Classification, Parameter Estimation and State Estimation, 2nd Edition

3Detection and Classification

Pattern classification is the act of assigning a class label to an object, a physical process or an event. The assignment is always based on measurements that are obtained from that object (or process, or event). The measurements are made available by a sensory system (see Figure 3.1). Table 3.1 provides some examples of application fields in which classification is the essential task.

Figure 3.1 Pattern classification.

Table 3.1 Some application fields of pattern classification

Possible
Application field	measurements	Possible classes
Object classification
Sorting electronic parts	Shape, colour	‘resistor’, ‘capacitor’, ‘transistor’, ‘IC’
Sorting mechanical parts	Shape	‘ring’, ‘nut’, ‘bolt’
Reading characters	Shape	‘A’, ‘B’, ‘C’
Mode estimation in a physical process
Classifying manoeuvres of a vehicle	Tracked point features in an image sequence	‘straight on’, ‘turning’
Fault diagnosis in a combustion engine	Cylinder pressures, temperature, vibrations, acoustic emissions, crank angle resolver	‘normal operation’, ‘defect fuel injector’, ‘defect air inlet valve’, ‘leaking exhaust valve’,
Event detection
Burglar alarm	Infrared	‘alarm’, ‘no alarm’
Food inspection	Shape, colour, temperature, mass, volume	‘OK’, ‘NOT OK’

The definition of the set of relevant classes in a given application is in some cases given by ...