Geography Reference
In-Depth Information
but unknown coordinate system. These data were the latest and most complete
information we could obtain. Obviously, these data can cover most of the trip
purposes we need to classify, including home, work, school, shopping, recreation,
personal affairs and welfare/religion activities. Among these seven trip purposes,
the first two categories can be identified according to participants' basic data
filled in the questionnaire; schools are comprised of elementary and secondary
school, universities and colleges; shopping places consist of markets and shopping
malls; public places of recreation refer to restaurants, sports courts and parks;
facilities for personal affairs includes post offices, banks, governments and medical
establishments, while the last category refers to welfare and religion institutions.
13.5
Detecting Activity Types and Trip Purposes
13.5.1
Data Preprocessing
13.5.1.1
Adjusting All Data into the Same Coordinate System
As mentioned earlier, some of the POI layers are referenced in a coordinate system
different from the WGS-84 system in which the GPS tracking data and other GIS
data are referenced. We first need to transform these POI layers to the WGS-
84 coordinate system so that a composite map that includes the base map of
Guangzhou boundary, land use layers and GPS track points in the WGS-84 system
are developed for further analysis. This transformation is accomplished through
spatial adjustment in ArcMap by linking and comparing a number of control points
having the same relative positions in the source and destination layers so as to derive
the transformation formulas from one coordinate system to another. To minimize the
correction errors, a total of eight control points scattered at different corners of
the urban area were chosen. The residual errors associated with each link and the
overall Root Mean Square (RMS) error were estimated and the results indicate that
the transformation meets the accuracy requirement. After the establishment of the
transformation formulas, the feature layers can be adjusted accordingly.
13.5.1.2
Preprocessing GPS Tracking Data
The original GPS tracking data contains only the spatio-temporal trajectories of
objects, including latitude and longitude coordinates, date, time, speed information,
temperature as well as some indicators concerning the status of the device. The
original GPS data may contain some noise data (tracking points that go beyond
the boundary of the study area or drift far away from other points) resulting from
many reasons including the lack of sufficient number of visible satellites because of
urban canyon effects. These noise data were first removed through a data filtering
process. Information on HDOP value, the number of visible satellites, speed and
Search WWH ::
Custom Search