基于宽带检测放疗X-光光声效应仿体实验

X-ray photoacoustic phantom experiment based on wideband detection

目的：搭建一个宽带检测的单振元光声信号采集系统采集仿体的光声信号,探讨不同介质X-光光声信号差异和人体切片体模光声信号的去噪方法。方法：带宽为0.10～1.00 MHz的单振元宽带换能器垂直于X-光照射方向在样本一侧采集超声信号,超声信号通过34 d B带宽为0.05～5.00 MHz的前置放大器放大后由数字采集卡（采样率=40.00 MHz）采集并存储到计算机中。首先通过光声信号在整体包络、信号强度、振荡趋势、上下包络所围成的面积等方面差异对4种仿体进行分析。另外对人体切片模体的X-光声信号在有限长单位冲激响应低通滤波的基础上进行小波平移不变量去噪与只对模体光声信号进行小波平移不变量去噪,比较这两种去噪方式的效果。结果：不同仿体光声信号有差异,这种差异与物质的成分和比例相关。对人体切片体模光声信号通过Db3小波基进行5层分解的小波平移不变量去噪,比有限长单位冲激响应低通滤波能更好地提高信号信噪比。结论：利用6 MV医用直线加速器的短脉冲X-光光声信号对介质的差异性及小波去噪有助于提高其对放疗的研究与应用价值。

Objective To discuss on the difference of X-ray photoacoustic signals in different media and different denoising methods for the photoacoustic signals of human body slice model by establishing a single element photoacoustic signal acquisition system based on wideband detection for capturing photoacoustic signals of different phantoms. Methods A single element wideband transducer with a bandwidth of 0.10-1.00 MHz was used to collect photoacoustic signals at the side of the sample perpendicular to the direction of X-ray beam. The signals were amplified by a preamplifier with a bandwidth of 0.05-5.00 MHz at 34 dB, and the amplified signals were acquired by a data acquisition board with a sampling rate of 40.00 MHz, and stored in a computer. Differences of photoacoustic signals among 4 kinds of phantoms in the envelope, signal intensity, oscillation tendency, and the area between the top and bottom envelopes were analyzed. Based on the finite impulse response （FIR） low pass filter, the X-ray photoacoustic signals of the human body slice model were treated by wavelet translation invariant denoising method. As an reference, the photoacoustic signals of model were treated by wavelet translation invariant denoising method alone. The denoising performances were compared between the two denoising methods. Results The photoacoustic signals of different phantoms were different, which were related to the composition and ratio of the substance. The wavelet translation invariant denoising of 5 layer decomposition of the Db3 wavelet base was performed on the photoacoustic signal of human body slice model, which improved the signal to noise ratio better than the FIR low pass filter. Conclusion The differences in different media of short pulse X-ray photoacoustic signals generated by 6 MV clinical linear accelerator and wavelet denoising can be used to improve the research and application value of X-ray photoacoustic signals in radiotherapy.