Geoscience Reference
In-Depth Information
A good example of an app in which this particular type of operation has been implemented is
LillyPad (Rogers et al., 2010); it was designed to allow users to record the state of trees in an
urban riverside regeneration environment. LillyPad aided researchers in recording information
about the trees; they could also access past work, and crunch statistics on the fly, in order to bet-
ter understand the state of individual trees in the context of previous years. End users, equipped
with a comprehensive suite of available records, could indeed more fully appreciate if such trees
were better or worse off than in the previous year and perhaps speculate as to why that might be.
Moreover, given observations of many trees, such researchers could also gain a better idea of the
overall pattern involved.
Automatic execution of a service for a user could be set up as simple triggers in an LBGC app.
For example, in the planning phase of a field exercise, there might be some parts of the field site
that look like they may contain certain points of interest (POI). The
triggering
of the device at this
point to prompt the user to perform an action could be used to help tie the background analysis with
data collection in the field. This type of triggering has been seen in media consumption experiences
such as mobile tour guides (Priestnall et al., 2009) and augmented reality (AR) experiences, such as
Layar (2013) and Wikitude (2013).
Perhaps the most important feature of context-aware apps that apply to LBGC is the ability
to tag information for later retrieval. We take this point of
later retrieval
to mean any time after
the collection of a piece of data, which may be in the field or later in a different location. An
app for doing this has been described as a method for picking out POI in the field with a view to
reconstructing glacial models in AR (Meek and Priestnall, 2011). Here, the smartphone running
an app called
Zapp
was used as a
point-and-capture
mobile device for picking out points in the
field from a distance using an app in which the user looks through the device's camera preview
with a crosshair augmented over the top. The user then aligns the crosshair with the point that
they wish to capture and presses a button which activates a vector line-of-sight algorithm (Fisher,
1996) using the device's sensor as inputs and underlying raster data to calculate the raster cell
being looked at. Additionally, Zapp has been used as a platform for consuming media in the field,
through running the same LOS algorithm several times per second and using a look up raster for
each cell with an underlying index of POIs for POI identification (Meek et al., 2013). This approach
could be combined with other features of context-aware apps to provide a method for performing
LBGC by directly tying together the data collection phase with the landscape. The use of AR in
this instance encourages the user to evaluate the piece of data collected in situ and offers insight
into any shortcomings of the underlying data itself. Methodologically, there are two categories of
current approaches to filtering information on location that mirrors types of data available to GIS.
These are as follows:
1. Ve ctor
Vector-based approaches are discussed at length in Yin and Carswell (2012) and include
both 2D and 3D querying information imbedded in an urban environment. Due to the cur-
rent availability of geodatabases designed for mobile devices, these approaches require a
client-server relationship to process a query and produce results. In these relationships, the
mobile device acts like a data-gathering tool for sensor information using sensor readings
to determine where the user is and what they are pointing the device at. There are different
approaches to this method of querying that offer advantages of high accuracy and the abil-
ity to report not only what building is being pointed at but also where on the building this
is being done.
2. Raster
This approach offers advantages as all of the calculations can be carried out on the
device and does not require a client-server relationship. This is useful in rural or upland
areas that often do not have access to a reliable 3G connection. This approach also offers
greater flexibility since it can employ many resolutions of base data; indeed, whilst vector
