issues in high data transmission environments (Wang et al. 2010). Private clouds are
built by enterprises to fill this gap. Some open source architectures like CloudStack
(Kumar et al. 2014) and OpenNebula (Milojičić et al. 2011) could be reasonable
choices. Enterprises have full control of the network and storage of private clouds so
that they can apply any suitable security mechanisms upon it. The disadvantage is
that the cost is higher than public clouds.
Either public or private clouds serve an efficient way to host big data which is
impossible on single servers or PCs. Clouds can avoid many annoying problems like
server malfunction, disk failure, memory leak, I/O slowness, high-cost maintenance,
etc. The data are uploaded and downloaded via the Internet. The nodes in the cloud
physically possess the data. Users can assess them by the management console of
clouds. The downloading is mostly through the HTTP and FTP protocols. The speed
depends on the network of cloud nodes and client devices. However, the search
button for the datasets in clouds is not as mature as the storage. The native searching
function of cloud platform is imperfect. In operational systems of datasets, a
separated register service which archives all the metadata for searching is often
established. When users find a product in the register, the individual link to the data
in cloud host will be returned to the users.
4.3.3 Crowdsourcing
Crowdsourcing is a new sourcing model in which individuals or organizations obtain
the needed data and services from the Internet contributors (Doan et al. 2011). It
forms distributed labor networks to exploit the spare processing power of millions of
the human brains via the Internet (Howe 2006). Crowdsourcing is a rising strategy
for data collection (Hirafuji 2014; Kanhere 2011). It is producing geospatial data
using informal social networks and state-of-the-art Web technologies (Heipke 2010).
The potential user groups collaborate voluntarily with very little monetary support to
make the result datasets free online. Key differences are that the contributors to
crowdsourcing datasets may lack formal training on collecting and organizing the
data. The citizens could be involved via their smartphones equipped with low-cost
GPS receiver, accelerometer, ambient temperature sensor, gyroscope, light sensor,
magnetometer, barometer, proximity sensor, humidity sensor, audio sensor, fingerprint identity sensor, moisture sensor, and camera.
In the agro-geoinformatcs, crowdsourcing has been used to collect data. For
example, Geo-Wiki, a crowdsourcing tool, is developed to collect in situ data to
improve the global land cover products (Fritz et al. 2012). The web-based system
integrates access to high-resolution satellite imagery from Google Earth with
crowdsourcing to vastly increase the available amount of information on the land
cover. The information can be used for training, cross-checking, calibration, and
validation of the land cover products (See et al. 2015). In addition, a smartphone app
is made to intuitively retrieve the exact geometry of smaller objects to access
agricultural entities like fields or ponds (Frommberger et al. 2013). Diseased leaf
60
Z. Sun et al.
built by enterprises to fill this gap. Some open source architectures like CloudStack
(Kumar et al. 2014) and OpenNebula (Milojičić et al. 2011) could be reasonable
choices. Enterprises have full control of the network and storage of private clouds so
that they can apply any suitable security mechanisms upon it. The disadvantage is
that the cost is higher than public clouds.
Either public or private clouds serve an efficient way to host big data which is
impossible on single servers or PCs. Clouds can avoid many annoying problems like
server malfunction, disk failure, memory leak, I/O slowness, high-cost maintenance,
etc. The data are uploaded and downloaded via the Internet. The nodes in the cloud
physically possess the data. Users can assess them by the management console of
clouds. The downloading is mostly through the HTTP and FTP protocols. The speed
depends on the network of cloud nodes and client devices. However, the search
button for the datasets in clouds is not as mature as the storage. The native searching
function of cloud platform is imperfect. In operational systems of datasets, a
separated register service which archives all the metadata for searching is often
established. When users find a product in the register, the individual link to the data
in cloud host will be returned to the users.
4.3.3 Crowdsourcing
Crowdsourcing is a new sourcing model in which individuals or organizations obtain
the needed data and services from the Internet contributors (Doan et al. 2011). It
forms distributed labor networks to exploit the spare processing power of millions of
the human brains via the Internet (Howe 2006). Crowdsourcing is a rising strategy
for data collection (Hirafuji 2014; Kanhere 2011). It is producing geospatial data
using informal social networks and state-of-the-art Web technologies (Heipke 2010).
The potential user groups collaborate voluntarily with very little monetary support to
make the result datasets free online. Key differences are that the contributors to
crowdsourcing datasets may lack formal training on collecting and organizing the
data. The citizens could be involved via their smartphones equipped with low-cost
GPS receiver, accelerometer, ambient temperature sensor, gyroscope, light sensor,
magnetometer, barometer, proximity sensor, humidity sensor, audio sensor, fingerprint identity sensor, moisture sensor, and camera.
In the agro-geoinformatcs, crowdsourcing has been used to collect data. For
example, Geo-Wiki, a crowdsourcing tool, is developed to collect in situ data to
improve the global land cover products (Fritz et al. 2012). The web-based system
integrates access to high-resolution satellite imagery from Google Earth with
crowdsourcing to vastly increase the available amount of information on the land
cover. The information can be used for training, cross-checking, calibration, and
validation of the land cover products (See et al. 2015). In addition, a smartphone app
is made to intuitively retrieve the exact geometry of smaller objects to access
agricultural entities like fields or ponds (Frommberger et al. 2013). Diseased leaf
60
Z. Sun et al.