一种用于表面重建的外边界轮廓线提取算法

An Algorithm of Outside Boundary Contour Tracking for Surface Reconstruction

医学图像处理中边界轮廓线的提取是科学计算可视化技术中的一个研究热点,是计算机图形学和图像处理技术在生物医学工程中的重要应用。在医学图像的三维重建、模式识别等的应用中,常常需要保留感兴趣的物体区域,从而必须用外边界作为轮廓线。在分析Log算子作用原理的基础上,提出并实现了一种新的按光栅顺序扫描取点然后进行边界跟踪的轮廓线提取算法,最后按照最短对角线法进行轮廓连接。实验证明,算法可以有效地消除噪声,获得更加精确的象素级边界轮廓,重建效果较为满意。

The extraction of contours from medical images is a hot subject in the Visualization in Scientific Computing, and it is also an important application of Computer Graphics and Image Processing technology in the biology medical project. In the application of medicine image 3D reconstruction and pattern recognition, the interested object areas often need to be reserved. Thereby, the outside boundary must be used as contour line. Based on the analysis of the Log operator, a ereative algorithm, fetching points by means of raster scan and extracting contours by border tracking, is provided and realized, and lastly the shortest diagonal method is used for contour connection. Experiments prove that the new algorithm can remove noise, obtain accurate pixel edge contour and achieve satisfactory effect.