UDC 004.932.2
UDC 528.88
CSCSTI 28.23
CSCSTI 47.49
This study is dedicated to addressing the challenges associated with enhancing the processing and analysis of open data related to digital terrain models, with a specific focus on urbanized areas. The research investigates contemporary technologies aimed at advancing the accuracy and quality of presented geospatial data. The study conducts a comprehensive review and analysis of existing global digital terrain models. Through the application of the proposed evaluation method, the ALOS dataset is selected as the primary data source. Based on this dataset, a meticulously constructed training set is developed to facilitate the experimentation of an innovative neural network approach. The research delves into the implementation of a hybrid neural network architecture, strategically composed of distinct modules for extracting coordinate information and subsequent processing through a deep model. This approach synergistically leverages the strengths of each module to more effectively account for the intricate features of urbanized landscapes, thereby yielding substantially enhanced precision in analysis outcomes. The practical implications of this research are profound and extend across various domains, notably urban planning and territorial management.
convolutional neural network (CNN), DEM, radar survey, SAGA GIS, data cleaning
1. Rozenberg IN. Geoinformatsionnaya sistema — instrument tsifrovoi transformatsii geodannykh [Geographic information system — tool for digital geodata transformation]. Systems and Means of Informatics. 2022;32(1): 46–54. (In Russian). DOIhttps://doi.org/10.14357/08696527220104.
2. Onkov IV. Otsenka tochnosti vysot SRTM dlya tselei ortotransformirovaniya kosmicheskikh snimkov vysokogo razresheniya [Estimation of the accuracy of SRTM heights for the purposes of orthotransformation of high-resolution satellite images]. Geomatics. 2011;3: 40–46. (In Russian).
3. Yifan Z, Wenhao Y. Comparison of DEM Super-Resolution Methods Based on Interpolation and Neural Networks. Sensors. 2022;22(3): 745. DOIhttps://doi.org/10.3390/s22030745/.
4. Karachevceva IP, Dubov CC, Andreev MV., et al. Otkrytye prostranstvennye dannye dlya issledovaniya territorii i tsifrovye servisy dostupa k nim [Open spatial data for the study of territories and digital access services to them]. Spacecrafts & Technologies. 2023;7:2(44):142–152. (In Russian). DOIhttps://doi.org/10.26732/j.st.2023.2.07.
5. Golitsyn GS. Osobennosti spektra rel'efa poverkhnosti Luny I planet [Features of the surface relief spectrum of the Moon and planets]. Solar System Research. 2021;55(1): 34–37. DOIhttps://doi.org/10.31857/S0320930X21010035.
6. Sergeev IS, Egorov IV, Glebova AB. Spektral'nyi analiz rel'efa dlya resheniya prognozno-poiskovykh zadach na primere riftovoi zony Sredinno-Atlanticheskogo khrebta [Spectral analysis of the relief for solving predictive search tasks using the example of the rift zone of the Mid-Atlantic ridge]. Geomorfologiya. 2020;4: 34–44. (In Russian). DOIhttps://doi.org/10.31857/S0435428120040094.
7. Wang L, Liu H. An efficient method for identifying and filling surface depressions in digital elevation models for hydrologic analysis and modelling. International Journal of Geographical Information Science. 2006;20: 193–213. DOIhttps://doi.org/10.1080/13658810500433453.
8. Chaohui T, Qingxin Z, Wenlin H, et al. PLANET: Improved Convolutional Neural Networks with Image Enhancement for Image Classification. Mathematical Problems in Engineering. 2020;2020: 1–10. DOIhttps://doi.org/10.1155/2020/1245924.
9. He K, Zhang X, Ren S, Sun J. Deep Residual Learning for Image Recognition. Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2015: 770–778. DOIhttps://doi.org/10.1109/CVPR.2016.90.
