利用三维重建进行粒子统计的方法  

Statistic Means of Particle Using 3D Reconstruction

作　　者：原飞[1] 肖德成[1] 任新宇[1] 刘畅[1] 王宗祺[1] 范勇[1] 

机构地区：[1]西南科技大学计算机科学与技术学院,四川绵阳621010

出　　处：《西南科技大学学报》2013年第1期70-74,98,共6页Journal of Southwest University of Science and Technology

基　　金：四川省教育厅基金项目(No.10ZC010);西南科技大学青年基金(11zx3117);西南科技大学研究生创新基金(12ycjj20)

摘　　要：针对单张二维图片难以准确完成粘连粒子计数与体积计算的问题,提出了一种基于重建的空间统计算法。首先对CT切片图像采用数学形态中的腐蚀处理,得到粘连粒子分离图像,然后采用Marching Cubes算法对切片图像进行三维重建,再将重建的粒子以STL数据格式保存,最后根据STL数据的特征和粒子的连通性完成统计。实验结果表明,当球形粒子随机噪声分布概率在20%以内时该方法能够准确完成粘连粒子计数和粒子体积估算。Aiming at the difficulty of calculating the volume and count of the adhesion particle based sin- gle 2D-image, a space statistical algorithm based on reconstruction was presented in this paper. Firstly, taking the erosion based on mathematical morphology of the CT slice images, the images of the separated adhesive particles were obtained. Secondly, 3 D reconstruction of the slice images was taken by using the Marching Cubes algorithm. Then, these reconstructed particles were saved with the STL data format. Fi- nally, the statistic was completed according to the features of STL data format and the connectivity of par- ticles. The experimental results show that this algorithm can accurately count the number and estimate the volume of the spherical particles within 20% random noise.

关 键 词：三维重建 MARCHING CUBES STL 连通性 统计 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]