激光散斑信号处理方法研究

发布时间：2018-04-19 13:17

本文选题：激光散斑成像 + 动静脉分割　；参考：《华中科技大学》2012年博士论文

【摘要】：了解脑组织不同成分在正常生理状态下或者病理状态下的动态响应,对研究神经血管耦合机制、脑部疾病的诊断预防、药物评估以及术中监测等都有重大意义。激光散斑衬比成像技术以其简单的成像系统、高时空分辨率、无需扫描的且非接触的二维流速全场成像等特点,在脑皮层功能研究、视网膜疾病和皮肤疾病的诊断和治疗、药物评价等领域的血流动态变化检测中取得了重大应用,是生命科学基础研究和临床诊断的重要工具。但是该技术本身不具备动静脉分割能力,不能独立的实现脑皮层不同成分的血流变化分析。基于光学影像的动静脉分割技术是提取组织不同成分的一个有效技术手段,对于基础研究、临床诊断、术中监测等有很重大的意义。不过,目前的动静脉分割技术和激光散斑衬比成像技术结合,会出现很多问题,比如降低系统时间分辨率、使成像系统结构变得复杂。因此,本文对激光散斑数据所携带的信息进行进一步分析：通过分析激光散斑强度的分布,提出了基于单波长激光散斑相对最小反射率(Relative temporal minimum reflectance, RTMR)分析的动静脉分割法,该方法具有简单、自动以及准确率高等优点。该技术结合激光散斑衬比成像,在不改变激光散斑成像系统地前提下,实现了脑皮层不同组织成分独立的血流变化分析。主要内容如下： (1)提出一种自动有效的基于单波长激光散斑相对最小反射率分析的动静脉分割法。衬比值比较小,并且CCD记录的散斑图像互相统计独立时,Rayleigh分布函数是实际的积分散斑强度概率密度函数的一个很好的近似。通过Rayleigh分布函数得出激光散斑最小光强的表达式,发现它是平均光强和血流速度的整合参数。在激光散斑最小光强图内,根据激光散斑最小光强值,对血管进行分类。不过由于平均光强分布和入射光强有关,因此激光散斑最小光强存在背景不均匀现象,为此引入激光散斑相对最小反射率这个参量。空间任意一点的激光散斑相对最小反射率定义为：该点的激光散斑最小光强除以该点在激光散斑时域内均图内的邻域的非血管脑皮层组织的平均光强。在RTMR图中：动脉血管区域值最大；静脉血管区域值相对最小,部分可能和非血管皮层区域重叠；非血管脑皮层区域的RTMR值在了两类血管之间。为了避免误判静脉区域,所以先通过RTMR值采用多阈值法逐个像素的提取动脉网络(即对于血管内的一个像素,如果它的RTMR值比它的邻域内非血管脑组织像素的RTMR值平均值大,则该像素属于动脉),然后从血管网络图中减去动脉结构,得出静脉网络。实验证明基于单波长激光散斑相对最小反射率分析的动静脉分割法的准确率高,动脉的PTR (True Positive rate)是98.5%,静脉的PTR是95%,动脉的误判率是1.5%,静脉的误判率是5%。同时分析了激光散斑相对最小反射率分析的有效波长,成像深度等参数,其中最佳波长可能是600nm,成像深度大概是几百微米。 (2)基于单波长激光散斑相对最小反射率分析的动静脉分割法和激光散斑衬比成像技术结合,实现大鼠脑皮层不同成分组织在CSD模型中的独自的血流变化分析。准确地显示了微小区域内,微动脉、微静脉以及非血管脑皮层组织各自在CSD过程中的血流变化情况。
[Abstract]:It is of great significance to understand the dynamic response of different components of the brain under normal physiological or pathological conditions. It is of great significance for the study of the mechanism of the neurovascular coupling, the diagnosis and prevention of brain diseases, the evaluation of drugs and the monitoring of the operation. The non contact two-dimensional full field imaging is an important application in the research of cerebral cortex function, the diagnosis and treatment of the retina disease and the skin disease, the drug evaluation and so on. It is an important tool for the basic and clinical diagnosis of life science. However, the technique itself does not have the ability to divide the veins. The analysis of blood flow changes in different components of the cerebral cortex can not be realized independently. The technique of arteriovenous segmentation based on optical images is an effective technique for extracting different components of tissue. It is of great significance for basic research, clinical diagnosis, and intraoperative monitoring. However, the pre eye arteriovenous segmentation and laser speckle contrast imaging With the combination of technology, there will be many problems, such as reducing the time resolution of the system and making the structure of the imaging system complicated. Therefore, the information carried by the laser speckle data is further analyzed in this paper. By analyzing the distribution of laser speckle intensity, the relative minimum reflectivity (Relative temporal mini) based on the single wavelength laser speckle is proposed. Mum reflectance, RTMR) analysis of the method of arteriovenous segmentation. This method has the advantages of simple, automatic and high accuracy. This technique combines laser speckle contrast imaging and without changing the laser speckle imaging system, and realizes the analysis of the independent blood flow of different tissues in the cerebral cortex. The main contents are as follows:
(1) an automatic and effective dynamic vein segmentation method based on the relative minimum reflectance analysis of single wavelength laser speckle is proposed. When the contrast ratio is small, and when the CCD recorded speckle images are independent of each other, the Rayleigh distribution function is a good approximation of the actual intensity probability density function of the integral speckle intensity. The Rayleigh distribution function is obtained. The expression of the minimum light intensity of laser speckle shows that it is an integral parameter of the average intensity of light and the velocity of blood flow. The minimum intensity of laser speckle, Gon Jon Tu Hei, is classified according to the minimum intensity of laser speckle. However, the minimum intensity of the speckle is inhomogeneous due to the average intensity distribution and the incidence of the incident light intensity. The relative minimum reflectivity of the laser speckle is introduced. The minimum reflectance of the laser speckle at any point of space is defined as the minimum intensity of the laser speckle intensity divided by the mean light intensity of the non vascular cortical tissue in the neighbourhood of the laser speckle in the time domain of the laser speckle. In the RTMR diagram, the regional value of the arterial blood vessel is the largest; The regional values of venous vessels are relatively minimal and partially overlap with the non vascular cortex regions; the RTMR value of the non vascular cortical regions is between the two types of blood vessels. In order to avoid misjudging the venous area, the arterial network is extracted pixel by pixel by the multi threshold method by the RTMR value (that is, a pixel in the blood vessel, if its RTMR ratio is compared. " The average value of the RTMR value of the adjacent non vascular brain tissue is large, and the pixel belongs to the artery. Then the arterial structure is subtracted from the vascular network map and the vein network is obtained. The experimental results show that the accuracy of the arteriovenous segmentation based on the relative minimum reflectance analysis of single wavelength laser speckle is high, and the PTR (True Positive rate) of the artery is 98.5%, The PTR of the vein is 95%, the error rate of the artery is 1.5%, the error rate of the vein is 5%. and the effective wavelength of the relative minimum reflectance analysis of the laser speckle, the imaging depth and other parameters, of which the best wavelength may be 600nm, and the imaging depth is probably hundreds of microns.
(2) based on the combination of the single wavelength laser speckle relative minimum reflectivity analysis and the laser speckle contrast imaging technique, the individual blood flow changes in the CSD model of the rat cerebral cortex were analyzed. The microarteries, the venules and the non vascular cortical tissues were respectively in the CSD. The change of blood flow in the process.

【学位授予单位】：华中科技大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：R318.51

【引证文献】