时域光学相干层析(OCT)系统通常采用短时傅里叶变换(STFT)完成干涉信号的解调和图像重构。短时傅里叶变换算法简单, 但是在干涉信号解调时难以获得好的去噪效果, 通常还需在二维(2D)图像域对重构图像进行去噪。该方法数据运算量大, 集成度不高。将一维(1D)小波变换(WT)应用于时域光学相干层析成像技术, 同时实现干涉信号解调、去噪和图像重构。算法将时域光学相干层析的干涉信号分解到各个不同的频率空间, 保留包含调制频率的频率空间的小波系数, 对保留的小波系数进行滤波去噪后进行逆变换即可实现对干涉信号的解调和去噪, 对解调信号等间距采样实现图像重构。该方法数据运算量小, 集成度高, 结合先进的小波去噪技术可以大大提高重构图像的分辨率, 具有良好的应用前景。

Short time Fourier transform (STFT) is usually used to demodulate the interferometric signal and image reconstruction for time domain optical coherence tomography (OCT). STFT algorithm is simple, but cannot obtain good de-noising performance when the interferometric signal is demodulated. Therefore an additional de-noising in two-dimensional (2D) image domain is needed for reconstructed image. This method is time-consuming and incompact. In this paper, one-dimensional (1D) wavelet transformation (WT) was applied to demodulate signal and image reconstruction, furthermore, de-noise in wavelet domain could be completed at the same time. The interferometric signal was decomposed into different frequency ranges by WT and reconstructed only with these filtered coefficients whose frequency range covers the modulating frequency. After demodulating, tomographic image of sample could be reconstructed. Comparing with STFT followed by image de-noising, this method is more compact and efficient. Combining advanced de-noising technique in wavelet domain, this method has great potential in application.