Latest Headlines / Surveying Basics / Remote Sensing / Government

An Automated Solution for Vegetation Mapping

August 29, 2014
/ Print / Reprints /
/ Text Size+

With its 27 member states, the European Union (EU) has some 12 million farmers who strive to reliably produce crops to feed about 500 million consumers, while simultaneously protecting the environment.

Despite adhering to established environmental standards, European farmers are significantly challenged to produce high yields within densely populated regions and equally protect the environment.

As a result, the EU operates the Common Agricultural Policy (CAP), which provides agricultural subsidies and other financial programs to both support farmers and the landscapes on which they work. Launched in 1962, the CAP has undergone many reforms to better ensure farmers can yield a living, to protect the environment and to ensure Europeans have a reliable, affordable food supply.

As part of this changing environment, the EU adopted an Integrated Administration and Control System (IACS) to improve the efficiency and accuracy of aid applications for direct CAP payments to farmers. These direct payments –– estimated to reach more than EUR 277 billion in 2014 –– are targeted rewards for both crop production and meeting “cross-compliance” rules, which stipulate efficient agricultural practices aimed at preserving biodiversity, soil quality and the environment in general. 

To comply with IACS, member states need to create a geodatabase that uniquely identifies each farmer, their individual parcels of agricultural holdings and aid applications specific to each parcel. A critical element of IACS is that States develop a Land Parcel Information System (LPIS) to accurately map their agricultural land at a very high resolution, as well as classify all vegetative features on each parcel by type and height. Without an LPIS, states cannot apply for CAP aid payments; and without a way to effectively archive and update the information to ensure claims can be validated, farmers and states risk financial penalties.

With deadlines looming, and significant subsidies on offer, Germany’s RLP AgroScience saw the opportunity to use advanced spatial technology to automate this monumental task in order to help its local authorities meet the EU’s requirements. Using available aerial imagery, digital surface models and image analysis software, a small RLP AgroScience team, together with local authorities, has created an operational system that completely automates the process of mapping and classifying vegetation, to quickly produce precise, standardized classification datasets––the root layer of the vegetative features in the LPIS.

The first of its kind in Germany, RLP AgroScience has not only proven that large-scale, automated and repeatable landscape-feature classification is possible, it has the operational seeds to possibly grow this system beyond its regional borders.


The Challenge

State authorities need to map their landscapes well enough that they can prove–– from their computer screen––that any farmer’s aid claim is accurate. This requires that every bush and tree on the ground has its geospatial counterpart in the LPIS. 

For RLP AgroScience that meant inventorying and classifying individual vegetative features across 19,000 square kilometers (7,336 square miles). It estimated it would need 15 full-time staff and a full year to manually digitize that volume of vegetation, a timeline that would jeopardize meeting the application deadline.

RLP AgroScience needed an intelligent, flexible and efficient image analysis tool that could objectively and automatically identify and classify vegetation. And since the claims application deadline is yearly, the solution for the landscape-feature classification needed to offer repeatable and adaptable workflows that could quickly integrate new data, run new classifications and allow for any unexpected CAP-compliance rules issued by the EU.

To fully meet the EU classification requirements, the system also needed to produce vector datasets that could seamlessly integrate with the OGC- and Inspire-compliant geodata infrastructure in Rhineland-Palatinate. 


The Solution

Having identified that simplicity, reliability and flexibility would be the three critical elements needed to develop their automated vegetation-mapping solution, RLP AgroScience chose Trimble’s eCognition technology to provide them with the image-analysis tools to identify, delineate and classify landscape features as well as the adaptable framework to integrate regular data updates and deliver standardized results.

Aptly titled “ALEK,” (Automatic Landscape Feature Classification), RLP AgroScience’s automated classification system combines customized eCognition and ESRI workflows to classify and map the entire region. Using existing 20-cm-resolution, orthorectified aerial images  and digital surface models, eCognition methodically and automatically analyzes the imagery to identify and separate vegetation from non-vegetation. Based on physical properties and pre-defined, region-specific rules, it then determines each vegetation type such as trees or hedgerows. And finally, it delineates each vegetative object and produces georeferenced vector datasets of all classified vegetation. Those vector classifications are then ingested into ESRI ArcGIS to create EU-compliant data for the local ministry of agriculture’s LPIS.

With the ALEK system, RLP AgroScience was able to precisely classify and map the entire 19,000-square-kilometer Rhineland-Palatinate region in three months, significantly reducing the time, resources and costs that would be needed to manually produce the required datasets.

“Manual digitization is not only incredibly tedious, it’s subjective––15 people can interpret the same object 15 different ways––and prone to error,” said Dr. Matthias Trapp, RLP AgroScience’s head of environmental systems. “With eCognition’s objective image analysis, we create standardized, reproducible results in a fraction of the time. Its speed, accuracy and data flexibility allowed a small team of image analysts to develop a fully automated, repeatable large-scale vegetation mapping system at no additional data cost to the ministry.”


The Benefits

By transforming months of manual classification work into an automated exercise in keystrokes––without taxing human or financial resources, or sacrificing quality––RLP AgroScience can now enable the local authorities to build their landscape feature layers of the EU-required LPIS, verify farmers’ claims and submit accurate applications for CAP aid on time. And with ALEK’s repeatable platform, it can ensure it can continually and reliably classify the changing landscape of Rhineland-Palatinate.

Since German law dictates that Ordnance Surveys must acquire aerial images of their respective states at regular intervals, RLP AgroScience can not only assure it will routinely update the ALEK system, it can continue to efficiently and accurately create land classifications at little to no cost. The automation, combined with regularly updated standard imagery, makes the system viable for other applications such as change detection and environmental monitoring.

With the batch processing features of eCognition technology, the ALEK system can handle significant volumes of data as well as automatically repeat the classification workflows each time a new dataset is introduced. In addition, should the EU issue new IACS data specifications, RLP AgroScience can adjust the system to meet those requirements with one simple change in the eCognition workflow. 

Without the simple, yet powerful developing environment of eCognition, RLP AgroScience would have had to more than triple the size of its team to produce the same amount of output––at potentially lesser quality. The increased productivity and improved efficiency enables the organization to continue to innovate and expand ALEK applications.

Anchored by eCognition’s automation and its building-block nature of repeatable workflows, ALEK is helping to ensure RLP AgroScience can continue to efficiently and reliably deliver a crop of classifications for Rhineland-Palatinate, and possibly for other EU regions as well.



A non-profit institute owned by the Federal State of Rhineland-Palatinate, RLP AgroScience has been undertaking applied research in the areas of agriculture and the environment for more than 20 years. It is based in Neustadt an der Weinstrasse, Germany.

Did you enjoy this article? Click here to subscribe to POB

Recent Articles by Mary Wagner

You must login or register in order to post a comment.



Image Galleries

HxGN Live

More than 3,500 attendees from more than 70 countries attended HxGN Live, the annual Hexagon AB user conference, at the MGM Grand Hotel & Casino in Las Vegas on June 3-6. About 450 keynotes and panel discussions were held, and several companies from around the world exhibited their geospatial products. Here are a few snapshots from the event.


POB January 2015 Cover

2015 January

In this January 2015 issue of POB, we take a look at integrated technologies that provide real-time 3D measurements of a Brisbane construction site.

Table Of Contents Subscribe

Unmanned Aerial Systems (UAS)

With the unmanned aerial system (UAS) presence growing in surveying, you will in 2015:
View Results Poll Archive

Point of Beginning Store

M:\General Shared\__AEC Store Katie Z\AEC Store\Images\POB\epubsite\Statues-pic-large.gif
Surveyor Statues

The perfect gift or award for any special occasion.

More Products

Clear Seas Research

Clear Seas ResearchWith access to over one million professionals and more than 60 industry-specific publications, Clear Seas Research offers relevant insights from those who know your industry best. Let us customize a market research solution that exceeds your marketing goals.


Facebook logo Twitter logo  LinkedIn logo  YouTube logoRPLS small logo

Google +

Geo Locator

Buyers Guide

The #1 buyers' guide for land surveyors and geomatics professionals. Search listings for software and equipment manufacturers, equipment dealers and professional services. CLICK HERE to view GeoLocator.