Scientific journal
Scientific Review. Technical science
ISSN 2500-0799
ПИ №ФС77-57440

Personal portfolio

DATA MINING IN PROBLEMS OF TRASNPORT INFRASTRUCTURE DEVELOPMENT FORECASTING

Fedoseev A.A. 1 Mikheev S.V. 1 Golovnin O.K. 1
1 Samara State Aerospace University
The analysis of Data Mining technology utilization has been done for showing up of undetected rules in problems of transport infrastructure detection. The results of Earth remote sensing (by spacecrafts or by aircrafts) was offered as initial data (transport stream intensity data and stream intensity static features condition data. Furthermore the initial data for static and dynamic features estimation are hyperspectral images of transport infrastructure features. Hyperspectral imagery is a composition of spatial and spectral data which gives a possibility to analyze of spectral characteristics (signatures). The possibility of classification, which provide of imagery segmentation, has been shown. The «soft» classification algorithms with «fuzzy» logic utilization implementing has been shown to meet the complexity conducted with specificity of processing and remote sensing process.