摘要
提出了一种基于视频图像序列的人体微小管状血管血液流速自动测量方法。在提取血管中心线时采用基于图像帧序列方差的种子点自动选取方法，设计了基于Hessian矩阵的血管中心线提取的迭代算法；提取血管中心线后，根据连续帧图像的中心线灰度统计曲线，计算出血液流速。结果表明：该方法测量速度快、精度高，与专家测量结果之间的相关系数为0．9014。
An automatic measurement of micro-vessel blood velocity based on video image frames is proposed. In order to extract vessel centerline automatically, the method of seed point auto selection is proposed using the pixels variance between image frames, then the centerline detection iterative algorithm based on Hessian matrix is designed. After the extraction of vessel centerline, the blood velocity distribution is calculated accurately with knowledge of the graylevel curve of centerline and time cousmnption between frames. Experiment result shows the proposed method is a rapid and precise detection one, which compared with expert detection result with 0.901 4 correlation coefficient, and could be a useful technique for human micro-vessel blood velocity rreasurement.
出处
《计量学报》
CSCD
北大核心
2010年第2期123-126,共4页
Acta Metrologica Sinica
基金
国家自然科学基金（10172043）
教育部博士点专项基金（20040287012）
国际合作项目（BZ2008060）
