IEEE Geoscience and Remote Sensing Magazine - September 2016 - 42

(a)

(b)

FIGURE 1. (a) The iSPEX device and (b) a demonstration of its use.

FIGURE 2. The Cropland Capture game in which players were

asked whether they could see evidence of cropland in images and
photographs.

validation of EO data sets. In the "Case Studies" section,
three case studies, based on our experiences, are provided
that illustrate different ways in which CS and crowdsourcing have been used with EO. The main challenge is scaling
up these efforts to collect larger quantities of data and to
expand the spatial outreach.
CASE STUDIES
THE iSPEX PROJECT
Aerosol measurements from professional ground-based stations
suffer from a lack of spatial coverage. Satellite measurements provide spatial coverage but an insufficient temporal
resolution, often providing only one measurement per day. The
iSPEX project is an example of a large-scale CS project that was
42

developed as a way to collect aerosol data at a much higher
spatiotemporal resolution. An inexpensive add on to a smartphone that uses a multitude of built-in phone functionalities
and a simple protocol for taking the measurements has enabled wide-scale citizen participation in providing crucial
information about atmospheric particulates that affect air
quality and influence our climate (Figure 1).
In 2013, experiments were run on three separate days
in The Netherlands where more than 10,000 measurements were collected. The quality of the measurements
when averaged over many devices at the same location
was extremely high, proving that measurements from an
army of iSPEX observers can complement professional
measurements from the ground [30]. The iSPEX measurements were also collected across The Netherlands
and compared with data from the moderate resolution
imaging spectroradiometer (MODIS), showing good correspondence in the spatial patterns. Thus, large amounts
of measurements with a sufficient resolution in time and
space can provide data that are complementary to products derived from EO. Moreover, the upward-looking
iSPEX measurements can be used to calibrate and validate satellite products that need complex corrections for
ground scenes.
Future plans include European-wide experiments in
aerosol data collection via the iSPEX add on and further citizen involvement. The enthusiasm of the Dutch
volunteers in taking measurements has been one of the
main factors in the success of this CS project. Motivating larger crowds on a European-wide scale will be a
future challenge.
THE GEO-WIKI PROJECT
Geo-Wiki was established in 2009 as a tool for involving citizens in determining whether global land-cover maps accurately
characterize the earth's surface based on Google Earth imagery
[31]. Since then, there have been a number of crowdsourcing
competitions that have collected approximately 250,000 pixels of land-cover classifications, which have been used to both
train and validate land-cover maps (e.g., [32], [33]). More recently, the Geo-Wiki game Cropland Capture (Figure 2) resulted
in the collection of around 4.5 million pixels from 3,000 players over a six-month period [34]. These data were used to develop a hybrid global cropland map for 2010.
As part of the ESA-funded Education for EO Capitalizing on a Citizen Science Approach (EducEO) project,
a new game, Picture Pile, has been developed that uses
many of the same game mechanics as Cropland Capture.
This game broadens image classification to any type of
land cover organized into different piles of pictures for
interpretation (Figure 3). Picture Pile has the potential to
become a new generic tool in the ESA toolbox that will
be devoted to validating Sentinel and other imagery. The
idea is to develop a large community of citizen sentinel truthers, similar to the communities that have been
built up by the Zooniverse project for the classification
ieee Geoscience and remote sensinG maGazine

september 2016

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - September 2016