Prologue
10 February 2012
This version : http://observedchange.com/sego/ns-10022012
Latest version : http://observedchange.com/sego/ns
Revision : 3.0
Author : Anusuriya Devaraju, Institute for Geoinformatics, University of Muenster.
Copyright @ 2008 - 2012 the author above
This work is licensed under a Creative Commons License. This copyright applies to the SEGO Ontology Specification and accompanying documentation. |
Note : The visual layout and structure of the specification was adapted from Parrot, a RIF and OWL documentation service and the TopBraid Composer.
Abstract
SEGO, the Sensing Geographic Occurrences Ontology provides a formal vocabulary to describe the relation between geographic occurrences and properties observed by sensors. This documents specifies the classes and properties introduced by the ontology. One may regard the proposed ontology as a component of the larger sensing domain ontology.
Status of this document
NOTE: This section describes the status of this document at the time of its publication. Other documents may supersede this document.
This specification is an evolving document. This document may be updated or added to based on implementation experience, but no commitment is made by the authors regarding future updates.
The author welcomes suggestions on the SEGO Vocabulary and this document. Please send comments to anusuriya.devaraju@uni-muenster.de
Introduction
Observations are fed into the Sensor Web through a growing number of environmental sensors, including technical and human observers. While a wealth of observations is now accessible, there is still a gap between low-level observations and the high-level descriptive information they reflect. The challenge is not to gather a vast number of observations, but rather to make sense of them in environmental monitoring and decision making. In order to infer meaningful information about geographic occurrences from observations, a description of how one gets from the former to information about the latter must be expressed. SEGO is developed to formally capture the relationships between geographic occurrences and the properties observed by sensors.
The ontology is anchored in the Descriptive Ontology for Linguistic and Cognitive Engineering (DOLCE ) (see Figure 1). It also incorporates the SWRL temporal ontology to express the temporal information associated with observations and geographic occurrences.
Figure 1. Aligning SEGO to DOLCE.The prefix dul indicates DOLCE basic categories; the prefix sego denotes domain categories; and the prefix tm specifies categories from the SWRL temporal ontology. Dotted arrows describe datatype properties.
The following Figure 2 provides an overview of the classes and properties specified in the SEGO Ontology. A sensor is regarded as an interface between the physical world and the information world. It responds to stimuli (e.g., geographic processes), and thereby allows the sensing of properties of a particular feature-of-interest. An observation-event is the central notion between categories in the developed ontology. A geo-event is related to the observation domain via its participants bearing certain observed properties. For an in situ sensing, the location of a geo-event coincides with its observation-ground.
Figure 2. An overview of SEGO ontology.
Example use of SEGO
A use case for reasoning about blizzards and their temporal parts from time series supplied by the Environment Canada illustrates the application of SEGO. Further details are available here.
The XML Namespace URIs that must be used by implementations of this specification are:
- http://www.loa-cnr.it/ontologies/DOLCE-Lite.owl# - DOLCE Foundational Ontology Namespace
- http://www.anusuriya.com/sego/SEGOv3.owl# - SEGO Core Ontology Namespace
- http://swrl.stanford.edu/ontologies/built-ins/3.3/temporal.owl# - SWRL Temporal Ontology Namespace
OWL file for local loading : SEGO.owl
Classes
feature-of-interest, geo-process, geo-stimulus, geographic-event, observation-event, observation-ground, observation-result, observation-time, observed property, sensor
Properties:
actuates, affects, bearer of, covers, facilitates, has-bearer, has-foi, has-obs-event, has-obs-ground, has-obs-property, has-obs-result, has-observation-value, has-physical-location, has-temporal-sub-event, hinders, initiates,obs-result-of, observed-by, observes, occurs-at, performs, perpetuates, has-physical-location, produces, temporal-sub-event-of, temporally-made-of, terminates
A short glimpse at the foundatonal ontology DOLCE
The top categories of DOLCE are endurant, perdurant, quality and abstract (Masolo et al. (2003)). Endurants exists as wholes at any time they are present. At different times the same endurant may lose or acquire new parts. Examples of these are geographic features such as forest, sediment, land parcel, and river. Perdurants extend over time; they consists of temporal parts succeeding one another. Thus, at any time at which they exist they are only partially present. Some of their temporal parts (e.g., their previous or future phases) may not be present. Examples of these are natural occurrences such as precipitation, volcanic activity, hurricanes, and soil erosion. Qualities are temporal or physical properties we perceive or measure (e.g., the water level and the precipitation duration).
Description
|
Example
|
|
physical-endurant |
|
|
physical-object |
|
|
amount-of-matter |
|
|
feature |
|
|
non-physical-endurant |
|
|
perdurant |
|
|
event |
|
|
stative |
|
|
physical- quality |
|
|
temporal-quality |
|
|
feature-of-interest |
Features-of-interest (foi) are conceptualized as real physical-objects or features as defined in the DOLCE, regarding which an observation is made. The observation can be past as well as scheduled observations. Another constraint is that a foi shall be an identifiable entity from the application domain. It can be an object itself (e.g., a river), a part of the object, (e.g., a branch of a river system), or features as defined in the DOLCE (e.g., a gulf and a cross-section of a river) that can be recognized, observed and re-identified over time. We do not regard an amount-of-matter as a feature-of-interest, but rather as a constituent of the latter. The premise is that amounts do not seem to be directly perceivable, since we cannot identify portions of matter as such.
rdf:type : owl:Class
rdfs:subClassOf
geo-process |
A geographic process satiesfies the following : (a) cumulative, (b) anti-atomic, (c) temporal shape: a geographic process goes on for at least some consecutive amount of time, (d) emphasis - a geographic process represents the temporal distribution of an occurrence, it lays emphasis on the mechanism or the path of becoming, (d) the relation to a sensor : a geographic process plays the role as a geo-stimulus by actuating a sensor; a sensor responds to a geographic process.
rdf:type : owl:Class
rdfs:subClassOf
-
dul:stative dul:has-participant some dul:endurant dul:has-quality only dul:temporal-quality
geo-stimulus |
A stimulus is a geographic process that actuates a sensor. A stimulus is a geographic process that actuates a sensor. Depending on the sensor configuration, a sensor may respond to one or more stimuli. A simple example is an ongoing rainfall that triggers a tipping bucket rain gauge. A complex example refers to a combination of processes that actuate a sensor. Consider, for instance, a lysimeter that estimates a water loss from a plant-covered soil. The sensor involves water inflow (e.g., irrigation) and water outflow (e.g., water percolation) as its stimuli. Not all geographic processes are necessarily stimuli. In some cases, processes of interest may be causally linked to a stimulus actuating a sensor. In this case, they indirectly actuate a sensor. For example, a sailor (a human sensor) classifying the wind strength (an observed property) based on the appearance of wind effects, such as the ripples or the waves on the sea. Stimuli of a certain type may be detected by different sensors. For instance, the presence of smoke in a building can be detected by a smoke detector or a human in a building.
rdf:type : owl:Class
rdfs:subClassOf
-
geo-process -
actuates some sensor
geo-event |
A geographic event satisfies the following : (a) anti-cumulative, (b) anti-atomic, (c) temporal shape : a geographic event is a demarcated episode, which takes place in a specific interval, (d) Emphasis : a geographic event emphasizes a summary of what has happened, (e) the relation to a sensor : a geographic event is a notable occurrence that is inferred from observations. It can be identified on the basis of conditions expressed in terms of spatial, temporal and thematic properties. Note : (a), (b), (c), (d) support (e). A feature-of-interest is a participant of an inferred geographic-event.
rdf:type : owl:Class
rdfs:subClassOf
-
dul:accomplishment has-observation-event only observation-event has-participant some endurant has-quality only temporal-quality
observation-event |
An act of observing a property of an observation target, with the goal of producing an estimate value of the property. Here, we allow for actual as well as scheduled observations. Since observation-event is a sub-class of accomplishment, therefore it is anti-cumulative (may have temporal sub-parts). For example, temporal parts of a daily water quality observation indicated by different sampling times. The observed properties, results, temporal details are associated with an observation event, not with a sensor. An observation-event may or may not be associated with a result or an observation-ground. For example, the irrigation water supply is scheduled by observing the changes in plant characteristics and observations performed by weather buoys.
rdf:type : owl:Class
rdfs:subClassOf
-
dul:accomplishment -
dul:has-t-quality some observation-time -
has-obs-property some observed-property -
has-obs-ground only observation-ground -
has-foi some feature-of-interest performed-by some sensor -
produces only observation-result
observation-ground |
An observation-ground refers to a piece of land (a site) where observations are assumed to be valid. Its spatial extent is defined empirically. For a ground observation, the sensor is usually deployed on the observation-ground, e.g., a weather station. This excludes the cases where the target of an observation is located underneath the ground, for example, in groundwater quality monitoring. The concrete value of an observation-ground can be specified via the datatype property has-physical-location. This property can be associated with certain datatypes, such as user-defined or XSD built-in. For an in-situ ground-based sensing, an observation-ground is conceptualized as (a) a background space in which a feature of interest is located; and (b) the location of an inferred event. It is understood that not all observation grounds have sharply determinable spatial location. For a station making surface synoptic observation, the dimensions of this region or the representative areas may range from 2 000 km2 to 10,000 km2 for a plane or homogeneous relief.
rdf:type : owl:Class
rdfs:subClassOf
observation-result |
Observation-result refers socially constructed abstract objects (i.e., information-object) produced by an observation-event. A result can range from numerical measurements (e.g., time-series), categorical measurements (e.g., human weather observations such as mild, windy and rainy), images (e.g., aerial photographs) and so on. An observation-result is related to its concrete data values through the has-observation-value datatype property.
rdf:type : owl:Class
rdfs:subClassOf
observation-time |
Two temporal information are considered. First, the time of an observation-event. Second, the temporal information of a geographic-event which is inferred from sensor observations. To represent these temporal information, the temporal quality observation-time is modelled as subclass of the category extended-proposition from the SWRL temporal ontology.
rdf:type : owl:Class
rdfs:subClassOf
observed-property |
In DOLCE, every entity comes with certain qualities (properties), which exist as long as the entity exists. A quality cannot inhere in two
different entities. Qualities can be observable qualities or non-sensorial qualities. An observed-property is a physical quality inheres-in a feature-of-interest. For example, the temperature, the dissolved oxygen and the water level of a water body. Non-sensorial qualities (i.e., abstract qualities such as foreign-exchange rate) are not considered.
rdf:type : owl:Class
rdfs:subClassOf
sensor |
Sensors are conceptualized as physical objects, including the devices/instruments (e.g., a wind profiler or a stream gauge) and human observers (e.g., a weather observer or citizens supplying data about noise level in their neighbourhood). Sensors are sensitive to the physical aspects of the environment. They detect the presence of stimuli, as a result they perform observations. We exclude virtual sensors such as numerical models since the focus is on making inferences of real world occurrences from actual observations. Virtual sensors usually use other physical sensor readings to estimate properties. Therefore, we regard them as post-operations rather than actual observations performed upon real-world entities.
rdf:type : owl:Class
rdfs:subClassOf
actuates |
A geographic process may act as a geo-stimulus that actuates a sensor.
rdf:type | owl:ObjectProperty |
rdfs:domain | geo-stimulus |
rdfs:range | sensor |
owl:inverseOf | actuated-by |
has-functional-participant |
There are several ways in which geographic entities can participate in a geographic occurrence. The guiding idea is that an event or a process requires some sources (inputs) to occur and that these produce certain results (outputs). The source of a geographic occurrence may play a primary or a secondary role. A participant with a primary role is directly associated with an occurrence, in the sense that it actively controls the course of the occurrence. We regard the participating entity as the performer of an occurrence. It may initiate, perpetuate or terminate the occurrence. A secondary participant indirectly participates in an occurrence, either by facilitating or hindering the occurrence. The difference between a perpetuating participant and a facilitating participant is that a perpetuator sustains an occurrence; its involvement is crucial for the course of an occurrence, although it should be noted that a perpetuator will not necessarily be the initiator of an occurrence. The involvement of a facilitating participant is not necessary; it indirectly influences the occurrence, either by participating temporarily or by participating in another related occurrence. The results of an occurrence can be divided into an ultimate product resulting from the occurrence or a secondary participant that is affected as a result of the occurrence. The secondary participant is also known as patient.
rdf:type | owl:ObjectProperty |
rdfs:subPropertyOf | dul:participant |
owl:inverseOf | functional-participant-of |
References | |
---|---|
|
performed-by |
A participant with a primary role is directly associated with an occurrence in the sense it actively controls the course of the occurrence. We regard the participating entity as the performer or the doer of an occurrence. It may initiate, perpetuate or terminate the occurrence. An amount of precipitation initiates an infiltration occurrence. Amounts of water and sandy soil perpetuate an infiltration occurrence. A roadblock stops the flow of traffic on a road.
rdf:type | owl:ObjectProperty |
rdfs:subPropertyOf | functional-participant |
owl:inverseOf | performs |
References | |
---|---|
|
has-bearer |
The bearer of an observed property is a feature-of-interest.
rdf:type | owl:ObjectProperty |
rdfs:domain | observed-property |
rdfs:range | feature-of-interest |
rdfs:subPropertyOf | dul:inherent-in |
owl:inverseOf | bearer-of |
has-foi |
The relation between an observation-event and a feature-of-interest.
rdf:type | owl:ObjectProperty |
rdfs:domain | observation-event |
rdfs:range | feature-of-interest |
owl:inverseOf | foi-observed-by |
has-observation-value |
An observation-result is related to its concrete data values through the has-observation-value datatype property.
rdf:type | owl:DatatypeProperty |
rdfs:domain | observation-result |
has-obs-event |
It is also possible that the same feature-of-interest is observed periodically, thereby resulting in inferences of different geographic events. In such cases, the inferred events are linked to their respective sensing events via the relation has-obs-event.
rdf:type | owl:ObjectProperty |
rdfs:domain | geo-event |
rdfs:range | observation-event |
has-obs-ground |
The relation between an observation event and an observation ground.
rdf:type | owl:ObjectProperty |
rdfs:domain | observation-event |
rdfs:range | observation-ground |
has-obs-property |
The relation between an observation event and the properties it observes.
rdf:type | owl:ObjectProperty |
rdfs:domain | observation-event |
rdfs:range | observed-property |
has-obs-result |
The relation between an observation event and the results it produces.
rdf:type | owl:ObjectProperty |
rdfs:domain | observation-event |
rdfs:range | observation-result |
owl:inverseOf | obs-result-of |
has-physical-location |
This relation is used to specify the concrete values of a location.
rdf:type | owl:DatatypeProperty |
rdfs:domain | owl:Thing |
rdfs:range | Any |
has-temporal-sub-event |
For a given observation site, a geographic event is sub-event-of another greater event, if the time occupied by the sub-event is proper-part-of the time of the greater event.
rdf:type |
owl:ObjectProperty owl:TransitiveProperty |
rdfs:domain | geo-event |
rdfs:range | geo-event |
owl:inverseOf | sego:temporal-sub-event-of |
rdfs:subPropertyOf | dul:proper-part |
temporally-made-of |
If we follow the principle that only occurrences that share a similar temporal shape can be related by the parthood relation, geographic processes ought not to be specified as parts of a geographic event (or vice versa) . Our proposal is that a geographic event may have other events as its parts, and each of these sub-events can be constituted by geographic processes. For a given observation site, a geographic event is temporally cosntituted from a geographic process when the event takes place during a time interval and a related process goes on at least throughout that interval.
rdf:type |
owl:ObjectProperty |
rdfs:domain | geo-event |
rdfs:range | geo-process |
owl:inverseOf | comprises |
rdfs:subPropertyOf | dul:specific-constant-constituent |
covers |
A physical-object or a feature covers an observation-ground if it is one-sided-specific-constant-dependent on the ground and is not a part of the ground. For example, a bush shelter covering the ground, snow pack covering a site or straw covering a seedbed.
rdf:type |
owl:ObjectProperty |
owl:inverseOf | covered-by |
rdfs:domain | dul:physical-object or dul:feature |
rdfs:subPropertyOf | dul:specific-constant-dependence-on |
observed-by |
A sensor observes an observed property.
rdf:type |
owl:ObjectProperty |
rdfs:domain | observed-property |
rdfs:range | sensor |
owl:inverseOf | observes |
occurs-at |
This relation is used to specify the location of an occurrence. For example, we may associate the approximate location of a geographic event to an empirically defined site like an observation-ground. This mainly applies to an in situ sensing. It should be also noted that the location of a geographic event may not be representative of a larger geographical area encompassing an observation-region. Conditions nearby may vary, due to local peculiarities of terrain, vegetation, human habitation, or other factors.
rdf:type |
owl:ObjectProperty |
rdfs:domain | dul:perdurant |