Machine learning methods for improving the quality of images from cctv cameras on railway transport
Issaikin D. Zhamangarin D. Akhmetov B. Lakhno V. Omarova B. Omarova G. Mailybaev E.
31 May 2021Little Lion Scientific
Journal of Theoretical and Applied Information Technology
2021#99Issue 102267 - 2279 pp.
The article has developed a methodology for changing the resolution (RS) of images obtained from CCTV cameras on railway transport. The research was carried out on the basis of the application of machine learning methods (MLM). Due to the implementation of this approach, it was possible to expand the functionality of the MLM. In particular, it was proposed to carry out the resampling process with the target frame information factor of the image. This coefficient is applicable for both increasing and decreasing of RS. This should provide a high quality of resampling and, at the same time, reduce the training time for neural-like structures (NLS). There was developed a method of changing the RS using the NLS. This contributes to the high efficiency of resampling of the images obtained from CCTV cameras, according to the criterion based on PSNR. The proposed solutions are characterized by a reduction in the size of the computing resources that are required for such a procedure. The proposed solutions are characterized by a reduction in the size of the computing resources that are required for such a procedure.
Image Quality Improvement , Neural Networks , Railway Transport , Video Surveillance Systems
Text of the article Перейти на текст статьи
Kazakh university ways of communications, Kazakhstan
Abai Kazakh National Pedagogical University, Kazakhstan
National University of Life and Environmental Sciences of Ukraine, Ukraine
Kazakh university ways of communications
Abai Kazakh National Pedagogical University
National University of Life and Environmental Sciences of Ukraine
10 лет помогаем публиковать статьи Международный издатель
Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026