anomaly-injection Filter

The flir-filter-anomaly-injection plugin is a simple filter that allows the injection of anomalies into the data pipeline for numeric data. In order to test the anomaly detection functionality of FLIR Bridge’s internal monitoring and of the various anomaly detection filters and rules; it is useful to have a means to inject anomalies into otherwise good data or simulated data. The anomaly inject filter is able to inject a variety of anomalies into the data stream of a pipeline.

When adding the filter a configuration page for the filter will be shown as below;

anomaly_injection_1

  • Anomaly Type: Type of anomaly to be injected

  • Anomaly Magnitude: The magnitude of the anomaly as a percentage of the signal value

  • Anomaly Probability: The probability of an anomaly expressed as a percentage

  • Minimum Duration: The minimum duration, in seconds, for an anomaly

  • Maximum Duration: The maximum duration, in seconds, for an anomaly

  • Noise Type: The type of noise to introduce into the signal. Supported noise types are Random and Gaussian

  • Datapoint Name: Name of the datapoint to inject the anomaly

  • Clean Period: An initial period, in seconds, during which no anomalies will be injected into the data pipeline