10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

MULTILEVEL "FOREST FIREFIGHTING - MANAGEMENT SYSTEM" FOR AN OPTIMIZED OPERATIONAL GUIDANCE OF GROUND AND AIR FORCES IN FOREST FIRE EVENTS Alexander Almer1, Thomas Schnabel1, Roland Perko1, Johann Raggam1, Armin Köfler1, Michael Schmidt1, Harald Schlemmer1, Eral Türkyilmaz1, Richard Feischl2 1. JOANNEUM RESEARCH Forschungsges.m.b.H, Graz, AUSTRIA, {firstname.lastname}@joanneum.at 2. Ing. Richard Feischl, Gumpoldskirchen, AUSTRIA, [email protected] ABSTRACT The dramatic increase in forest fires in Europe has made improvements in forest fire control a major national and international issue in the quest to protect human lives and resources as well as to reduce the negative environmental impact. Forest firefighting operations are in general very dangerous and regarding the involvement of comprehensive human resources as well as extensive use of ground-, air vehicles and equipment the event management is a crucial factor to reduce risks in firefighting missions. This paper describes the development of a multi-functional airborne management support system within the frame of the Austrian national safety and security research programme KIRAS. The objective was to assist crisis management tasks of emergency teams and armed forces in disaster management by providing multi spectral, airborne image data products in near real-time and analysis processes. This article includes further concepts and results of ongoing research activities which focus on the development of an innovative system for optimal deployment of firefighting teams and the generation of situational maps for a common operational picture in near real-time using airborne image acquisition. Furthermore it includes an effective management solution to optimize communication and the command of operational human units and vehicles in the air and on the ground. The ongoing project involves the development of simulation-based modules for decision support and impact evaluation and an innovative development for the large-scale monitoring of affected forest areas once the fire has been extinguished.

Keywords: Airborne sensing, multi sensor imaging, line-of-sight communication, near real-time fire monitoring, simulation-based decision support, forest firefighting management, firefighting impact analysis.

INTRODUCTION Climate change will lead to a dramatic increase in damage from forest fires in Europe in this century. In the Mediterranean region, the average annual area affected by forest fires has quadrupled since the 1960s (24). Europe suffers approximately 65,000 fires every year, which burn, on average, half a million hectares of forested areas (10). The number of forest fires is also on the increase in Central

1

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

and Northern Europe. The Austrian forest fire database1 shows a total of 584 fires for the period 2012 to 2014. Improvements in forest fire control are a major national and international issue in the quest to protect human lives and resources and to reduce the negative environmental impact of these fires to a minimum. The key to optimal deployment of firefighting teams is to provide a situation map in near real-time and a powerful role and scenario-focused management solution. This will ensure that resources such as mobile firefighting and rescue teams, vehicles/equipment on the ground as well as firefighting planes and helicopters can be deployed efficiently and interactively. Firefighting measures differ substantially depending on the type of forest fire. Ground fires, crown fires and large-scale surface fires require different control strategies, which need to be taken into account in the management solution. Apart from the optimal deployment of emergency teams during the forest fire, it is also necessary to monitor the affected area for at least 36 hours after the fire has been extinguished in order to detect any hot spots that may rekindle the fire. Figure 1 gives an overview about the operation layers and the different types of resources which are involved in a recent forest fire situation.

Figure 1. Multi-layer forest fire management sketch. Forest firefighting operations are in general very dangerous and regarding the involvement of comprehensive human resources as well an extensive use of ground-, air vehicles and equipment, the event management is a crucial factor to protect human lives and resources and to reduce the negative environmental impact. Forest firefighting is commonly based on estimations made by 1

See http://fire.boku.ac.at/public

2

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

firefighting experts from visual observations. These estimations are subject to a great number of errors due to smoke occluding the flames, human inaccuracy in the visual estimation and errors in the localization of the fire (21). Within disaster management at wildland fire disasters, the key for effective decision support and the potential to support time critical decision processes is the ability to provide a common operational picture consisting of rapid available, up-to-date geo-data and related data analysis results (3). Solutions and concepts for airborne mapping and precise data processing already exist (e.g. 17) but they lack the possibility to deliver comprehensive geographical coverage data products in near real-time. To meet the needs for rapidly available ortho-rectified airborne data in the deployment area new methods and developments were required to ensure fast and costefficient usage (23, 1).

The airborne multi-functional management support system ARGUS, developed as part of the KIRAS2 project AIRWATCH 3 , enables near real-time monitoring of natural disasters (forest fires, earthquakes, floods, etc.) based on optical and thermal images. Airborne image acquisition and near real-time processing allow the generation of up-to-date situation maps as the basis for a common operational picture. ARGUS builds the base for innovative extensions and research developments designed to optimize command operations in national and international firefighting missions. The system can be integrated into different airborne platforms and can thus also be used for other applications of airborne surveillance like international environmental monitoring assignments. The research project used results and experiences from previous projects and focused on the development of a more flexible integrated concept taking into account both, technological aspects and extended requirements of the public stakeholders and industrial partners. It was carried out in close cooperation with the Austrian Defence Ministry and the civil protection departments from the regional governments of Styria and Lower Austria. This allowed to focus on a detailed definition of real world problems and needs as part of the requirement definition phase. One important objective was to develop a modular and easily transportable multi-sensor platform for efficient integration into various aircraft types like the Pilatus Porter PC-6 Austrian Army (AA), the DA 42 MPP (Diamond Aircraft Industries) and potentially various Cessna models. A further key objective was to realise fast scenario oriented processing workflows and an end user focused management solution. Therefore, the resulting system “ARGUS” (Airborne Real-Time Ground Units Support) provides highperformance data processing and the definition of different processing chains tailored to user requirements. A broadband line-of-sight (LOS) data link enables the transmission of acquired image data in real-time from the airborne to the ground segment and near real-time data processing (on board and/or on ground) enables an efficient support of time-critical decision-making processes in disaster management. The management solution allows a coordinated data acquisition, control over the performed geo-data processing modules, data archiving and multiple distribution channels to associated users as well as existing systems. With respect to the specific requirements of large wildland fire situations and a multi-layer forest fire management (see Figure 2) the ARGUS concept (4) is extended by further innovative developments. These will be realised in frame of an ongoing research project (3F-MS4) focusing on the following topics: 2

See http://www.kiras.at/home/?L=1 See http://www.kiras.at/projects/detail/?tx_ttnews[tt_news]=280&cHash=420a84fc92d9f5dcd318d2d2d17877a4&L=1 4 3F-MS (Multi layer Forest Fire Fighting management system) is an ongoing research project funded by the Austrian Security Research Programme KIRAS 3

3

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

    

Optimization of the thermal sensor data acquisition and treatment focusing on a rotating mirror system to enable a large-scale coverage and the required geometric treatment; Automated analysis of thermal data evaluating different fire types; Simulation of the efficient allocation of resources aligned with the requirements of different wildland fire scenarios; Support of task distribution and extended status information of connected operational units; Integration of wearable systems to assist ground teams in rescue operations as well as a mobile information system into innovative command and fire-fighting vehicles.

WILDFIRE MANAGEMENT CONCEPT Figure 2 shows an overall forest fire management concept which tries to include all essential aspects of a wildlandfire disaster management cycle (21), from forest fire prevention and preparedness to post-fire damage analysis. To reduce the risks for human lives and also environmental and infrastructure damages extensive research activities and innovative technological solutions are needed in all these areas. The ARGUS system enables to support (a) early detection of fire sources, (b) forest fire event management, (c) the post fire monitoring and (d) the damage documentation. Interfaces will allow to integrate risk maps and results of fire propagation models as well as to export a common operational picture (COP) as recent data input for fire propagation models and to distribute detailed information of an up-todate situation map to involved operational teams and decision makers.

V e g e t a t i on a n a l ys i s a n d b i o m a s s e s t i m a t i o ns Thread - „Risk Areas“ Forest fire risk index E a r l y d e t e c t i on o f f i r e s o u r c es F o r e s t f i r e – e ve n t m a n a g e m en t near realtime COP; optimized command and support Fire propagation modelling up-to-date meteorogical and local data for an risk forcast for the coming hours P o s t f i r e mo n i t o r i ng high performance related to area coverage D a m a g e d o c u m e n t at i on Land cover damage assessment and analysis of the environmental impact V e g e t a t i on r e g e n er at i o n

Figure 2. Overall forest fire management concept.

To allow an optimized support of the defined topics (a) – (d) the requirements for extending the ARGUS solution (see also 4) can be summarised as follows:     

Wide area monitoring of threatened forest surfaces on the basis of risk analyses; Extension and optimization of command and control functionalities regarding the requirements of wildland fire situations; Resource management for ground and aerial forces; Simulation of a fictitious use of resources and comparison of the simulated and real impact; Monitoring of doused forest fire areas after the “fire extinguished” (min. 24-36 hours) to enable a very early detection of fire sources and to manage targeted counter-measures.

4

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

SYSTEM COMPONENTS AND WORKFLOW ANALYSIS The ARGUS solution is designed as an airborne multi-functional management support system and meets fundamental requirements of customers and end users to support time critical disaster management processes. An important factor for the applicability is the modular and easily transportable multi-sensor data acquisition platform and the efficient integration of this sensor platform into different airplanes (Pilatus Porter PC 6 and DA42 MPP). The system enables an efficient data processing as well as the definition of different processing chains aligned to user requirements. For the use in the field of disaster management, time-optimized "near real-time" processing of data enables the support of time-critical decision processes. The management solution permits the coordination of the data acquisition, the management of the processing tasks, the data archiving and a corresponding data distribution into other systems that exist. To respond to the needs of rapid data acquisition and provision of accurate geo-products, the following key system components were developed: 

 

Airborne segment: Multi-Sensor platform including RGB, NIR, TIR as well as GPS/IMU sensors for the acquisition of image data and an application for operator control and pilot guidance. Real-time communication: Line-of-sight (LOS) connection for real-time data transfer to the ground station over a distance of up to 50 km. Ground segment: Processing and management modules including data reception, near real-time processing of image data, plus a visualisation and management solution including a simulation and task distribution module.

Figure 3. System concept. 5

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

The following chapters describe the main modules and functionalities of the ARGUS system and the ongoing developments. AIRBORNE SEGMENT The ARGUS airborne segment with its multi sensor platform is able to acquire RGB color image data as well as images in the near (NIR) and thermal (TIR) infrared range. TIR is required in particular for locating temperature maxima, which is of essential importance for the monitoring of forest fires and floods. Furthermore, TIR allows measuring temperature differences and so detecting drenching ground and seeds of fire which is relevant for monitoring and situation analysis tasks in different scenarios. The NIR sensor on the other hand is mainly used for environmental monitoring missions like the analysis of vegetation in agriculture domains and forestry. All components are integrated into a single case (see Figure 4) to be able to install it on the prepared plane (Pilatus Porter PC6 AA) within 20 minutes and therefore it can already be used in the air after a very short period of time. Apart from this solution, different camera systems (e.g. 9, 11, 12, 13, 19) for precise airborne mapping exist but these do not aim on and also do not allow near realtime availability of the images due to the amount of data and therefore are not appropriate in the field of fast support in crisis situations. Next to this, for most customers the cost factor as well as the need of flexibility of the sensor platform related to carrier platforms are important issues (2).

Figure 4. Multi-sensor components (left) and the fully integrated solution (right) of the airborne segment at a Pilatus Porter PC6 AA. The following sub chapters describe the components included in the airborne segment as well as their purpose. Camera Setups The installed RGB camera is a Prosilica GT6600C RGB with a resolution of 6576x4384 pixels (29MPx). Using a 35mm lens, a ground resolution of 16cm at a standard flying height of 1000m above ground level (AGL) and a swath width of 1030m is possible. The TIR camera is an InfraTec VarioCAM hr head 600 with a resolution of 640x480 pixels (0.3MPx) operating between 7.5 and 14 µm. As shown in Figure 4, a panning mirror is mounted in front of the TIR camera with a longer lens to ensures that it covers the same area on ground as the optical camera but still get a ground resolution of 50cm per pixel. Thereby, 12 TIR images are taken within the cycle time of one RGB image. The installed NIR camera is a Prosilica GC 2450 (5MPx) covering the same area as RGB 6

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

with the used 8mm lens. It provides a ground resolution of 44cm at 1000m AGL. If higher ground resolutions are required, a 50mm lens can be mounted on the RGB camera so that a GSD of 11cm and a swath width of 720m at 1000m AGL can be reached. In this configuration only 6 TIR images are taken for one RGB image. GPS/IMU Sensor Setup To allow a near real-time ortho-rectification, a direct geo-referencing approach is employed by using high-precision IMU and GPS values. Thereby, the inertial system is fixed to the reference system of the sensor platform and measures the rotation about all 3 Cartesian axes. Using a control PC, the GPS position and attitude data as well as the orientation data from the IMU synchronized with the image acquisition to provide accurate data packages. One aspect on IMUs is the drift error which tends to increase over time. To ease this problem, the IMU positioning data are supported by GPS and so enhance the overall accuracy to positional errors in the subdecimetre range. Also special flight manoeuvres can be used to recalibrate the IMU from time to time to keep up the needed accuracy. ARGUS uses a Novatel GPS (L1/L2) receiver and an iMAR IMU-FSAS IMU to meet the requirements of near real-time applications. On-board Control and Processing The on-board computing devices are used to fulfil the following tasks:   

Direct control of the different sensors (cameras, GPS, IMU), image capture and subsequent synchronisation between image and metadata and data storage. Data compression depending on the available data link bandwidth and subsequent data transfer. Optional on-board data processing capabilities to generate ortho-images, previews or situation specific data analysis results for flexibility and mission depending usage. Especially for missions with a low speed communication channel, this allows to download only dedicated data.

The components are mini box PCs (Pokini, ZOTAC ZBOX) including a fast, reliable and easy switchable SSD store to meet the requirements for using standard hardware within planes. REAL-TIME COMMUNICATION Transmitting sensor data during flight can be achieved at high data rates at minimal capital expenditure and operational costs by using a Line-of-sight (LOS) radio link. One main complication is that national radio frequency regulations differ widely on permissible frequencies. No dedicated frequencies have been reserved on an international or European level for data transmission yet. A key requirement for the operation in various countries is therefore carrier frequency agility. JOANNEUM RESEARCH developed a bi-directional IP traffic LOS communication system (22) that supports adjustable carrier frequencies in the range from 2 GHz to 6 GHz. The used TDD (Time Division Duplex) medium access has the advantage that only one frequency for up- and downlink is required. A further advantage is configurable allocation of the available carrier bandwidth between the downlink to the ground and uplink to the aircraft. The maximum throughput in the downlink is 8 Mbit/s.

7

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

The sensor data from the cameras on board have to be transferred in near real time to the ground station. The architecture in Figure 5 shows how the Line-of-sight (LOS) is embedded between the sensors on board of the aeroplane (optical and thermal cameras) and the ground processing centre where the near real time processing takes place. The tracking antenna on ground is directed by the received ADS-B telemetry information from the airplane. ADS-B (Automatic dependent surveillance – broadcast) is a standardized system where the aircraft determines its current position, altitude, heading and other information and broadcasts this information in regular intervals over an ADS-B transponder. The received information on the ground is used to calculate the pointing of the directional antenna on ground (Figure 5).The antenna is of nomadic nature which means that transportation in small volumes and fast installations are possible.

Figure 5. Architecture of the integrated communication system The system is based on a software defined radio (SDR) approach (see Figure 6) which enables high flexibility due to faster development times of signal processing algorithms. This circumstance allows cost effective adaption of the system to different user requirements.

Figure 6. Airborne Line-of-sight modem (left) and integrated in a 19” housing for plane integration (right). 8

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

The functionality of the LOS system was proofed in several test flights and exercises, e.g. for an international exercise with the red-cross in Austria, the system has been integrated and optimized for a bandwidth of 8Mbit/s and an required operational range of 50 km. The actual benefits like high carrier frequency agility and single carrier operation with flexible up and downlink bandwidth configuration are further extended with features like adaptive range and bandwidth extension. This allows higher downlink data rates of up to 20Mbit/s at low distances as well as dynamic data rate adaptation without service interruption for distances of up to 100km. A further enhancement planned is to use an inbound telemetry channel enabling to transmit telemetry data (current position and altitude) on top of the downlink data. This step allows replacing the ADS-B frontend on ground and providing an even more compact system which is still able to use a tracking antenna. GROUND SEGMENT Data Pre-Processing A comprehensive and modular software infrastructure, devoted to geometric image processing as well as image analysis, is available at JOANNEUM RESEARCH. This modular concept enables problem- and client-specific generation and implementation of end-to-end processing chains. Implementation and triggering of such processing chains is realized via self-developed process monitoring and controlling tools, which are based on a dedicated scripting language. The process controlling system enables image processing on a single system as well as on distributed systems including a geo-oriented priority system and data warehouse solution. This is of special significance when data processing needs to be done rapidly or even in near real-time. As an example, direct ortho-rectification can be performed synchronously to image acquisition when utilizing such distributed data processing capabilities. Thus, for this core data pre-processing task the given requirement to get ortho-rectified products in near real time available can be fairly well met. On the other hand, the software also supports post-processing modes including direct orthorectification, ‘enhanced mapping’ and ‘precision mapping’, which are based on utilization of tiepoints and ground control points. The system also features methods for automated mosaicking from numerous ortho-photos to provide a comprehensive overview of extensive areas (rivers etc.), thus substantially facilitating subsequent data processing steps. Also workflows for hot spot detection and thematic mask generations provide valuable benefits and an enhanced support of decision making processes in the field of crisis management. This is also valid for analysis for damage assessment actions, e.g. in the case of floods or forest fires (e.g. 18). Geocoding Processing Chains In natural disaster scenarios where situations and associated spatial changes evolve rapidly, the time factor is more important than absolute geometric accuracy. An absolute positional accuracy better than three meters is absolutely acceptable for real-time ortho-rectification for the involved relief forces and commanders. With an eye on the speed factor, direct geo-referencing appeared to be a good approach but results mainly depend on the accuracy of the position and orientation sensors (7). Direct geo-referencing involves combining the measured GPS and IMU data with the image data without subsequent post-processing. In principal, there are different approaches and quality levels of geo-referencing possible. This includes the mentioned direct geo-referencing, enhanced mapping and precision mapping. Enhanced mapping includes automatically found tie 9

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

points as well as an optional post processing of GPS/IMU data. This increases the geo-referencing accuracy at the expense of processing time. Precision mapping uses ground control points to enhance the absolute positional accuracy of the resulting data product. Table 1 shows the accuracies achieved by the system for the different processing modes based on the used sensors. The accuracy changes linearly with the flying height above ground level (AGL). Table 1. Geo-position accuracy of differet processing aproaches. Geo-referencing level

Flying height AGL

Positional accuracy

Direct geo-referencing

1000m

~ 1m

Enhanced mapping

1000m

~ 0.5m

Precision mapping

1000m

~ 0.15m

Image Data Analysis The acquired image data should be used for automated image analysis tasks. Figure 7 shows subsets of the thermal information (left) and the according RGB color image (right) of an artificially set forest fire situation. The thermal image is visualized in pseudo colors whereas the ground temperature (blue) is at 13 degrees and the hottest fire spot (red) is at 259 degrees. These geometrically coherent multi-spectral data is the basis for further detailed image analyses.

Figure 7. Annotated TIR (left) and optical (right) images of a forest fire situation. In the past the thermal information was often manually interpreted (5, 7, 8), which however is inefficient for the envisioned forest fire management system. Obviously, the detection of forest fire hotspots is a rather simple computer vision task as such regions can be determined by appropriate thresholding operations of the TIR image. Such processing was e.g. presented by (20) or by (14) where the latter extracts fire contours employing learning based thresholding. However, the envisioned computer vision system should be capable of a much more detailed analysis of the current state, i.e. in particular a classification of different forest fire types and situations, which was not tackled in recent literature. For efficient fire management it is not only important to differentiate between burning areas and burned areas, but also between forest fire types. Most important are burning ground (including underground hotspots and wildland fire), crown fires and full developed or surface fires. Such fire types vary in size and extend as well as their peak temperature. 10

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

Our work in progress is focused on classifying the TIR and optical information jointly within a training based classification scheme. Appropriate TIR features are size and shape of the hotspots and statistics of the temperature within hotspots (minimal, maximal, mean, and fluctuation values or histograms). For training real data will be used to learn classification models, e.g. support vector machines (SVN), deep neural networks (DNN) or convolutional neural networks (CNN). Moreover also optical features will be incorporated in the learning as e.g. a burning region can be very well distinguished from a burned region by their optical appearance. Those methods will be fined tuned to exactly match the frequency band of the applied TIR camera. The underlying algorithms and processing chains will be related our previous works (15, 16). The classification results will then be the basis and pre-requisite of the management systems and the simulation. Visualization and Management Module After the automatic ortho-rectification of the gathered images, the results are available for visualization and further on-demand processing (e.g. mosaicking, generation of hotspot maps, etc.). The following figure shows examples of the geo-oriented data visualization in the management module.

Figure 8. Visualisation and data management module.

The geo-oriented management interface of ARGUS offers the following features, which are extended for the use in forest fire situations:     

Data acquisition planning and real-time mission control; Situation monitoring based on images of the affected areas; Temporal/spatial search functions; Comparison of historical and current images for assessing changes; Trigger semi-/automatic data analysis processes and visualization the results; 11

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

       

Visualisation available geo-data and positions of operational units; Creation of value added products (e.g. mosaics, hotspot maps, 3d reconstruction); Thermal image analysis and target spotting; Enhanced selection and data distribution to external systems (e.g. Intelli R.4C); Forest firefighting simulation module. Text based communication with the operator on board; Communication with connected operational units (e.g. plane, helicopter, field forces); Command units and task allocation;

Management of Tasks and Resources To support the already described concept of a multi-layer management in forest fire scenarios, ARGUS will be extended with modules that allow the operator to perform the needed actions. Concrete these are possibilities to define areas which should be monitored based on actual events. A simplified exemplary workflow consists of seven main steps. At first (step one) the ARGUS operator plans the area and selects the available units and has an overview of available infrastructure objects like water pickup locations, landing places and possible no-fly zones (like neighbouring hospitals or motorways) for firefighting aircrafts. Step two is planning the initial route including flight/camera relevant parameters for the surveillance plane (layer 3). This step performs an optimized flight plan to monitor the area in an efficient way. The calculation also considers no-fly zones (see Figure 9). This route is sent to the surveillance plane and it starts to cover the whole area of interest sending aerial images back to the operator (step three). The fourth step includes semi-/automatic data analysis processes for the support of the operator to assign new tasks to different operational units of layer two. While or at latest when these tasks were performed (step five/six), the operator might send a surveillance plane to the defined area for inspecting the impact and to decide whether Figure 9. Flight track considering no-fly zones. further firefighting actions are necessary or not (step seven).

12

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

The steps four to seven are repeated until all fire spots are extinguished. Figure 10 illustrates this course of action.

Figure 10. Processes in airborne forest fire monitoring Another extension is a task distribution module. It includes the visualization of completed and still open tasks of operational units. Based on a situational map including infrastructure objects (landing strips, extinguishing water reservoirs, etc.), operational units can be ordered to monitor areas, pickup water, move to targets etc. Thereby every operational order is recorded and an event log is created automatically to support documentation requirements (see Figure 11 and Figure 12).

Figure 11. Module for task allocation and management.

Figure 12. Example of a digital mission protocol.

To achieve the described goals, also a messaging module is needed to exchange data between ARGUS ground segment and the operational units in the air and on the ground. This aims at getting an optimized feedback from currently performed task of the different involved parties. Furthermore it is also requires a communication channel to the used types of ground vehicles (forest fire engines, control units as well as support units). The basic possibility to solve this is using TETRA which allows using a small bandwidth to exchange command and control messages. 13

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

For solutions with higher bandwidths (e.g. WLAN, LTE, Tetra over LTE), the systems supports using more bi-directional data exchange between the control point and vehicles and operational units in the field. These data include telemetry data, water tank status, availability status information etc. which is used to generate an extended operational picture and allows an efficient usage of available resources. Simulation Module Currently many research approaches for simulation modules exist, which focus on the simulation of fire spread (e.g. TechForFire-Service5) including the usage of meteorological data, vegetation and biomass data. The goal of the simulation module in frame of 3F-MS is to simulate different situations of forest fire and analyse the potential impact of different available firefighting resources. Based on these results, the most efficient way to extinguish the fire will be derived. On the basis of up-to-date situation data from a surveillance plane using the ARGUS airborne segment, the available ground based and airborne firefighting resources (water bomber planes, helicopters, fire trucks, and others as well as operational units, a simulated impact will be calculated and a proposal for an efficient usage of forces is provided to the commanding officer. The officer in charge will command the units to fight the fire where also the surveillance plane will be ordered to record the actual impact of the measures (see also Figure 10). Based on the analysis between calculated impact and real impact, the system will compute an updated suggestion of the intervention tactic and the needed allocation of firefighting resource. The module can use different sources as basic data for the simulations. These are on the one hand the mentioned analysis from the ARGUS system providing near-real-time situational data and on the other hand simulation results from external data sources like fire spread models, local operational units etc. This also allows to covers the temporal aspect of predicting needed resources for the situation in e.g. one hour, when firefighting units are available on site. EXTERNAL INTERFACES The generated information products of the ARGUS system are available in standardized data formats to allow an easy exchange with other, existing geo-information systems (e.g. local/regional GIS systems, geo-information systems for disaster management, etc.). Also, the results of the data processing and also analysis processes can be distribute offline via standard file transfer as well as online via ftp or OGC standard compliant interfaces. This means that staff experts always have access to relevant up-to-date data on site and are able to use their existing command and control system. Import interfaces allow the usage of available local raster data like risk maps, basic data products and data services like Copernicus EMS6. In addition results can be displayed using a web based geo-view including basic data from Open Street Maps (OSM) or Goggle Maps / Bing Maps and are available on different platforms – web, desktop, mobile devices. The distribution system also includes services for mobile units to provide up-to-date situation maps to involved operational units.

5 6

NOVELTIS – TechForFire; an operational Service for Tactical Firefighting; see also www.noveltis.com See http://emergency.copernicus.eu/mapping/

14

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

CONCLUSIONS AND OUTLOOK The number of forest fires increase in all regions in Europe. The risks for human lives and also environmental and infrastructure damages are becoming larger and reached a very critical level in some regions of Europe. To reduce the negative impact of wildfire and forest fire situations it will be very important to realize extensive research activities and innovative technological solutions to enable a comprehensive support of the full wildlandfire disaster management cycle. In general, to manage disaster events it is a fundamental requirement to offer timely and holistic information products to allow the support of time critical decision processes of the involved organisations and persons. This paper presented a multi-functional airborne management system which allows to support crisis management tasks of emergency teams and armed forces in disaster situations based on the generation of a near real-time situation map and a powerful role and scenario-focused management solution. In a close cooperation with respective first responder groups, governmental civil protection organisations and departments as well as the Austrian Armed Forces (Ministry of National Defence and Sport) the ARGUS system were developed and tested in exercises and real operations. The results of these cooperations have continuously flown into the developments. The evaluating results concerning the management support of firefighting scenarios documented the need for further developments focusing on the optimization of the thermal sensor data acquisition and automated processing as well as developments for an innovative multilevel management approach. This multilevel approach considers the involvement of comprehensive human resources as well as extensive use of ground-, air vehicles and equipment. The development of simulationbased modules for decision support and impact evaluation and an innovative development for the large-scale monitoring of affected forest areas will enable an efficient resources management during an event situation. In the frame of this article also first results of the ongoing research project were presented. An important part for the future activities will be the realisation of interfaces to allow the integration of risk maps and results of fire propagation models as well as to export a COP (current situation map) as up-to-date data input for fire propagation models and to distribute detailed information of an up-to-date situation map to involved operational teams and decision makers. In frame of the 3FMS projects cooperations with the European Commission Joint Research Centre (IES - Forest Resources and Climate Unit), NOVALTIS (TechForFire) will be achieved to work specifically on these objectives. In general, a very important question for future activities will be to bring together the already existing effective solutions to offer an entire service regarding the issues of forest fires for the national civil protection organizations. ACKNOWLEDGMENTS The research project was carried out in close cooperation with the Austrian Defence Ministry. The presented research activities were embedded into a project running within the Austrian national promotion programme for security research (KIRAS; see also http://www.kiras.at/) and funded by the Austrian Research Promotion Agency (FFG).

15

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

REFERENCES 1.

Almer A., T. Schnabel, H. Raggam, KH. Gutjahr & M. van Dahl, 2007. Rapid Information Flow within a Crisis Management System. 27th Earsel Symposium.

2.

Almer A., H. Raggam, T. Schnabel, M. Schardt, R. Wack, KH. Gutjahr, M. Schmidt & O. Koudelka, 2011. Flugzeugbasierte Low-Cost-Aufnahmeplattform für den internationalen Katastrophenschutz. AGIT.

3.

Almer A. 2013. Airborne data acquisition and processing technologies for improved situational awareness. PSC Europe Forum Conference.

4.

Almer A., T. Schnabel, A. Köfler, H. Raggam, R. Wack & R. Feischl, 2015. Airborne multi-sensor management support system for emergency teams in natural disasters, ISCRAM 2015.

5.

El-Sheimy N. & B. Wright, 2004. Real-time Airborne Mapping System for Forest Firefighting (F3) System. PERS, Vol. 70, No. 4, 381-383.

6.

Esposito F., G. Rufino, A. Moccia, P. Donnarumma, M. Esposito & V. Magliulo, 2007. An integrated electro-optical payload system for forest fires monitoring from airborne platform. Aerospace Conference, IEEE, 1-13.

7.

Gutjahr KH., P. Hafner, M. Ofner., K. Längauer, M. Wieser & N. Kühtreiber, 2010. Performance of GNSS/IMS Integration Methods in Context of a Near Real-Time Airborne Mapping Platform. EuroCOW 2010: the Calibration and Orientation Workshop, ISPRS Commission III, WG III/1.

8.

Halikias G., G. Leventakis, C. Kontoes, V. Tsoulkas, L. Dritsas & A. Pantelous, 2011. Design Issues of an Operational Fire Detection System integrated with Observation Sensors. Chapter 5 in Advances in Satellite Communications. Edited by Masoumeh Karimi and Yuri Labrador, ISBN 978-953-307-562-4, 206 pages, Publisher: InTech. DOI: 10.5772/838.

9.

Hinz A., C. Dörstel & H. Heier, 2000. Digital Modular Camera: System Concept and Data Processing Workflow. Archives for Photogrammetry, Remote Sensing & Spatial Information Sciences Vol. XXXIII.

10. JRC European Commission, 2010. Forest Fires in Europe 2009, JRC Scentific and Technical Reports, no 10, European Union, EUR24502EN, ISBN 978-92-79-16494-1, ISSN 1018-5593, doi:10.2788/74089. 11. Kremer J., 2010. The Quattro DigiCAM – IGI’s Versatile Aerial System for Various Aerial Imaging Tasks. DGPF Tagungsband 19/2010, 623-630. 12. Leberl F., M. Gruber, M. Ponticelli, S. Bernögger & R. Perko, 2003. The UltraCam large format aerial digital camera system. Proceedings of the American Society for Photogrammetry & Remote Sensing, Anchorage, Alaska. 13. Leica, 2011. Leica ADS80 Airborne Digital Sensor, Datasheet, http://www.leicageosystems.com/downloads123/zz/airborne/ads80/brochures-datasheet/ADS80_datasheet_en.pdf. 14. Merino L., F. Caballero, R. J. Martinez-de-Dios, I. Maza & A. Ollero, 2012. An unmanned aircraft system for automatic forest fire monitoring and measurement, Journal of Intelligent & Robotic Systems, vol. 65 (1-4), 533-548. 15. Perko R., T. Schnabel, G. Fritz, A. Almer & L. Paletta, 2013. Airborne based High Performance Crowd Monitoring for Security Applications Scandinavian Conference on Image Analysis (SCIA), Springer LNCS, 2013, 7944, 664-674.

16

10th EARSeL Forest Fire Special Interest Group Workshop Limassol, Cyprus, November 2-5, 2015

16. Perko R., T. Schnabel, A. Almer & L. Paletta, 2014. Towards View Invariant Person Counting and Crowd Density Estimation for Remote Vision-Based Services IEEE Electrotechnical and Computer Science Conference, 2014, B, 80-83. 17. Petrie G., 2011. Tiltan’s Automated Geo-Map - A Highly Integrated Airborne Image Data Acquisition & Processing System. GeoInformatics. 18. Raggam H., R. Wack & KH. Gutjahr, 2006. Assessment of the Impact of Floods using Image Data acquired from a Helicopter. 26th Earsel Symposium, “New Developments and Challenges in Remote Sensing”. 19. Raggam H., R. Wack & KH. Gutjahr, 2007. Mapping Capability of a Low-Cost Aerial Data Acquisition Platform - First Results. ISPRS-Workshop, Commission VI, WG VI/4, “High Resolution Earth Imaging for Geospatial Information”. 20. Salami E., S. Pedre, P. Borensztejn, C. Barrado, A. Stoliar and E. Pastor, 2009. Decision Support System for Hot Spot Detection. Intelligent Environments, 277-284. 21. San-Miguel-Ayanz J., E. Schulte, G. Schmuck, A. Camia, P. Strobl, G. Liberta, C. Giovando, R. Boca, F. Sedano, P. Kempeneers, D. McInerney, C. Withmore, S. Santos de Oliveira, M. Rodrigues, T. Durrant, P. Corti, F. Oehler, L. Vilar & G. Amatulli, 2012. Comprehensive Monitoring of Wildfires in Europe: The European Forest Fire Information System (EFFIS), Approaches to Managing Disaster - Assessing Hazards, Emergencies and Disaster Impacts, Prof. John Tiefenbacher (Ed.), ISBN: 978-953-51-0294-6, InTech, DOI: 10.5772/28441. 22. Schlemmer H., E. Türkyilmaz, M. Schmidt, J. Ebert & H. Mayer, 2015. Design and development of a high throughput airborne line-of-sight communication link, e & i Elektrotechnik und Informationstechnik: volume 132, issue 6, 282-288. 23. Schnabel T., 2013. Rapid Mapping - Multisensorale Bilddatenprozessierung in Krisensituationen. Innsbrucker Hofburggespräch. 24. WWF, 2012. Wälder in Flammen – Ursachen und Folgen der weltweiten Waldbrände; WWF Deutschland, Berlin; July 2012, 6th edition, http://www.wwf.de/fileadmin/fm-wwf/PublikationenPDF/120809_WWF_Waldbrandstudie.pdf.

17