盲反褶积
反褶积
计算机视觉
人工智能
计算机科学
图像复原
图像(数学)
图像处理
算法
作者
Jonnadula Narasimharao,P. P. Deepthi,Bolla Aditya,C. R. Shruthi Reddy,Avala Raji Reddy,Ganpat Joshi
出处
期刊:Lecture notes in networks and systems
日期:2024-01-01
卷期号:: 371-389
被引量:2
标识
DOI:10.1007/978-981-99-9704-6_35
摘要
Due to improper camera settings and atmospheric turbulence, captured scene images are usually degraded by blurring. Image restoration is the process of extracting a blurred image's original image. Image restoration is an articulate technique that facilitates the retrieval of original images in a one-step manner. First blind deconvolution is among the most significant restoration methods in which the blurring operator is not identified. Blind system deconvolution is a very complicated process that restores objects without knowing what they were like before they started to deteriorate PSF. The PSF is an impulse response of a point source. This paper introduces a novel approach of blind deconvolution for blurred color images. Here, first perform the degradation process, then use the Wiener filter to reduce the noise from the input blurred image using the degradation function. Then the restoration process is performed using blind deconvolution, involving the estimation of PSF kernels through a thresholding mechanism. Following this, deblurring can be carried out for different sizes of PSF images. The deblurred images are restored with some ringing effects. In conclusion, the Canny edge detector is utilized to reduce noise and achieve the optimal restored image. The cost of MSE and PSNR was calculated for each method and compared in the results.
科研通智能强力驱动
Strongly Powered by AbleSci AI