metadata (i.e., classification, valid time, characteristics, and position).
This meta-information of heterogeneous sensors can be obtained using the DescribeSensor interface of SOS. Therefore, clients
on the web can access this sensor and its data through its SOS
interface and by using these important meta-information.
Figs. 8 and 9 show a GetObservation request and response sample.
After receiving the data request from the client, the SOS uses
spatial–temporal queries inside the PostgreSQL database. If the
request is valid, the SOS returns data accordingly. Fig. 9 shows that
the sensor data was encoded in O&M format as the request
indicated.
Subsequently, a test client was implemented to retrieve and
visualize different real-time and historical agro-meteorological
measurements in this field using the primary monitoring pattern.
Fig. 10 shows that this client continuously retrieves real-time sensor
data from the SOS and displays the data on the left. These
observations are refreshed at a certain frequency, i.e., every 10 s.
On the right, the Tianditu (Map World), which is the public version
of the national platform for common geospatial information service
of China, was used. Tianditu provides JavaScript-based web
APIs to construct a geospatial client quickly. In our experiment,