BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//talks.staging.osgeo.org//foss4g-europe-2025//speaker//U3
 HPFY
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-foss4g-europe-2025-EU3CB3@talks.staging.osgeo.org
DTSTART;TZID=CET:20250717T140000
DTEND;TZID=CET:20250717T143000
DESCRIPTION:Introduction and Research Objective\nUnder the Renewable Energy
  Sources Act (EEG 2023)\, Germany's installed capacity of solar photovolta
 ic (PV) systems is projected to increase from 81.7 GW in 2023 to 400 GW by
  2040\, making them a key pillar of the country's energy transition. This 
 rapid expansion of ground-mounted PV systems necessitates monitoring tools
  to assess their environmental impact and ensure compliance with regulator
 y frameworks. One such framework is Section 37 of EEG 2023\, which mandate
 s e.g. that solar modules must not cover more than 60% of the total solar 
 park.\nHowever\, there is currently a lack of precise spatial data on PV i
 nstallations across Germany\, which poses major challenges for those invol
 ved in quantitatively assessing the conflicting goals of nature conservati
 on and energy use. This missing data limits the ability to evaluate compli
 ance with regulations and assess the effectiveness of conservation measure
 s implemented alongside renewable energy development.\nThe Marktstammdaten
 register (MaStR)\, an open registry managed by the German Federal Network 
 Agency\, provides point-based location data for PV systems but lacks essen
 tial spatial details\, such as 1) proportion of the total area covered by 
 solar modules\, 2) distance between module rows and 3) orientation in degr
 ees to which the solar modules are aligned. These spatial parameters are c
 rucial for understanding the ecological and regulatory impacts of PV syste
 ms\, such as their effects on biodiversity and compliance with ecological 
 guidelines. In this study\, we aim to derive information from orthophotos 
 about the listed parameters for all ground-mounted PV systems in Germany. 
 Specifically\, we employ the Segment Anything Model (SAM) (Kirillov et-al.
 \, 2023)\, a state-of-the-art zero-shot segmentation model\, in combinatio
 n with digital orthophotos (DOP20) with a ground resolution of 20 cm per p
 ixel. SAM enables precise segmentation of objects or regions in images\, a
 llowing us to identify and delineate the components of PV plants with high
  accuracy.\nData and Methodology\nFor this study\, we utilized an open-acc
 ess dataset by Manske et al. (2022) available at Zenodo. This dataset cont
 ains manually digitized areas of 7\,839 ground-mounted PV plants across Ge
 rmany\, serving as a reference to identify and locate corresponding DOP20.
  The DOP20\, featuring red\, green\, blue (RGB)\, and near-infrared bands\
 , are made publicly available under Germany’s Second Open Data Act (effe
 ctive July 2021)\, offering the necessary spatial resolution for detailed 
 mapping. The temporal resolution is dependent on the federal state and var
 ies between 1 and 3 years.\nTo prepare the DOP20 images for use with SAM\,
  we cropped the images intersecting with PV plants into 640x640 pixel imag
 e chips with an overlap of 340 pixels. This process resulted in the creati
 on of a dataset comprising over 350\,000 image chips\, which formed a grid
  for segmentation. The SAM segmentation process was conducted using only t
 he RGB bands of the orthophotos. After generating segmentation masks\, we 
 extracted the pixel values for all available bands and derived spectral in
 dices\, including the Normalized Difference Vegetation Index\, Photovoltai
 c Index\, and Normalized Impervious Surface Index. The pixel values for ea
 ch mask were aggregated into median values for each segment. Subsequently\
 , we conducted iterative unsupervised clustering using the DBSCAN algorith
 m. The clustering process comprised two main steps:\n    1. clustering bas
 ed on spectral indices and NIR bands to filter out vegetation\, shadows\, 
 and other non-PV objects as outliers. \n    2. geometric property clusteri
 ng\, including segment size\, rectangularity\, orientation\, and percent a
 rea difference to their oriented bounding box\, to remove additional outli
 ers (e.g.\, irregular shapes overlapping with pathways or transformer stat
 ions). \nIn the final post-processing step\, overlapping segments (due to 
 overlapping images) were merged and\, where sufficient rectangularity (≥
 0.8) was achieved\, rectangular bounding boxes were generated. Only segmen
 ts classified as PV module rows were retained for further analysis.\nOne o
 f the primary challenges in this pipeline was the unsupervised filtering o
 f PV modules\, as the DOP20 images were captured under varying seasonal\, 
 solar\, and viewing angle conditions. Fixed clustering rules were unsuitab
 le due to the diverse spectral properties of PV modules\, making unsupervi
 sed clustering the only viable approach to classify PV rows.\nResults\nPre
 liminary results from this ongoing work demonstrate that it is highly feas
 ible to extract high-quality PV module rows from DOP20 images using SAM. O
 nly about 5% of the results exhibited unsatisfactory segmentation\, where 
 module rows and inter-row spaces were not properly separated and grouped i
 nto the same segment.\nThe details of the PV plants analyzed to date are a
 s follows:\n    • The proportion of the total area covered by solar modu
 les is approximately 65%. However\, these installations were constructed b
 efore the EEG 2023 regulation came into effect on January 1\, 2023\, and a
 re thus exempt from the new coverage limit. \n    • The distance between
  module rows averaged 3 meters\, with considerably larger spacing observed
  in PV sun-tracking systems. \n    • The orientation of solar modules pr
 edominantly faced to the south (180° azimuth). Approximately 80% of modul
 e rows deviated by up to 20° east (azimuth -20°) or 20° west (azimuth +
 20°). \nSince the analysis is still ongoing\, additional results will be 
 added in the presentation.\nDiscussion and Outlook\nThe vector products ge
 nerated through our pipeline have significant implications for both policy
 -making and research. Derived plant details will be made freely available 
 under an open licence\, enabling further studies on the environmental impa
 cts of PV systems and supporting decision-making by regulatory bodies.\nBy
  applying this segmentation pipeline to newly recorded DOP20 datasets\, ty
 pically updated every two years across Germany\, our approach offers a sca
 lable solution for long-term monitoring of PV system dynamics. This capabi
 lity is critical given the rapid expansion of PV systems and the potential
  conflicts between renewable energy development and nature conservation.\n
 Our methodology aligns with the goals of the FOSS4G Europe Academic Track\
 , emphasizing the use of open data\, open-source tools\, reproducible work
 flows and therefore ensures full transparency. The complete workflow will 
 be made publicly available under an appropriate open-source license to fos
 ter collaboration and innovation within the geospatial and renewable energ
 y communities.
DTSTAMP:20260527T125422Z
LOCATION:PA01 (Quarticle)
SUMMARY:Spatial Parameter Analysis of Ground-Mounted Photovoltaic Systems U
 tilizing Orthophotos and the Segment Anything Model - Philipp Gärtner
URL:https://talks.staging.osgeo.org/foss4g-europe-2025/talk/EU3CB3/
END:VEVENT
END:VCALENDAR