Citizen Science (continued)

The ever-growing circulation of multi-media-capable smartphones creates the opportunity for a “dense network of observers”, enabling collection of a wide variety of data at minimum cost. Scientists can take advantage of this new source of data by involving people willing to contribute requiring them minor efforts while still providing potentially valuable data.

The concept introduced in this paper is related to an app designed for automated integration and validation of land cover layers, especially those provided in the Copernicus framework, with crowdsourced data. The final aim of the app is to sustain a live-updated map based on the submitted reports, made available to anyone in the form of open data. Information extracted from the collected points could indeed merge into the data flow behind Copernicus Land Cover layers (e.g. CORINE [1]) and enable validation of automated production results from Sentinel imagery. Users will be asked to submit a geocoded picture taken on the spot through their smartphone, and to select the observed land cover, from a proposed list, enriched with additional features, also picked from a context-sensitive option list. Each newly registered user will go through a brief tutorial to ensure he/she understands the procedure. Then each selection will be guided through pre-determined example pictures detailing land cover hypotheses and possible additional features.

The idea turned out to be sufficiently convincing to sit among the winners of the third MyGEOSS contest [2], an open call for innovative ideas related to changes affecting local environments. The proposed mobile app can fulfil the intent thanks to the active involvement of European citizens in monitoring land cover changes. The app will be developed following the open-source framework designed by our group for the distributed collection of reports [3], e.g. crop-related data [4][5].

A beta version of the app will be presented at the EOScience 2.0 conference with the aim to collect feedback and include improvements for the final version.

References

[1] CORINE Land Cover, available online at: http://www.eea.europa.eu/publications/COR0-landcover, accessed on 20/05/2016.

[2] MyGEOSS call, available online at: http://digitalearthlab.jrc.ec.europa.eu/mygeoss/info_thirdcall.cfm, accessed  on 20/05/2016.

[3] D.A. Galeazzo, D. De Vecchi, F. Dell’Acqua, P. Demattei, “A small step towards citizen sensor: a multi-purpose framework for mobile apps”, International Geoscience and Remote Sensing Symposium IGARSS, 26-31 July, Milan, Italy.

[4] D.A. Galeazzo, D. De Vecchi, F. Dell’Acqua, “Citizens as Sensors: from a multi-purpose framework to app implementation”, Earth Observation Open Science 2.0, 12-14 October 2015, Frascati, Italy.

[5] D. De Vecchi, F. Dell’Acqua, “Citizens as source of “fresh” information: a mobile app for updated vegetation status”, ESA Living Planet symposium, 9-13 May 2016, Prague, Czech Republic.

The “Citizen Scientists” (CS) concept is one of the emerging fields recognized as a key element to increase EO capabilities. This technology and culture can be used to enable communities of citizens to provide essential information for deeper analysis of remotely sensed information [1]. However, in such a context, metrics and processes for data quality assessments are issues of crucial importance to document the usefulness of the results that can be produced.

In this work a CS project addressing accurate classification of Very High Resolution (VHR) optical imagery, is presented. Besides the data, the “scientists”  are provided a neural network based toolbox for image processing plus additional scripts to test the accuracy of the obtained results.The considered land cover classes are: buildings, asphalted areas, natural, water, bare soil, vegetation. Each tile is assigned to the “scientist” who mainly has to perform two tasks: image classification and accuracy evaluation.  The automatic image classification is performed using the Neumapper Toolbox which is a freely distributed package enabling image classification via neural networks [2]. Neural networks have great capabilities as a pattern recognition method for multi-source remotely sensed data because of the parallel nature of the processing. In particular, it has been shown that multilayer perceptrons (MLP) maybe an efficient alternative to conventional statistical approaches for automating image classification [3],[4]. With Neumapper the CS can easily manage the whole processing in a unique user friendly software environment. Moreover, a few scripts in python language have been developed implementing the following tasks: generation of a given number of validation points randomly distributed over the image, localization of the generated points in Google Earth, realization of a confusion matrix based on the generated (manually labeled) points and the classified image. An assessment of the global result can be carried out by means of an external validation. The overall accuracy so far is around 95%. After completing the classification of Rome, new Italian cities are currently being processed. A web platform has been set up to let the users interact remotely by selecting the tile they want to elaborate and by uploading the final results. The platform is continuosly updated in terms of data and functionalities, in particular, a service to support the participants in their operations has been also implemented.The presentation will show how the platforms works and the main results obtained so far.

[1] Goochild, M.F., “Citizen as sensors: the world of volunteered geography,” GeoJournal, 69:211-221, 2007

[2] Del Frate, F., I. Fabrini, M. Penalver, M. Iapaolo, “NEUMAPPER: a Neural Networks Software for Image Classification,” Proc. of the ESA-EUSC-JRC 2011 Image Information Mining conference, ISPRA (VA), Italy, 30-31 March, 2011

[3] Pacifici, F., F. Del Frate, W. J. Emery, P. Gamba, J. Chanussot, “Urban mapping using coarse SAR and optical data: outcome of the2007 GRS-S data fusion contest,” IEEE Geoscience and Remote Sensing Letters, vol 5, n. 3, pp. 331-335, July 2008

[4] Del Frate, F., F. Pacifici, G. Schiavon, C. Solimini, “Use of neural networks for automatic classification from high resolution imagery,” IEEE Transactions on Geoscience and Remote Sensing, vol. 45, n. 4, pp. 800-809, April 2007.

sTREEt HEALTH project funded by ESA aims at developing a full service that will provide geolocated Tree Health Indicators to support tree management in cities to municipalities and other target markets such as real state, construction or even retail market. The trial service will be deployed within the STREEt HEALTH project in Barcelona city from June 2016 to May 2017

The information needed to produce such products would be extracted and processed from high-resolution optical images from the Pleiades in combination with Sentinel data. The service would then benefit from new EO sensors capabilities with a resolution that correspond to the size of the targeted objects: single urban trees.

 Additionally, with a daily repeat cycle, the PlanetLabs products will be analyzed in order to improve the current concept based on Pleiades and Sentinel images, providing auxiliary information (seasonal trends) to better estimate the risks.

The service will be deployed and run in a cloud environment: Amazon Web Service. Both data processing and data access will be supported by a scalable cloud system that will allow us to process very high- resolution data covering entire cities and surroundings. The possibilities offered nowadays by this service include a complete informatic infrastructure for EO product generation and accessibility.  Moreover, the interface to end-users will be designed using an emerging visualization service enhanced by EO product providers: CartoDB API.

Starlab counts with some great experts on Machine Learning (ML) due to its other department; neuroscience.  Lately this knowledge is being applied to the processing of space data mixed up with the Big Data coming from OpenData source. The possibility of doing a fully automated inventory of tree species in cities where there is no tree data recorded is just an example of the capabilities that ML together with EO and OpenData will bring to the service.

The use of geospatial data is crucial for effective disaster risk reduction (DRR), disaster management and climate change adaption. The application of freely available geospatial data and freeware is particularly useful for sustainable development and reducing the global digital divide. Free geoinformatics can help to optimize the limited financial, technological and manpower resources that many organisations face, providing a sustainable input to analytical activities. The freeware and open source software (FOSS) applications examined here range from Geographical Information Systems, to the processing of remotely sensed images, crowd-source mapping, web applications and content management systems. Free geoinformatics can help to optimize the limited financial, technological and manpower resources that many organisations face, providing a sustainable input to analytical activities.

Digital information technologies, such as the Internet, mobile phones and social media, provide vast amounts of data for decision-making and resource management. We present a time series analysis (2009-2014) for a modern evaluation of the “Digital Divide” concept that originated in the 1990’s. The Data Poverty Index (DPI) provides an open-source means of annually evaluating global access to data and information. The DPI can be used to monitor aspects of digital data and information availability at global and national levels, with potential application at local (district) levels. Of particular interest are digital datasets that are of use for disaster risk reduction, early warnings and improved preparedness, increased resilience and enhanced adaptation to climate change. In that context, the DPI could be a useful tool for monitoring the Sustainable Development Goals of the Sendai Framework for Disaster Risk Reduction (2015-2030). The effects of severe data poverty, particularly limited access to geoinformatic data, free software and online training materials, are discussed in the context of sustainable development and disaster risk reduction. Examples of free geoinformatic applications from developing countries are presented: for the management of water resources and flood risk (Sierra Leone), and for coastal flood risk assessment (Sri Lanka).