Search academic texts

Information × Registration Number 0222U005341, 0121U113956 , R & D reports Title Machine learning based processing of multichannel remote sensing data popup.stage_title Head Uss Mykhailo L., к.т.н. Registration Date 27-12-2022 Organization National Aerospace University "Kharkiv Aviation Institute" popup.description2 Scientific research report: 83 pages, 22 figures, 11 tables, 141 sources. The object of research is multi-channel images formed and processed by remote sensing systems. The purpose of the research is to develop methods based on machine learning for effective automatic and semi-automatic processing of multi-channel remote sensing data. The report presents the results of the development of combined quality metrics for three-component images of multi-channel remote sensing data using several elementary metrics fed to the input of a trained neural network. The analysis proved that this significantly improved the Spearman's rank correlation between the metrics and the average human rating of quality. A method and tool for predicting the efficiency of lossy image compression has also been developed. The prediction is based on the training of a neural network whose inputs are a parameter that controls the compression and several parameters that describe the complexity of the image. It is shown that it is possible to predict the presence of an optimal operating point when compressing images distorted by noise, and the probability of correct classification of the image. Product Description popup.authors Ieremiev Oleg I Vasilyeva Irina K. Kovalenko Bogdan V Lukin Volodymyr Vasylovych Makarichev Viktor O Chernova Galina A popup.nrat_date 2022-12-27 Close