近红外静脉图像识别及处理算法研究

发布时间：2018-04-11 00:40

本文选题：近红外技术 + 静脉成像　；参考：《西安科技大学》2017年硕士论文

【摘要】：静脉穿刺是现代医疗技术中最常见的一种医疗手段,广泛应用于医疗行业中,包括静脉注射、静脉输液、静脉采血与输血等。静脉穿刺的首要条件是能够观察到较为清晰的静脉血管,而现实生活中由于患者个体皮肤颜色、静脉深浅以及脂肪厚度不同的影响,常常造成静脉穿刺失误,过度依赖医护人员的经验。因此,研究静脉图像识别及处理算法并用于辅助静脉穿刺等医疗应用领域具有很重要的实际应用价值。利用近红外光线的敏感性,可得到近红外静脉图像,但原始图像存在光照不均匀、对比度低、边缘不够清晰等问题,原始图像不能直接应用于辅助静脉穿刺,本文通过搭建近红外静脉图像采集装置,采集静脉图像,并运用改进图像处理算法进行处理,提取了静脉边缘图像,又提出了应用灰度对应方法测量手背静脉血管皮下深度的测量方法。本课题首先利用静脉及周围组织对近红外光线不同程度的吸收与反射原理,搭建并完善基于FPGA的非介入式静脉成像装置,包括设计装置固定支架、改装双层亚克力光学扩散板、加装近红外滤光片,然后通过研究静脉图像采集条件,将采集到的较为清晰的近红外静脉图像传输到计算机系统,再对图像做进一步的实时处理。在图像处理时,一是提出自动提取ROI感兴趣区域算法,便于对提取的静脉图像进行滤波去噪、图像增强和边缘检测;二是提出了改进算法:限制性中值滤波去噪算法、基于幂律变换的直方图均衡化对比度增强算法、基于Otsu阈值分割的逻辑运算边缘检测算法;处理效果显示改进算法提取到了清晰的静脉边缘;三是通过图像质量评价标准对改进算法做出评价,评价结果表明改进算法与传统算法比较具有一定的优势,改进算法简单灵活,其处理结果达到预期效果;四是提出了通过灰度对应方法测量人体手背静脉血管皮下深度的测量方法。通过自行设计制作的人体手背静脉模型,将手背模型图像的静脉灰度值与手背模型静脉血管皮下深度值进行标定、曲线拟合及校准,最终可定量的给出手背静脉血管皮下深度的直观差异,并且对测得血管深度进行实时的标注,方便医护人员进行静脉穿刺时参考判断,达到辅助医疗应用目的。
[Abstract]:Venipuncture is the most common medical method in modern medical technology. It is widely used in medical industry, including intravenous injection, intravenous infusion, venous blood collection and blood transfusion.The primary condition of venipuncture is to be able to observe a clearer vein vessel, but in real life, due to the influence of the skin color, the depth of vein and the thickness of fat, the venipuncture is often caused by the mistake of venipuncture.Excessive reliance on the experience of health care workers.Therefore, it is of great practical value to study the algorithms of vein image recognition and processing and to apply them to the medical applications such as venipuncture.The near infrared vein image can be obtained by using the sensitivity of near infrared ray, but the original image has some problems, such as uneven illumination, low contrast and not clear edge, so the original image can not be directly applied to assist venipuncture.In this paper, a near infrared vein image acquisition device is built to collect vein image, and an improved image processing algorithm is used to extract the vein edge image.A method for measuring the subcutaneous depth of the dorsal hand vein is proposed.Based on the principle of the absorption and reflection of near-infrared light from veins and surrounding tissues, a non-interventional venous imaging device based on FPGA is built and perfected, including the design of a fixed support and the modification of a double-layer subcrine optical diffusion plate.The NIR filter is added, and then the clear NIR venous image is transmitted to the computer system by studying the condition of venous image acquisition, and then the image is processed in real time.In image processing, an algorithm for automatically extracting ROI region of interest is put forward to facilitate filtering and denoising, image enhancement and edge detection of extracted venous images; second, an improved algorithm: restrictive median filter denoising algorithm is proposed.The histogram equalization contrast enhancement algorithm based on power law transformation, the logic operation edge detection algorithm based on Otsu threshold segmentation, the improved processing effect display algorithm extract the clear vein edge;Third, the improved algorithm is evaluated by image quality evaluation standard. The evaluation results show that the improved algorithm has some advantages compared with the traditional algorithm, the improved algorithm is simple and flexible, and its processing results reach the expected results.Fourth, the method of measuring the subcutaneous depth of human dorsal hand vein vessel by gray correspondence method is put forward.Through the self-designed human dorsal hand vein model, the grayscale value of the vein image of the back of hand model and the subcutaneous depth value of the vein vein of the back of hand model were calibrated, fitted and calibrated.Finally, the visual difference of the subcutaneous depth of the blood vessel of the dorsal vein of the hand can be quantitatively given, and the measured depth of the blood vessel can be labeled in real time, which is convenient for the medical staff to make the reference judgment when the venipuncture is carried out, so as to achieve the purpose of auxiliary medical application.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R318;TP391.41

【参考文献】