The Research On Continuous-wave Terahertz Image Denoising And Contrast Enhancement

Terahertz(THz) wave, electromagnetic radiation between far-infrared and millimeter wave, has low photon energy but strong penetrability. In recent years, due to its good characteristics, THz had attracted attention in areas such as biomedical science, non-destructive evaluation, security control and had great potential application prospect. Continuous wave terahertz(CW-THz) imaging is an important way of THz imaging, unlike the pulsed wave terahertz imaging, CW-THz imaging has a series of advantages, such as simple structure, low cost, high reliability and control accuracy. All the parts of imaging system had been in-depth research in recent years. As an important part of CW-THz imaging, image denoising and image contrast enhancement would be the key to practical and commercial application. High signal-to-noise ratio, edge protection, accord with human visual effect of the CW-THz images have the decisive significance to image segmentation and image recognition. However, due to the milliwatts level of CW-THz power, large terahertz spots, the pyroelectric detector or the pyroelectric cameras, or the two-dimensional scanning platform, would cause significant influence to the quality of CW-THz images. Large noise and low contrast, blurred in details are the common characteristics of most CW-THz images. It is hard to get ideal effect by using some common image processing methods to process CW-THz images.In this paper, for the purpose of denoising and contrast enhancement, the CW-THz images, as the main research object, would be processed by a few of image processing algorithms. Qualitative and quantitative methods were used to evaluate the processing results.This paper mainly includes the following contents:(1) We would first introduce the mechanism of CW-THz imaging, and we would focus on the factors that affect the quality of the images. The CW-THz images noise, contrast, background of the stripe would be analyzed, so as to determine to direction of image processing algorithm.(2) We will use bilateral filter to CW-THz image denoise. We can see the feasibility and advantages of the CW-THz image processing by using this method. Experimental results show the effectiveness of the proposed method.(3) In order to enhance the contrast of CW-THz image, a few typical contrast enhancement algorithms would be use to enhance the CW THz images. At the same time, an optimization of homomorphic filter would be proposed to enhance the CW-THz images. And results show that the improved method has a better effect.(4) An approach of combination of the discrete wavelet transform, Retinex algorithm, and a designed nonlinear function would be proposed and use to increase CW-THz image contrast.(5) We would combine the MFC and OpenCV to integrate the algorithm of this paper to software, so as to facilitate in the CW THz image processing.