将线性回归理论应用到光声(PA)图像的分析过程中,构建了多波长组分浓度分布算法,并用计算机模拟和生物组织样品实验来证明该算法能够使PA图像反映更多的生物组织信息。实验中,同一个样品多次成像,只改变激发光波长,其他参数保持不变。实验结果表明,单一波长的PA图像反映的信息与混合物组成成分在该波长处吸收强弱相吻合,当混合物各个组分的吸收差异比较大时,单一波长PA图像不能充分反映混合物中每种组分的分布信息;在此基础上,对不同波长下获得的PA图像用组分浓度分布算法处理之后,图像中各个成分分布的信息明显改善,与实际样品吻合较好。
Linear regression theory is introduced to analyze the process of photoacoustic image and the algorithm of multi-wavelength component concentration distribution is constructed. Computer simulation and biological tissue sample experiment are carried out to prove that the algorithm can make photoacoustic image reflect more biological tissue information. During the experiment, all parameters are kept as constant except the wavelength of exciting pulsed laser. The result shows that the information reflected by photoacoustic image of single wavelength matches well with the absorption of mixture component at excited wavelength. Photoacoustic image of single length cannot fully indicate the distributions of each component when the component absorptions are quite different. Based on this, the distribution information of image components changes and matches well with real sample after being treated by the algorithm.