浅表软组织超声信号处理与皮下脂肪厚度测量研究

发布时间：2018-08-11 16:19

【摘要】：超声波技术近年来迅速发展并广泛应用于医学诊断、治疗和工业检测领域。与生物组织作用后的超声波，是反映生物组织性质的一个信息载体。在生物医学超声基础研究中，探索超声波信息载体与生物组织间的联系，寻找超声波中体现组织的一些结构特征参量，一直都是人们研究热点课题之一。实际中的超声诊断技术，例如超声诊断仪，探头从体表接收到的体内超声信号是浅表软组织（皮肤、脂肪和肌肉等）、体内组织和超声系统相互作用的结果，这就为体内组织结构特征识别带来困难；另一方面，超声诊断仪获得的信号大都是经过了检波处理，是过滤了的信号，丢失了很多组织性质和结构特征信息。因此，，对于浅表软组织的超声检测信号处理，不管从理论还是实际运用来讲均应属于研究的热点。针对浅表软组织结构的超声检测研究并未深入，使得浅表界面结构识别、定位还不准确情况，本文提出了“皮肤-浅层脂肪-浅筋膜-深层脂肪-深筋膜-肌肉”的浅表分层结构，利用信号的卷积模型分析浅表界面信号特征。研究课题的目的有两方面，一方面，本文对浅表软组织分层结构的超声检测课题，利用信号处理技术探讨界面的信号特征提取和识别；另一方面，研究课题尝试使用超声波检测技术研制脂肪厚度测量仪，测量人体局部皮下脂肪厚度，从而随时检验运动和节食效果，更期望为超声吸脂提供一种监测手段。本课题完成的主要工作、成果和创新点总结有以下几点：对浅表（皮肤、脂肪、筋膜和肌肉）的生物组织特性及结构分布特征进行分析，结合浅表超声检测的射频回波信号，找出各反射回波代表的浅层组织界面。利用脉冲反射法测量界面之间的距离，即各组织层厚度，将测量结果与B超诊断仪、直尺作对比，从而确定组织界面回波，为进一步界面信号特征提取做准备。虽然软组织间的声阻抗相差不大，但浅表软组织有其特殊的结构特征，特别是浅筋膜和深筋膜处的界面薄而且界面数等因个体不同各异，造成超声回波在这两处信号特征明显，因此我们可以通过这两处的回波信号特征识别来测定人体脂肪厚度。为了提取筋膜界面信号特征，本文从理论上利用信号卷积模型对界面信号相互叠加效应进行解释。在此基础上，通过Matlab软件对信号卷积模型进行模拟验证，并将卷积模型应用于离体猪浅表组织界面信号的特征识别。实验结果证实了振荡波数作为信号特征识别参量的可行性。超声射频信号携带了浅表软组织的大量结构特征信息，但由于探头发射波形有一定的持续时间以及受检测系统的影响，而在实际浅组织中筋膜界面又相距很近，在界面处反射波就会不同程度的相互叠加，造成界面波形的不易识别。本文应用小波变换具有恒Q（品质因数）性质、对信号的时宽和带宽的局部分析能力，对浅表信号在不同频率范围内进行多分辨分析，实现了筋膜界面回波的时域信号重构。从重构界面信号能更清晰看出应用振荡波数作为界面特征量的可行性，证实卷积模型解释信号叠加的正确性。本文中的信号处理方法是第一次应用于浅表软组织结构的超声检测课题，研究工作还未见相关文献报道。实验卷积模型及其对振荡波数的解释可用于人体软组织间界面或其它相似界面结构特征的识别。离体猪肉脂肪厚度实验证明仪器能较好的测量较厚且组织结构均匀的脂肪厚度，而对于人体腹部实验，由于呼吸运动，超声回波信号伴随着噪音干扰，这使得浅表软组织的回波信号难以检测。如何消除测量干扰，提高信号的稳定性，是测量仪器设计无论从软件还是硬件都是进一步深入研究的课题。
[Abstract]:Ultrasound technology has developed rapidly in recent years and is widely used in medical diagnosis, treatment and industrial detection. Ultrasound after interaction with biological tissue is an information carrier reflecting the nature of biological tissue. Some structural parameters of tissues have always been one of the hotspots of research. In practice, ultrasonic diagnostic techniques, such as ultrasonic diagnostic apparatus, the ultrasonic signals received by the probe from the body surface are the results of superficial soft tissues (skin, fat, muscle, etc.), the interaction between the tissues in the body and the ultrasonic system, which is the structure of the body. On the other hand, the signals obtained by ultrasonic diagnostic instruments are mostly filtered signals, which lose a lot of information about the nature and structure of tissues.
In view of the fact that the ultrasonic detection of superficial soft tissue structure is not thoroughly studied, which makes the identification and localization of superficial interface structure inaccurate, the superficial layer structure of skin-superficial fat-superficial fascia-deep fat-deep fascia-muscle is proposed in this paper. The convolution model is used to analyze the signal characteristics of superficial interface. There are two aspects. On the one hand, the ultrasonic detection of superficial soft tissue layered structure, using signal processing technology to explore the interface signal feature extraction and recognition; on the other hand, the research topic attempts to use ultrasonic detection technology to develop a fat thickness measurement instrument, measuring the thickness of human local subcutaneous fat, so as to detect movement at any time. And dieting results are more likely to provide a monitoring tool for ultrasound liposuction.
The main tasks, achievements and innovations of this project are summarized as follows:
The biological tissue characteristics and structure distribution characteristics of superficial tissues (skin, fat, fascia and muscle) were analyzed, and the superficial tissue interfaces represented by the reflected echoes were found out by combining with the radio frequency echo signals detected by superficial ultrasound. By contrast, the echo of the tissue interface is determined, so as to prepare for further feature extraction of interface signals.
Although there is little difference in acoustic impedance between soft tissues, superficial soft tissues have their special structural characteristics, especially the thin interface between superficial fascia and deep fascia, and the number of interfaces varies from individual to individual, resulting in the obvious characteristics of ultrasonic echoes in these two places. Therefore, we can identify the characteristics of echo signals at these two places to determine the human body. Fat thickness.
In order to extract the signal characteristics of the fascial interface, the signal convolution model is used to explain the superposition effect of the interface signals in theory. Based on this, the convolution model is simulated and verified by Matlab software, and the convolution model is applied to identify the characteristics of the interface signal of the pig superficial tissue in vitro. The wave number is the feasibility of signal recognition.
Ultrasound radio frequency signal carries a lot of structural information of superficial soft tissue, but the waveform emitted by the probe has a certain duration and is affected by the detection system. In the actual superficial tissue, the fascial interface is very close, and the reflected waves will overlap at the interface to different degrees, which makes the interface waveform difficult to identify. The wavelet transform has the property of constant Q (quality factor) and the ability of local analysis of the time-width and bandwidth of the signal. The superficial signal is analyzed by multi-resolution in different frequency range. The time-domain signal reconstruction of the fascial interface echo is realized. It is proved that convolution model can explain the correctness of signal superposition.
Experimental convolution model and its interpretation of oscillatory wavenumber can be used to identify the structural features of human soft tissue interfaces or other similar interfaces. In vitro pork fat thickness tester It is difficult to detect the echo signal of superficial soft tissue because of the noise disturbance accompanied by breathing movement in abdominal experiment. How to eliminate the measurement interference and improve the stability of the signal is the design of the measuring instrument whether from software or hardware. These are all topics to be further studied.
【学位授予单位】：重庆医科大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R310

【参考文献】