The evolution of unmanned aerial vehicles and the data from their sensors is noticeable in most of the areas of environmental sciences, forests being inevitable part of them. The interest in ultra-high resolution raster data for forestry purposes ranges from optical imagery and high detail mapping, through three-dimensional point clouds from Structure from Motion techniques for geometrical information including tree amount, diameters, height and stand volume till spectral features extraction where it is expected to obtain information about species composition, species characteristics (for example in our case ecotype origin) and finally health status or fragility.
In this work we present preliminary results of three studies deploying Micasense Rededge camera for automatic differentiation of individual regions of interest (ROI) in raster mosaics from unmanned aerial systems. We describe the data processing workflow, from acquisition of the images, through processing and calculation of individual band reflectance, with special attention to spectral calibration and reflectance normalization of individual bands. Finally we compare different statistical methods for evaluation of the reflectance (and spectral indices) their feasibility and practical outputs for forest inventory and planning.
The case studies presented include species classification in mixed stands in central part of Czech Republic, differentiation of various Norway spruce ecotypes and evaluation of damaged trees by bark beetle in Bohemian Switzerland National Park and south Moravia region. The special attention is payed to time changes in the same class (the same ecotype, the same specie and the same healthy/damaged tree) over time and mostly with respect to the ratio indices and their fluctuations.

Area B - Board 202

Bioclimatic models forecast redistribution, assemblage change and extinction risk for many species as a result of the anthropogenic climate change. However, these predictions mostly rely on coarse interpolated standard meteorological measurements and neglect microclimate heterogeneity relevant for species living near the ground. To not take into account that aspect can lead to overestimation of potential impacts of climate change.

In our study, based on very detailed 38 microclimatic measurements of soil, near the ground and air temperature, we point out that even within a 20-ha temperate forest air temperature varied greatly, showing nearly 2° C differences for a vegetation period average, and even more for temperature extremes. Contrary to expectations, geomorphometric characteristics showed minimal variability, suggesting the dominant role of forest canopy cover in influencing land-atmosphere fluxes of energy and water, and thus significantly modifying observed surface temperature.

To confirm the link between canopy cover and forest microclimate, we assessed canopy openness properties from the analysis of hemispherical photographs. However, this method provides purely discrete information about the local forest canopy architecture, moreover is laborious and difficult to scale up. Therefore, we discuss several alternative forest structural metrics (e.g. canopy closure, canopy cover, canopy density, and others) derived from UAS laser scanning and UAS optical-based methods as spatially continuous and highly detailed proxies of surface temperature.

Interestingly, certain predictability was found for soil temperature average and extremes, whereas air temperature was difficult to quantify by remote sensing microclimate proxy variables caused by lack of information or excessive complexity of microclimatic processes.

Overall, though our results showed the medium strength of statistical relationship (probably because of partial stochasticity of microclimate), we consider the remote sensing of microclimate proxy variables capable to determinate a critical function of forest in buffering regional climatic condition and improve future bioclimatic models at a landscape scale.

Area B - Board 203

Unmanned aerial vehicles (UAVs) or drones have been providing benefit to the RSPB’s nature conservation work for several years. This presentation highlights a collection of case studies where we’ve successfully used drones to improve the natural environment.

Surveys by drone have many advantages. Often drone surveys are quicker than field surveys. They can reduce disturbance in surveys of bird colonies. Remote and inaccessible locations are more easily accessed by drone, and the survey can be considerably cheaper than a ground survey.

The conservation value of drones has increased recently, due to the availability of low cost consumer-grade drones. Over the last couple of years, the RSPB has worked to improve the quality of the data we collect with drones, and ultimately get more data mapped and used in Geographic Information Systems (GIS). To date, most of our drone data has been collected with RGB cameras, allowing us to create on-demand, high resolution aerial imagery. We have used drone imagery to plan conservation work, including habitat creation, and peatland restoration work. Drone imagery allows us to document changes, for example, before and after habitat management. Using standard RGB imagery, we’ve used semi-automated classification to map exposed peat at a peatland restoration project site.

Drones are invaluable for surveying locations that are difficult to access, such as seabird cliffs and small islands.

Using photogrammetry, we have used drone imagery to create digital surface models (DSMs), which we’ve used to model topography, and drainage networks. Moving forward, we plan to extend our use of drone-derived DSMs to hydrological engineering.

We’re now working to further develop our conservation use of drones. We’re looking to investigate the advantages of more technologically advanced sensors, including multispectral and thermal cameras. Thermal cameras, which detect heat emitted from objects, can be mounted on drones. In preliminary tests, we've used them to locate camouflaged bird nests. We’re also keen to experiment with multispectral cameras, particularly those that detect infrared light. Incorporating infrared data allows semi-automated classification of many types of vegetation in a landscape, which in turn permits quantification of vegetation change over time. It also allows calculation of vegetation indices, such as NDVI, providing a quick assessment of vegetation vigour. With vegetation indices, we will have a robust measure to quantify change over time, and thus to assess the effectiveness of our conservation work.

Area B - Board 204

“Earth observation is diversifying […]” (Grainger 2017). This statement by Alan Grainger condenses recent and future developments incorporating in situ, unmanned aerial vehicle (UAV), airborne, spaceborne, and citizen observatory based measurements to monitor state and processes of the earth. In the past years citizen science for Earth Observation gained great recognition by policy. E.g., the Horizon 2020 Programme is funding several citizen science observatory projects. These developments are the consequence of the main bottle neck in Earth Observation: missing reference data. Foresightful, some scientists (e.g. Dell'Acqua and De Vecchi 2017; See et al. 2016) discuss opportunities to incorporate volunteered geographic information (VGI) in political frameworks such as the Copernicus program to complement the in situ component. Successful projects such as Geo-Wiki or Crowd4Sat demonstrate the vast potential of citizen science in Earth Observation. However, to date all projects and initiatives focus on 2D information such as single photographs or the manual classification of satellite imagery. What is entirely missing is the third dimension which is mandatory for several observation applications such as vegetation biomass estimation or the identification of damage in buildings for disaster management.
Currently, up-to-date 3D reference information is hardly available. The acquisition of such data causes high efforts either in terms of labour or in terms of equipment (laserscanners etc.) – or both.
The advent of low cost (< 1000 €) and easy to fly light weight drones (< 5 kg) has the potential to supplement the need for 3D reference information. The heavily growing interest in UAVs can be illustrated by the following figures. For 2013 (MarketsandMarkets 2013) assessed that the global UAV market is worth 5,400 M€. The forecast for 2018 (made only 5 years ago) was a value of 6,350 M€. Now in 2018 the new assessment says that the global UAV market is worth 15,420 M€ and will grow up to 44,460 M€ in 2025 (MarketsandMarkets 2018).
Although the great potential of Structure-from-Motion (SfM) based methods for 3D reconstruction has been demonstrated for a wide range of applications, several issues need to be solved to foster trust and acceptance in this kind of data by the scientific community and administrative bodies, in particular if this data are acquired by citizens. The issue with the potentially highest priority is the quality of the generated 3D point clouds derived from multiple views.
To address this issue, on 1st November in 2018 a comprehensive UAV campaign was accomplished close to the city of Jena, Germany. During this campaign, eight drones equipped with dissimilar cameras operating in different modes acquired data over an area featuring different land cover types including forest, buildings, grassland, and water surfaces. This UAV campaign, which included among others/also citizen scientists as drone pilots flying their own UAVs, was accompanied by DLR’s 3K sensor to gather a high grade reference dataset. Additionally, airborne LiDAR data with a nominal point density of 4 points per m² are available as well. The flight altitude of the drones was 100 m above the ground, the image overlap along and across track was 75%. The processing of the data was accomplished by applying Agisoft Photoscan 1.4.4. The quality assessment of the resulting point clouds comprised the vertical (height) and horizontal (2D location) accuracy for different objects and surfaces within the test site. Although the evaluation of the results is still ongoing, first results promise a wide usability of low cost drone data acquired by citizens.


References
Dell'Acqua, F., & De Vecchi, D. (2017). Potentials of Active and Passive Geospatial Crowdsourcing in Complementing Sentinel Data and Supporting Copernicus Service Portfolio. Proceedings of the IEEE, 105, 1913-1925
Grainger, A. (2017). Citizen Observatories and the New Earth Observation Science. Remote Sensing, 9, 30
See, L., Fritz, S., Dias, E., Hendriks, E., Mijling, B., Snik, F., Stammes, P., Vescovi, F.D., Zeug, G., Mathieu, P.P., Desnos, Y.L., & Rast, M. (2016). Supporting Earth-Observation Calibration and Validation A new generation of tools for crowdsourcing and citizen science. IEEE Geoscience and Remote Sensing Magazine, 4, 38-50
MarketsandMarkets (2013): Unmanned Aerial Vehicle Market (2013-2018). Technical report. MarketsandMarkets. Dallas, TX, USA.
MarketsandMarkets (2018): Unmanned Aerial Vehicle Market (2018-2025). Technical report. MarketsandMarkets. Dallas, TX, USA.

Area B - Board 205

The Water Frameworf Directive (WFD) requires Hydromorphological assessments (HYMO) of rivers and streams to be part of the operational surveillance programs of each Member State at six-year intervals, which forms the basis of an extensive surveillance network. In this context, the EU-funded REFORM project has developed a process-based, multi-scale river hierachical framework (RHF) to help river managers assess the HYMO character of EU rivers (Gurnell et al. , 2016). One of the most important results of the RHF has been the development of key indicators of process control and HYMO forms at different spatial scales (from the catchment scale to the reach scale) (del Tánago et al., 2016 ). The main objective of the MONHYSENSE project is to develop an appropriate operational methodology using new remote sensing (RS) technologies to monitor and characterize the HYMO status of different river systems at the reach scale. More specifically, the project aims to determine the most cost-effective technology between high-resolution (mainly hyperspectral and LiDAR) data obtained from airborne platforms or Unmanned Aerial Vehicles (UAV) to monitor some of the HYMO indicators identified as important at the reach scale by the RHF (del Tánago et al., 2016) and therefore define an operational methodology for mapping these indicators.

Area B - Board 206

Monitoring soil consumption due to marble quarries and the assessment of the compliance with licence conditons are of paramount importance for enforcing a sustainable development of timing activities in the Apuan Alps district. According to the Tuscany Region administration, since the last two years the Environmental Protection Agency of Tuscany Region (ARPAT) has developed an innovative approach for monitoring quarries by using satellite images (Sentinel-2/1, Pleiades and Cosmo SkyMed), aerial data, UAV surveys and terrestrial laser scanner.
The extremely-varying quarry types - underground, on top of peaks, with high-sloped wells - make remote sensing data not very suitable for precision monitoring: on the other hand, UAV and terrestrial lidar surveys are very expensive for a basin-wide monitoring plan including more than 80 quarries.
Different resolutions and regions of interest have been investigated, and a reliable procedure for monitoring both the enlargement of mining areas and the extracted volumes has been defined by integrating optical and lidar data, leading to the definition of 2d/3d change indicators: a general environmental impact indicator has been derived from these ones. In addition to a priori knowledge from authorized technical documentation, these indicators may be used to plan precision UAV or terrestrial lidar surveys on surface quarries by impact severity classification.
In the present work the opportunities of the integration of different approach and the main results with the achieved precisions have been investigated; costs and benefits of the adopted solutions are also presented, taken into account both human and processing resources ones. Six UAV surveys and one mixed underground/surface lidar survey have been used for both assessing remote sensing techniques and results and precise 3d monitoring of various authorized excavation phases in support of control activities. Both methods showed a great capability of achieving the required precision in volume change detection and quantification, making them suitable for public control activities.

Area B - Board 207

Urban green areas and green infrastructure are becoming of greater importance as we have to take into account to much water and at the same time with water scarcity as result of noticeable changes at the seasonal level (decrease of precipitation in the summer and increase in winter) which are especially alarming.
The predicted changes in climate extremes, among others, point to substantial warming in temperature extremes by the end of the 21st century and to the likelihood that the frequency of heavy precipitation or the proportion of total rainfall from heavy falls is increasing. The simulations of climate scenarios for Slovenia until the end of the 21st century show a significant increase in the average annual temperature for the whole country in all seasons and the intensification of precipitation changes.
The challenges associated with extreme climate events are numerous and complex, for instance, ageing urban infrastructure commonly unprepared for extreme storm events. I such circumstances, urban green areas becomes of even greater importance.
We are using Unmanned Aerial Systems (UAS) RGB imagery to detect urban green areas using different RGB based indices. By processing visible spectrum orthophotos, we can obtain some vegetation indices (VARI, TGA, Green Leaf Index, etc.) for visual (RGB) data. When we combine colour bands from red-green-blue RGB photos, we can get some graphical indicators to separate green areas (woods, meadows, gardens) from the built environment (buildings, roads, pavement areas).
We compere UAS, airborne and satellite orthophotos as a source of our investigation.
The field of our case study is part of the city of Maribor together with several visual spectre indices.
We see UAS as an affordable and convenient tool for small area observations.
UAS are optimal sources of spatial data especially in the case of semi size project like ours. They are filling the gap between satellite and aerial data on one side and common terrestrial geodetic survey data on the other hand. Even prosumer drones are giving us excellent results under some restrictions (Ground Control Points (GPC), permissions, weather conditions ...).

Area B - Board 208

To detect inconsistencies in any type of data, first it is necessary to define what is considered to be anomalous. Usually, elements with characteristics opposed to those reputed as normal of local and/or global neighborhood elements in the dataset are deemed anomalies. Some spots in grapevines leaves can signal the existence of biotic or abiotic problems, which may negatively influence both yield and quality. Several techniques are used for detecting those spots. Traditional techniques for diseases detection in viticulture are conducted by in-field visual inspection or through laboratory analysis. Even though these techniques can have good results, they are a very time-consuming and demanding process when considering agricultural fields with considerable dimension. Indeed, crop diseases are responsible for major economic losses in the agriculture worldwide. However, treatments are the most effective on early detections because future spreading can be stopped and appropriate treatment can be applied. Therefore, having a crop consistently monitored can be very important to detect the first signs of a disease. As such, their early detection and mitigation is a key to make agricultural production more efficient and sustainable. A more efficient, cheap and quick alternative is by means of automatic processing. Due to their flexibility, adjustable granularity and low operational costs, Unmanned Aerial Vehicles (UAVs) are the most suitable remote sensing platform to accomplish the task.
In this study, computer vision techniques using images from the visible (RGB) and near-infrared (NIR) portion of the electromagnetic spectrum were explored for this purpose. A comparison between three automatic techniques - unsupervised K-means clustering, supervised artificial neural network detection and vegetation indices – is also presented. Results show that anomalies that can be detected in RGB/NIR part of electromagnetic spectrum are well identifiable by using vegetation indices, however, when neural network is well trained it seem to have a great potential for addressing anomalies in grapevines leaves as well.

Keywords: Precision viticulture, Grapevines leaves, Anomalies detection, Machine learning, Digital image processing, Vegetation indices, K-mean clustering

Acknowledments
The authors would like to acknowledge ERDF and North 2020 - North Regional Operational Program, as part of project “INNOVINE&WINE - Vineyard and Wine Innovation Platform" (NORTE-01-0145-FEDER-000038) for partial funding as well as the ERDF - European Regional Development Fund through the Operational Programme for Competitiveness and Internationalisation - COMPETE 2020 Programme within project POCI-01-0145-FEDER-495

Area B - Board 209

Radar Interferometry (InSAR) can provide measurements of surface displacement from Space, with millimetric accuracy. These measurements are used in natural hazards analyses but also for monitoring anthroprogenic activities. In the last few years, the number of SAR satellites with shorter repeat intervals and higher resolutions make increase significantly SAR data volume.
Multi-temporal (MT) InSAR techniques provide time series analyses of SAR data. In others words, MT InSAR enables the deformation monitoring of natural phenomena and anthropogenic activities/structures at an unprecedented scale, with hundreds (even thousands) of “measurements” (scatterers) per km2. These observations can be updated weekly, making possible to detect potential hazards at an early stage. However, this is only possible with the ability to dynamically process and update these billions of measurement points, each time a new observation (SAR image) is available. Moreover, MT InSAR observations are ambiguous, yielding erroneous interpretations making such analyses numerically challenging.
On the other hand, and despite many applications and improvements offered by new SAR missions, standardized procedures allowing for the effective use of InSAR data are lacking. Displacement mapping methodologies are still frequently based on the visual inspection of average line-of-sight velocities. However, the spatial distribution of velocities as visible from a standard point cloud map, is often scattered and affected by noises and offers no reliable information on the probability of movement for intervening areas. Moreover, the information extraction from MT InSAR techniques is sometimes difficult due to a possible large number of PSs, thereby entailing long interpretation times.
In this work, we use Machine Learning Algorithms to extract relevant information from MT InSAR time series. We focused on developing and demonstrating methodologies to identify InSAR deformation patterns with ML algorithms. In fact, the success reached by MT InSAR techniques, when applied for monitoring of natural/anthropogenic hazards, was reached mostly after the disaster/collapse consummated. However, it seems that some of the hazards exhibit precursory deformation patterns, which could serve as a warning. Identifying and classifying those patterns in the data therefore is of importance to prevent the catastrophic events and the associated casualties and damage.
In these ongoing project we present the preliminary results of our ML approach applied to different scenarios of deformation patterns to systematically detect patterns in the data which are indicative of impending hazards. We also present some qualitative measurements related to the efficiency of the algorithms when applied to large data volumes.

Area B - Board 210

Unmanned aerial vehicles (UAVs) have received increasing attention as a platform for assessing vegetation canopy biophysical properties in recent years, particularly as a means of providing detailed information on vegetation condition to support applications in precision agriculture. A variety of studies have investigated the retrieval of variables such as leaf area index (LAI) from the spectral information provided by UAV-mounted digital cameras and multispectral sensors, which can provide very high spatial resolution visible and near-infrared (VNIR) imagery at the field scale. Nevertheless, several challenges remain, including variable illumination and bidirectional reflectance distribution function (BRDF) effects, in addition to issues related to image calibration, mosaicking and normalisation. In this contribution, we investigate an alternative method for deriving spatially explicit estimates of LAI, making use of UAV-based digital hemispherical photography (DHP). Used to obtain multi-angular measurements of canopy gap fraction, DHP is a standard and widely used technique for ground-based LAI measurement, but has so far received little attention from practitioners working with UAVs.

Data collection was carried out over a winter wheat field in Southern England at three different growth stages in 2018. A 3DR Solo quadcopter was equipped with a GoPro HERO 4, whose fisheye lens provides angular sampling over a 149° field-of-view. The quadcopter was also equipped with a MAPIR Survey2 digital camera, enabling the capture of VNIR imagery. The acquired DHP data were automatically classified and processed to yield point-based estimates of LAI, which were then interpolated to produce a gridded LAI dataset. To facilitate comparison, vegetation indices were also calculated from MAPIR Survey2 VNIR orthomosaics, enabling empirical LAI retrieval. To validate the UAV-based LAI estimates, independent ground measurements were performed, using both DHP (acquired with a Nikon Coolpix 4500 and FC-E8 fisheye lens) and a LI-COR LAI-2200C Plant Canopy Analyser. Using these datasets, we evaluate the UAV-based DHP approach and compare the results to empirical LAI retrievals obtained using the VNIR orthmosaics. Implications for applications in precision agriculture and earth observation (EO) product validation are discussed.

Area B - Board 211

Detailed description of plant characteristics and its dynamics is highly desired for agriculture and environmental management. They include canopy structure based on the green fraction, green area index, plant height, leaf orientation, as well as plant density or stem and ear density. They may also include leaf characteristics such as leaf chlorophyll content and disease infestation. Further, the dynamics of some of these variables allows to access to the plant phenology such as emergence, maturity and flowering.
The sensors used, the operating mode and interpretation algorithms needed to estimate the former cited plant characteristics from UAV observations are described. Emphasis is put on the exploitation of the very high spatial resolution accessible from UAV, i.e. from millimetre to centimetre, either on empirical machine learning methods including convolutional neural networks, or from the physical principles based on radiative transfer model inversion. The several methods are described, with due attention on their validation and the corresponding accuracy is reported.
Finally, several approaches used to scale-up with satellite data these local measurements to larger domains that are not easily covered by UAV systems because of the low altitude and autonomy are described.

Area B - Board 212

We settle an UAS survey above the Longitudinal Valley Fault zone (LVF) which is one of the activest tectonic place of Taiwan (average 3cm/year interseismic shortening velocities) in order to better analyse, characterize and quantify this major Taiwan active tectonic geodynamic suture situated in between both Eurasian and Philippine Sea Plates.
We acquired 17483 high resolution photographs through two drones flying at 350 meters above ground level, with a total of 23 flight missions, that cover a 195km2 total studied area. After classical photogrammetric processing, we calculated the high resolution Digital Terrain Model (HR-DTM) of the LVF zone (with a 11cm planimetric resolution and 40cm vertical accuracy) as well as its immediate surroundings along a "buffer" zone of 2.5km. It gives us a precious high resolution intial "T0" DTM in order to better constrain the LVF topographic evolutions through time under those so active tectonic deformations.
Consequently, this UAS HR-DTM enables us through classical morphostructural interpretation (Deffontaines et al. 1991, 1993...) to map into much details the active tectonics structures of the Longitudinal Valley Fault that lead us to up-date pre-existing published works (e.g.: CGS Coastal Range geological map, 1993; Lin et al., 2009; Shyu et al., 2005, 2006, 2007, 2008).
Moreover, we used and combined this new morphostructural interpretation with long-time levelings (Hsu Y.-J. et al., 2018), GPS monitoring (Yu S.-B. et al., 1997, Hsu Y.-J. et al., 2018) and 2007-11 PSInSAR datasets (Champenois et al., 2013, Deffontaines et al., 2016) in order (1) to locate active tectonic structures; (2) to characterize them using also validation through field studies; and (3) to quantify them using also various complementary geodetic measurements (e.g. Yu S.-B., et al., 1997; Lee et al., 2008; Hsu et al., 2009, 2018; Huang et al., 2010).
Local studied applications gives us keys information to better understand the LVF earthquake cycle and the active tectonic processes.

Area B - Board 213

Aquatic vegetation (macrophytes) is important component of the littoral zone, which provide multiple ecosystem services: high primary productivity and biomass, high biodiversity, substrates and shelters for large numbers of organisms, play a pivotal role in element cycling and comprise in a diet for other aquatic biota. Besides that, macrophytes reduce wave effect, sediment resuspension and coastal erosion. Due to these aspects and constant pressures from both aquatic and terrestrial ecosystems, the monitoring of vegetation in a littoral zone is important. The choice of methods for mapping of macrophytes is relatively high; however, there is a lack in understanding which set of them is the most suitable for different water bodies.
The use of unmanned aerial vehicles such as drones for environment research is fast growing with various applications including mapping of macrophytes. The biggest advantages of using drones are very good image quality and ability to get images during cloudy conditions and on demand. However, there are also many obstacles considering image acquisition and processing. Although, limited mapping aria by drones is another shortage of this method, it can be solved by use of satellite images and supplementary environmental data (e.g. bathymetry, sediment map) to some extent.
Therefore, in this study we tried to assess the applicability of both drone and satellite images for mapping of macrophyte distribution in inland and transitional waters. For transitional waters, a study area was the north-eastern coast of the Curonian lagoon. This part of lagoon is estuarine, where is a constant outflow of the riverine waters from the Nemunas river, while the northern part of the study area is often affected by water inflow from the Baltic sea. The second study area, Plateliai lake, was selected from inland waterbodies. For image acquisition in May-July and November 2018, Phantom 4 advanced was used together with smart GPS rover for more accurate orthophoto georeferencing and DroneDeploy app to make flight plans. For a creation of image mosaic, an open source program Open Drone Map was used. The cover of macrophytes was visually determined and automatically classified from the mosaics. These results were validated with in situ data. The Sentinel-2 MSI data was obtained for upscaling the distribution of macrophytes, where Normalized Difference Vegetation Index (NDVI) was calculated and validated with results from the drone images.
From the drone images, it was possible to get image quality as good as 0.007 m per pixel resolution flying at 10 m, while flying at 70 m it was possible to get a resolution of 0.01 m per pixel but larger area was covered. Perfect conditions for the flight were during completely cloudy days or sunny days with correct flight angles, times and polarized lenses. However, a sun glint was still visible in the images due to water waves. Several macrophyte species could be distinguished such as Patomageton perfoliatus, Cladophora glomerata and Phragmites australis and could be estimated their areas. The accuracy between orthophotos and in situ measurements was 94%. Other results will be presented and discussed in the poster.

Area B - Board 214

Currently, two main in situ methods are applied to monitor the water quality of a specific site: 1) Sailing out with a vessel and use it as a platform to perform several measurements (e.g. taking water samples) and 2) Deployment of one or more buoys with sensors at a specific depth below the water surface.

Although both methods provide useful data, they are often expensive and/or cumbersome resulting in sporadic data or data limited to few locations. Another approach is the use of Earth Observation satellites but the main drawbacks of this technique are the temporal resolution especially if high spatial detail is needed and strong dependency on cloud-free images.

The areas under consideration in water quality monitoring are often part of a dynamic environment subjected to important and continuous changes due to tide, current, waves, etc. Therefore, an easy-to-adopt and flexible method like drone-based images might be a useful addition to the range of existing techniques. They have some important advantages and are characterized by a rapid evolution from a technical point of view.

At the moment, drone surveys for land monitoring are already well developed and used in many industrial and environmental applications. The same approach cannot be extrapolated in a straightforward way to monitoring the water environment due to the continuous change of the scene and the specific characteristics of a water body (e.g. lower reflectance, glint effects or the lack of reference features). Therefore important modifications to the flight set-up and processing algorithms are required.

In this presentation, two case studies are discussed. One is about a Scottish lake and the other regarding a coastal area in the North Sea. The camera in use on both locations is a multispectral camera (MicaSense RedEdge M). The presentation will focus both on the data acquisition and data processing focusing on turbidity and Total Suspended Matter (TSM) retrieval. The results are validated with in situ measurements taken from several vessels. A comparison is done with appropriate satellite data like Sentinel-2.

Area B - Board 215

The Baltic Sea has a seasonal ice cover, which begins to form in October – November and the last pieces of
ice melts away by the beginning of June in the northern Bay of Bothnia. Maritime transportation around
the year is essential for the society – and there are many factories situated along the northern coast of the
Baltic Sea. Specially built ships - the icebreakers - are part of the Baltic winter navigation system that
enables merchant ships to visit northern ports also in wintertime.
The icebreakers get Sentinel-1 images in near-real-time and these images are visualized in their IBNet
information system. On a larger scale, the ice conditions can be assessed from the satellite images, but ice
drift and ambiguities in the interpretation are best resolved by airborne imagery. Previously, manned
helicopters were used for ice reconnaissance, but today these are considered too expensive to operate.
The assumption is that UAV:s can provide the icebreakers with up-to-date information on:
- position of navigable leads and registration of leads that are not usable anymore due to ice
convergence
- status of old leads
- ice drift (amount and velocity) to provide an update of the situation compared to the latest satellite
image
- finding best way through a ridged ice field
- inspection of nearby ships that report of problems or request aid
However, these needs have not been systematically documented nor evaluated. The capabilities of
modern UAV:s and instruments also have to be studied as well as processing needs to extract meaningful
information to the users helping them in their operations.
To shed some light upon these questions, VTT is contracted by the maritime authorities in Sweden and
Finland, to conduct a feasibility study where the use of UAV:s to support icebreaker operations, is studied
in detail. This study is done in winter 2019 (January -March). The methodology is based on interviews and
workshops. Also a simple proof-of-concept trial will be done as part of the study.
The poster will show the preliminary results of the study answering the following type of questions
User related: In which scenarios could a drone provide additional useful information - and for what
purpose? How should it be presented? Should it be combined with other information and
how? Should this information be shared? With whom?
Technical capabilities: What kind of sensors (e.g. cameras, ..) , what kind of UAV:s would be required - what
are the options and rough costs. What communication capabilities are required?
Operational aspects: Safety, operational arrangements, operator support, regulations
Roadmap: What is the long term roadmap including the suggested next steps regarding the utilization of
UAV:s in icebreaking operations?

Area B - Board 216

The European RPAS Steering Group (ERSG) received in 2013 by the European Commission the mandate to establish a Roadmap for the safe integration of civil RPAS into the European aviation system. The research study conducted up to now will finish briefly its activity, publishing the regulation by EASA; this will get under way the request of new services like airspace monitoring, territory monitoring, emergency and disaster monitoring, industrial services. To enable and catch this pushing request, the aerospace industry must enhance safety and efficacy performances of the RPAS, integrate them in the airspace and introduce new sensors and technologies for integrated data analysis.
The RPASinAir project is starting in November 2018 with a grant from the Italian Ministry of Scientific Research and University, and has been proposed by the DTA - Distretto Aerospaziale Pugliese (Aerospace District of Apulia), in partnership with Planetek srl, Leonardo spa, Enginsoft spa, Politecnico di Bari, CNR (Italian National Council of Research), ENEA (Italian National Agency for New Technologies, Energy and Sustainable Economic Development), Università di Bari, Vitrociset spa, ENAV spa, Aeroporti di Puglia spa, Università degli Studi di Enna KORE, Telespazio spa, and INFN (Italian National Institute for Nuclear Physics), and aims at developing knowledge and solutions in three technological fields strictly linked, enabling innovative services for the monitoring and control of the territory by RPAS and satellite Earth Observation:
1) design and planning of air mission, developing architecture and SW modules for the full-scale simulation of missions of one (or more) RPAS with 2d/3d visualisation within the air traffic;
2) Air Traffic Control, developing new ATM/C protocols, techniques, systems and interfaces to visualise the air traffic and communicate with the pilot on the ground;
3) Innovative sensors and technologies, and systems for data analysis, developing sensors for the monitoring of new risk factors (cryospheric, fire dust…), the relevant signal processing techniques, and a system for real and near real-time multisource analysis (air payloads, satellite data, in situ measurements), aiming at the identification of alert states.
The project belongs to the long term programme launched by the DTA aimed at providing the Italian aerospace system with an experimental airport to test new aeronautic products, particularly RPAS: the Grottaglie Airport Test Bed. This programme was accepted by the Italian Minister for Infrastructure (MIT), that in the National Plan for Airport (2013) declared the Grottaglie “airport of national interest”. In July 2014, ENAC (with directorial decree) authorized the use of the airport to test innovative aircraft (RPAS and General Aviation), successively ENAV designed the air space for experimentations. Since then, MIT coordinates the ‘technical committee for RPAS for Grottaglie airport’, that is participated also be Leonardo, ENAV, AdP (partner of this proposal). Several companies (Leonardo, IDS, Sistemi Dinamici, Blackshape) already exploited the Grottaglie airport for experimentation activities.
RPASinAir is to be considered part of a set of research projects devoted to the insertion of RPAS in the Air Traffic Management:
• DeSIRE 2 – fix wings (completed)
• CRUISE – cyber security (starting)
• PACIS-4 – rotary wings (in preparation)
• PRIME – service and support (in preparation)
• ASI call on “UAS/RPAS integrati nel sistema ATN Nazionale” (to be delivered in Grottaglie with AdP and DTA as partners

RPASinAir project is then not only a research and development initiative but is part of the several initiatives that the partners and institutions have planned and are implementing.
The solutions designed and implemented will converge into a demonstration, implemented as said in Apulia (both Grottaglie airport, and the INFN/UNIBA’s ReCaS infrastructure), including both the design and execution of various RPAS flights controlled by the ATM/C and the validation of the innovative sensors and of the multisource analysis systems.
The project concurs to various technologic fields in the aerospace sector: Advanced avionics in the field of HW modules network and Man-Machine Interface (MMI), Enhanced air traffic management system, Unmanned aerial vehicle (UAV) for civil use, EO systems, in the field of missions, instrumentations and data processing. The planning of missions with RPAS requires new simulation tools of the mission profile in the air space. The behaviour of the aircraft can be tested simulating both the avionic (on-board and ground) elements, devoted to the command, control and communication, and the air traffic control system. The simulator we intend to develop will allow 1) the study of the individual and systemic characteristics of the avionic modules in the different phases of the flight (e.g. the impact of the delay due to satellite communication), 2) the study of the impact on security of the satellite communication (Satcom) and satellite positioning services (GNSS), and 3) the study of the ground control stations (GCS) to verify their consistency with the workload expected for the operators. The reachability of the objective (the detection of the event) will be tested during the pre-flight phase, integrating in the network of the modules also the simulator of the sensors. Discrepancies between the planned and actual demonstration will be verified, individuating lack in the models, in the platform simulators and ATM/C systems. The simulation of the avionic system and the experimentation of the missions will allow the definition of the field of validity of the current solutions, identifying new technological needs.
New ATM systems that allow to manage in an integrated and homogeneous way both the manned and the unmanned aircrafts will be designed; in the frame of the SESAR project, it has been clarified that RPAS will not modify the regulation and standards of the piloted air traffic, but for the management of the air traffic of the RPAS just general guidelines have been developed (ICAO), while standards for the interaction among flight controller, ATM, pilot and aircraft do not exist. Among the goals of the project, there are then present 1) the exploration of the modalities, techniques, procedures and air control system of mixed air traffic and 2) the development and experimentation of an innovative ATM. In this case too, the MMI and the workload caused on the operator will be studied and measured.
The safe execution of RPAS missions is mainly finalised to the study and monitoring of events, phenomena or criticality of the territory, with the goal of assisting the operators in the monitoring (and control) of the events themselves. Air missions can have different goals: the identification of place and modalities of intervention in critical and catastrophic events, minimising the risks for the rescue team and enhancing the operational capability; detect short, medium and long periods phenomena relevant to the status of the land, water and air, to monitor and prevent the climate changes and environmental pollution; detect events on the ground to activate process and procedure for its control, e.g. in precision farming.
To understand the needs for such services, it is to note that climate changes and anthropic impact are enhancing the frequency and magnitude of the flooding, landslides and coastal swells. In Italy, especially in the southern regions, natural disasters are causing damages more and more evident to people, assets and cultural heritage. The precision monitoring of such hydrogeological phenomena and the evaluation of the relevant risk holds an important and crucial role in all the phases of the management of the disaster, from the alert to the emergency phase and civil protection, up to the assessment of effects and damages, for compensation and risk reduction. Among the risks, the project will take into account the sectors for which the Italian Civil Protection has established the Commissione Grandi Rischi (Greater Risks Commission). The results of the projects have the goal to identify radar devices capable of deep prospection on snow cover and ice sheets for stratigraphic studies and their real-time modifications, and then redesign such devices in order to be airborne and be able to monitor vast territories in a reasonable timing (even daily).
This requires the definition of a spectrum of decision support functionalities, based on the capability of integrate processing of massive and heterogeneous data fluxes, coming from several sources on the ground (in-situ sensors), in air (airborne RPAS sensors), and satellite (e.g. Copernicus). To properly address this Big Data problem, not only hi-performance and reliable infrastructure like ReCaS, but also methods and process based on the Artificial Intelligence strongly innovative and capable to combine effectiveness, flexibility and extensibility. The innovative technologies proposed and studied in the frame of the project will make a start to further engineering and industrialisation processes, up to the convergence towards interoperable protocols to promote the economy of scale and the market entry of new players and ideas, capable of expanding the type, improve the quality and reduce the cost of the services.

Area B - Board 217

Since 2012 more actively Ligther Than Air platforms are being developed for mointoring and reconnaisance purposess, both as low altitude platforms, as well as HAPS. What is the state of the art of these developments in Europe (Germany, France, Belgium)? What are the key impediments? From an overview of the most progressed intiatives, a listing of challenges to overcome will be described as well as possible strategies observed to overcome these challenges. Furthermore, aslo will investigated if the low(er) altitude intiatives (e.g. Airbus Altair programme), Flywin Project (Belgium) also take into accoutn possible synergies with HAPS under development (both lighter than air as the initatives relying on convential aviation).

Area B - Board 218

It’s easy to develop Air vehicles or satellites for each customer and its own requirements to fulfill the mission(s) from technical perspective. But the market for civil science and business missions will double in the next years. This growing customized market can supported by aircraft manufacturers who are able to manage projects under aeronautic conditions and extend their product portfolio by HAPS air vehicles. This kind of vehicles could be re-engineered, using existing aircrafts with modified lift- and gliding performances and with standard cutouts and flexible rack systems for the mission payload and for unmanned flight controls.
Project manager for platform development and system integration are in the position for typical customized platforms, with the result, that the diversity of engineering solution and configuration grow up rapidly. Pseudo satellites vehicles are categorized as flight vehicles and have to be approved by the aviation authorities (e.g. EASA and FAA). That means the administrative work and these costs for approving each changes increase fast. To stay with calculable costs for the future is mandatory for civil science projects and all other business opportunities.
To solve present situation and move to modular solutions or engineering concepts is a key to keep this market attractive and spread for more popular and business applications. There are a number of aircraft manufacturer able with developing and production skills at the market to modify an existing aircraft to a High Altitude Vehicle and their also a number of companies existing for unmanned flight control systems.
Intention is to open the mind for serial (for sure small serial) development and production for HAPS vehicles solutions with common standards, to learn from big aircraft manufacturer and automotive industry about standardization and handling with growing market opportunities. I know the difficulties from small manufacturer with good product basis and the huge requirement coming from the market to steer the turn-around into the new course with the experience from civil aviation and automotive industry project management.

Area B - Board 219