MOS Gas Sensors for Food Quality Monitoring using GC-MS and Human Perception as Reference

An approach for training of temperature cycled metal oxide semiconductor gas sensors to monitor food quality is presented. A setup including various gas sensors and a GC-MS is used to measure the headspace of foodstuffs like strawberries in food containers. The food quality is assessed by several persons with respect to appearance, smell, and overall edibility. The relation between the human perception, GC-MS component peaks and the gas sensor data is evaluated. Significant features are extracted from the temperature cycle to derive a model that is able to project the sensor data to the assessed edibility of the investigated food.