High-resolution remote sensing, deep learning and geomorphometry in analyses of mass movements and geological risk

• Grant number: 19/26568-0
• Support Opportunities: Regular Research Grants
• Start date: February 01, 2021
• End date: July 31, 2023
• Field of knowledge: Physical Sciences and Mathematics - Geosciences
• Principal Investigator: Carlos Henrique Grohmann de Carvalho
• Grantee: Carlos Henrique Grohmann de Carvalho
• Host Institution: Instituto de Energia e Ambiente (IEE). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated researchers: Daniel Hölbling ; Francisco Manoel Wohnrath Tognoli ; John Lindsay ; José Alberto Quintanilha ; Marcelo Fischer Gramani

Abstract

The development of remote sensing technologies in the last decade has lead to an exponential growth of available information about the Earth's surface. Among such advances, one can mention high-resolution orbital imagery with stereoscopic geometry allowing the generation of Digital Elevation Models (DEMs), airborne or terrestrial LiDAR (Light Detection And Ranging), and, more recently, applications of Structure from Motion--Multi View Stereo (SfM-MVS) to images acquired by Remotely Piloted Aircrafts (RPAs). In this research project, we propose the application of modern geospatial tools -- high-resolution Remote Sensing based on RPAs, terrestrial and RPA-borne LiDAR, SfM-MVS, Deep Learning and cloud computing, -- in the analysis of mass movements (landslides) and geological risk. The subjects selected for study include: multi-sensor (SfM-MVS, LiDAR-RPA) and temporal mapping of a landslide area active for over 20 years in the coastal city of São Sebastião (São Paulo State); comparative of airborne LiDAR and SfM-MVS 3D models from two distinct dates in two urban low-income areas identified as highly susceptible to mass movements (in collaboration with the Civil Defense of São Paulo); image classification and segmentation by GEOBIA and deep learning towards semi-automatic creation of landslides inventories. The project, with a two-year schedule, will be conducted by the principal researcher and his students, in collaboration with professors/researchers/students from USP, Unisinos, Salzburg University (Austria), and Guelph University (Canada). (AU)

Scientific publications

(8)

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

DIAS, HELEN CRISTINA; HOELBLING, DANIEL; GROHMANN, CARLOS HENRIQUE. Landslide Susceptibility Mapping in Brazil: A Review. GEOSCIENCES, v. 11, n. 10, OCT 2021. (19/17261-8, 19/26568-0)

GARCIA, G. P. B.; SOARES, L. P.; ESPADOTO, M.; GROHMANN, C. H.. Relict landslide detection using deep-learning architectures for image segmentation in rainforest areas: a new framework. International Journal of Remote Sensing, v. 44, n. 7, p. 28-pg., 2023-04-03. (16/06628-0, 19/26568-0)

GROHMANN, CARLOS H.; VIANA, CAMILA D.; GARCIA, GUILHERME P. B.; ALBUQUERQUE, RAFAEL W.. Remotely piloted aircraft -based automated vertical surface survey. METHODSX, v. 10, p. 6-pg., 2023-01-01. (16/06628-0, 19/26568-0)

DE SOUSA, AMANDA MENDES; VIANA, CAMILA DUELIS; GARCIA, GUILHERME PEREIRA BENTO; GROHMANN, CARLOS HENRIQUE. Monitoring Geological Risk Areas in the City of Sao Paulo Based on Multi-Temporal High-Resolution 3D Models. REMOTE SENSING, v. 15, n. 12, p. 19-pg., 2023-06-01. (19/26568-0)

HELEN CRISTINA DIAS; LUCAS HENRIQUE SANDRE; DIEGO ALEJANDRO SATIZÁBAL ALARCÓN; CARLOS HENRIQUE GROHMANN; JOSÉ ALBERTO QUINTANILHA. Landslide recognition using SVM, Random Forest, and Maximum Likelihood classifiers on high-resolution satellite images: A case study of Itaóca, southeastern Brazil. BRAZILIAN JOURNAL OF GEOLOGY, v. 51, n. 4, 2021-12-17. (19/17261-8, 19/26568-0)

DIAS, HELEN CRISTINA; SANDRE, LUCAS HENRIQUE; SATIZABAL ALARCON, DIEGO ALEJANDRO; GROHMANN, CARLOS HENRIQUE; QUINTANILHA, JOSE ALBERTO. andslide recognition using SVM, Random Forest, and Maximum Likelihood classifiers on high-resolution satellite images: A case study of Itaoca, southeastern Brazi. BRAZILIAN JOURNAL OF GEOLOGY, v. 51, n. 4, 2021. (19/26568-0, 19/17261-8)

GUILHERME PEREIRA BENTO GARCIA; CARLOS HENRIQUE GROHMANN; CAMILA DUELIS VIANA; ELTON BARBOSA GOMES. Using terrestrial laser scanner and RPA-based-photogrammetry for surface analysis of a landslide: a comparison. Bol. Ciênc. Geod., v. 28, n. 3, 2022-10-10. (19/26568-0, 16/06628-0)

SOARES, LUCAS PEDROSA; DIAS, HELEN CRISTINA; BENTO GARCIA, GUILHERME PEREIRA; GROHMANN, CARLOS HENRIQUE. Landslide Segmentation with Deep Learning: Evaluating Model Generalization in Rainfall-Induced Landslides in Brazil. REMOTE SENSING, v. 14, n. 9, p. 17-pg., 2022-05-01. (19/26568-0, 16/06628-0)