基于图像配准的激光散斑成像技术

Laser Speckle Contrast Imaging Based on Image Registration

激光散斑成像技术可以有效地监测血液流动,然而在活体实验当中由于动物的呼吸和心跳等因素,导致拍摄到的散斑图像产生晃动。针对上述问题,本文提出了一种图像配准算法,可以有效地消除图像抖动,不仅提高了细小血管的分辨能力,并且使图像信噪比提高了10%。此外,由于原有的算法运算太耗时,本文在对算法进行分析的基础上编写并行计算程序使运算速度提高了26倍,可以满足临床实时监测需求。

Laser Speckle Contrast Imaging shows great potential to monitor blood flow.However,fluctuation always happens due to animal breathing and heartbeat.An image registration algorithm was proposed to solve this problem,not only improving the resolution of small blood vessels,but also increasing the image signal-to-noise ratio by 10%.In addition,a parallel computing program based on the analysis of the algorithm was put forward to increase the operation speed by 26 times to meet the clinical needs of real-time monitoring.