基于盲去卷积算法的太赫兹时域光谱图像处理方法

Terahertz time-domain spectral image processing method based on blind deconvolution algorithm

为了减少太赫兹波束衍射对成像造成的影响,以及提高太赫兹时域光谱的空间分辨率,采用了一种基于盲去卷积算法的图像处理方法。通过建立物理模型,利用图像灰度范围调整、盲去卷积、频率域滤波去噪等方法对太赫兹时域光谱进行处理,使用不同的指标比较处理前后所得到的图像。结果表明:该处理过程能够有效地提高太赫兹时域光谱的空间分辨率,对太赫兹成像分辨率的提高具有良好的效果。

In order to reduce the influence of terahertz beam diffraction on imaging and improve the spatial resolution of terahertz time-domain spectra,an image processing method based on blind deconvolution algorithm is adopted.By establishing aphysical model,terahertz time-domain spectra are processed by means of image gray scale adjustment,blind deconvolution,frequency domain filtering and denoising.Finally,different evaluation indexes are used to compare the images before and after processing.The results show that the processing process can effectively improve the spatial resolution of terahertz time-domain spectrum,and has a good effect on the improvement of terahertz imaging resolution.