Russian Federation
Russian Federation
Machine learning methods for identification of territories near railway infrastructure facilities are considered. The main sources of initial data are presented, as well as algorithms for assessing the quality of the results for landscape elements automated identification using the forest vegetation example. Forest vegetation state identification is carried out based on spectral-brightness characteristics determined using multispectral aerospace imaging materials. Quality indicators for the results of survey materials automated processing and algorithms for calculating the considered indicators are proposed. Examples are given to assess the results quality of using machine learning methods.
multispectral remote sensing, spectral characteristics, machine learning methods, identification quality
1. Shovengerdt R. A. Distancionnoe zondirovanie. Metody i modeli obrabotki izobrazheniy / per. s angl. A. V. Kiryushina, A. I. Dem'yanikova, 3-e izd. M.: Tehnosfera, 2010. 560 s.
2. Koraboshev O. Z. Analiz i perspektivy primeneniya metodov mashinnogo obucheniya dlya chrezvychaynyh situaciy // Intellektual'nye tehnologii na transporte. 2024. № 1 (37). S. 12–17. DOI:https://doi.org/10.20295/2413-2527-2024-137-12-17
3. Rashka S. Python i mashinnoe obuchenie / per. s angl. A. V. Logunova. M.: DMK Press, 2017. 418 s.
4. Géron A. Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow. Second Edition. Sebastopol (CA): O’Reilly Media, 2019. 848 p.
5. Terebizh V. Yu. Vvedenie v statisticheskuyu teoriyu obratnyh zadach. M.: Fizmatlit, 2005. 376 s.
6. Mochalov V. F., Habarov R. S. Obrabotka materialov mul'tispektral'noy s'emki na osnove metodov mashinnogo obucheniya pri upravlenii sostoyaniem lesnogo massiva // Sbornik materialov IV Mezhdunarodnoy nauchnoy konferencii po problemam upravleniya v tehnicheskih sistemah (PUTS-2021) (Sankt-Peterburg, 21–23 sentyabrya 2021). SPb.: LETI, 2021. S. 269–272.
7. Maxwell A. E., Warner T. A., Guillén L. A. Accuracy Assessment in Convolutional Neural Network-Based Deep Learning Remote Sensing Studie. Part 1: Literature Review // Remote Sensing. 2021. Vol. 13, iss. 13. Art. 2450. 27 p. DOI:https://doi.org/10.3390/rs13132450
