|
This article is cited in 7 scientific papers (total in 7 papers)
IMAGE PROCESSING, PATTERN RECOGNITION
Technology of enhancing image detalization with nonlinear correction of highly gradient fragments
V. A. Fursovab, E. V. Goshinba, K. S. Medvedevaa a Samara National Research University, Moskovskoye Shosse 34, 443086, Samara, Russia
b IPSI RAS – Branch of the FSRC “Crystallography and Photonics” RAS,
Molodogvardeyskaya 151, 443001, Samara, Russia
Abstract:
The article is devoted to the problem of improving the quality of images recorded using low-resolution optical instruments, including diffraction-based cameras. A two-stage image correction technology is proposed. At the first stage, the correction is carried out using a linear FIR filter with a centrally symmetric frequency response in the form of quadratic and exponential functions. The resulting image is then processed with a non-linear filter that performs computer retouching of image areas characterized by a noticeable brightness difference. This procedure is only performed on those pixels in which the absolute value of gradients in different directions is sufficiently high, that is, they are located on the borders of areas with different intensity levels. This allows us to avoid noise amplification in the background, which is typical of traditional filters. The examples of the implementation are provided, showing the possibility of achieving high sharpness and illustrating how the filter can be adjusted by visual perception.
Keywords:
image processing, FIR filter, nonlinear filter, centrally symmetric frequency response, blind identification.
Received: 29.04.2019 Accepted: 03.06.2019
Citation:
V. A. Fursov, E. V. Goshin, K. S. Medvedeva, “Technology of enhancing image detalization with nonlinear correction of highly gradient fragments”, Computer Optics, 43:3 (2019), 484–491
Linking options:
https://www.mathnet.ru/eng/co668 https://www.mathnet.ru/eng/co/v43/i3/p484
|
Statistics & downloads: |
Abstract page: | 175 | Full-text PDF : | 91 | References: | 12 |
|