HedgeRows - Bavaria, 2019-2021

Hedgerows play an important role in maintaining biodiversity, carbon sequestration, soil stability and the ecological integrity of agricultural landscapes. In this dataset, hedgerows are mapped for the whole of Bavaria. Orthophotos with a spatial resolution of 20 cm, taken in the period from 2019 to 2021, were used in a deep learning approach. Hedgerow polygons of the Bavarian in-situ biotope mapping from 5 districts (Miltenberg, Hassberge, Dillingen a.d. Donau, Freyung-Grafenau, Weilheim-Schongau) as well as other manually digitized polygons were used for training and testing as input into a DeepLabV3 Convolutional Neural Network (CNN). The CNN has a Resnet50 backbone and was optimized with the Dice loss as a cost function. The generated hedgerow probability tiles were post-processed by merging and averaging the overlapping tile boundaries, shape simplification and filtering. For more details, see Huber Garcia et al. (2025). The dataset has been created within the project FPCUP (https://www.copernicus-user-uptake.eu/) in close cooperation with Bayerisches Landesamt für Umwelt (LfU).

Citation proposal
German Aerospace Center (DLR) - German Aerospace Center (DLR). HedgeRows - Bavaria, 2019-2021. https://gdk.gdi-de.org/geonetwork/srv/api/records/88438fae-9fe5-4f55-9f4d-1a73778bf62c

Simple

Date ( Creation ): 2025-07-01T09:00:00

Edition

Identifier: https://geoservice.dlr.de/catalogue/srv/metadata/88438fae-9fe5-4f55-9f4d-1a73778bf62c

Presentation form: Digital map

Other citation details

Purpose: The hedgerow dataset can form a data basis for informed decision making and for scientific applications in the field of sustainable agriculture and ecology.

Status: ongoing

Author

German Aerospace Center (DLR) -

Author

German Aerospace Center (DLR) -

Point of contact

German Aerospace Center (DLR)

Maintenance and update frequency: As needed

GEMET - INSPIRE themes, version 1.0 ( Theme )

Land cover

Spatial scope ( Place )

Regional

Keywords

DLR
EOC
opendata
Hedges
Hedgerows
Habitats
Linear Woody Vegetation
High-Diversity Landscape Features
Biotopes
Orthophotos
CNN
DeepLabV3

Use limitation: Nutzungseinschränkungen: Das DLR ist nicht haftbar für Schäden, die sich aus der Nutzung ergeben. / Use Limitations: DLR not liable for damage resulting from use.

Access constraints: Other restrictions

Use constraints: Other restrictions

Other constraints: Nutzungsbedingungen: Lizenz, https://creativecommons.org/licenses/by/4.0/ Terms of use: License, https://creativecommons.org/licenses/by/4.0/

Other constraints

{"id": "cc-by-4.0",

"name": "Creative-Commons - Attribution 4.0 International (CC BY 4.0)",

"url": "http://dcat-ap.de/def/licenses/cc-by/4.0",

"quelle": "Copyright DLR (year of production)"}

Spatial representation type: Grid

Metadata language: English English

Character set: UTF8

Topic category

Farming

Topology level: Full surface graph

Geometric object type: Surface

Geometric object count: 564339

Distribution format

Geopackage ()

OnLine resource: HEDGEROWS_BY_2019_2021
EOC Land Map Service

OnLine resource: HTTP Download
HTTP Download

OnLine resource: EOC Geoservice Dataset
EOC Geoservice Dataset

OnLine resource: Scientific publication
Scientific publication

Hierarchy level: Dataset

Domain consistency

Measure identification: INSPIRE / Conformity_001

Conformance result

Date ( Publication ): 2010-12-08

Explanation: See the referenced specification.

Pass: true

Statement: RGB orthophotos with a spatial resolution of 20 cm were used as earth observation data, which were taken between April 2019 and September 2021 by the Bavarian State Office for Digitization, Highspeed Internet and Surveying. Hedgerow data from the federal biotope mapping of the Bavarian Environmental Agency as well as manually digitalized hedgerow shapes are used for training, validation and testing.

Description

Processing: Orthophotos and hedgerow shapes were used in a CNN model to predict hedgerow objects. For details of preprocessing see Huber Garcia et al. (2025).

Quality Assurance: A comprehensive validation campaign has been performed to assess the accuracy of the dataset quantitatively base on 3732 image tiles in the districts Freyung-Grafenau, Miltenberg, Hassberge, Dillingen and Weilheim-Schongau. The accuracy assessment is described in detail in Huber Garcia et al. (2025)