[1] 谭久彬[1] 

机构地区：[1]哈尔滨工业大学超精密光电仪器工程研究所,黑龙江哈尔滨150001

出　　处：《光电子．激光》2005年第4期458-461,共4页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金资助项目(50275041)

摘　　要：为解决微小尺寸的快速精确测量问题,提出一种基于Shannonδ函数的多尺度图像边缘提取算法。该算法选定2个不同尺度的边缘提取算子,并利用自适应选取阈值算法,将不同尺度下的处理结果图像结合起来,得到最终的边缘图像。经过对大量的共焦显微图像的处理表明,该方法具有较高的抗噪性和准确性,可实现共焦显微图像尺寸的精确测量,性能优于传统微分算子,数据处理速度是Canny算法的2倍。In order to solve the problem of accuracy and speediness in Mico-structure measurements, an algorithm for multiresolution edge detection based on Shannon δ function was presented. The algorithm provides two different resolution operators for the detection of multiscale edge. The larger scale operator is effective against the image noise, and the other one is capable of extracting the fine details of the image. Furthermore, by using the self adapting threshold algorithm, edge images of different resolution are combined to attain the final edge image. Computer experiments were carried out for extracting edge information from many real images and sample images with the noise. Compared with other existing methods, It shows that the approach in the paper is effective and reliable. The anti-noise performance and accuracy of the algorithm is better. The processing data speediness of the method is two times of Canny algorithm's.

关 键 词：显微图像 多尺度边缘提取 算法研究 CANNY算法 激光 边缘提取算法 边缘提取算子 测量问题 微小尺寸 处理结果 边缘图像 精确测量 图像尺寸 共焦显微 微分算子 处理速度 Δ函数 自适应 准确性 抗噪性 

分 类 号：TN911.73[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]