BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//talks.staging.osgeo.org//foss4g-2024//talk//U7DXAQ
BEGIN:VTIMEZONE
TZID:-03
BEGIN:STANDARD
DTSTART:20000101T000000
RRULE:FREQ=YEARLY;BYMONTH=1
TZNAME:-03
TZOFFSETFROM:-0300
TZOFFSETTO:-0300
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-foss4g-2024-U7DXAQ@talks.staging.osgeo.org
DTSTART;TZID=-03:20241206T120000
DTEND;TZID=-03:20241206T123000
DESCRIPTION:Extreme precipitation events lead to rapid surface runoff\, cau
 sing sheet erosion and the formation of rills\, increasing the risk of fla
 sh floods. This combination of processes pose a threat to agricultural and
  rural areas and the sediment-laden water can infect even the urban zones 
 or cause damage to infrastructure. Detecting and predicting the formation 
 of erosive rills on agricultural land is\, therefore\, crucial for effecti
 ve land management and disaster prevention in rural areas.  \n\n  \n\nThe 
 contribution presents a research on the utilisation of convolutional neura
 l networks (CNN) to detect enhanced erosion using remote sensing data comb
 ined with the SMODERP hydrological and erosion model.   \n\n  \n\nWhile mo
 st tools for semantic segmentation (such as random forests) work only with
  single-pixel values\, CNNs consider also the relationship with its surrou
 ndings and between the bands. As the erosion rill patterns are visible esp
 ecially when compared to the surrounding soil\, it is a valuable feature f
 or their detection.   \n\n  \n\nHowever\, if we also have the digital elev
 ation model\, we can use geospatial tools and algorithms to enhance the im
 agery input to the neural networks with knowledge-based indices. In this c
 ase\, it is the SMODERP model.  \n\n  \n\nSMODERP is a hydrological model 
 designed to simulate surface runoff and erosion processes. It considers va
 rious factors such as soil type\, land cover\, slope\, and rainfall intens
 ity to predict the movement of water and sediments throughout the landscap
 e. The model calculates the critical height of sheet runoff as a rill form
 ation threshold\, which is essential to understand where erosion is likely
  to occur. The SMODERP is developed as a GIS tool\, available through GRAS
 S GIS and QGIS. More details about model on smoderp.fsv.cvut.cz or on GitH
 ub.  \n\n  \n\nThe methodology begins with data collection and preparation
 \, utilising high-resolution orthophoto aerial images of spatial resolutio
 ns of only a few centimetres. Additionally\, hydrological data from the SM
 ODERP model are incorporated to the CNN's input to capture erosion dynamic
 s. The talk will discuss the effect of the SMODERP's output inclusion on t
 he CNN's accuracy in rill detection.
DTSTAMP:20260428T230058Z
LOCATION:Room II
SUMMARY:Convolutional Neural Network-Based Detection of Erosion Rills on Ae
 rial Imagery Combined with Hydrological Model SMODERP Outputs - Ondřej Pe
 šek\, Petr Kavka
URL:https://talks.staging.osgeo.org/foss4g-2024/talk/U7DXAQ/
END:VEVENT
END:VCALENDAR