AI4IoT Sensor Calibration
This asset was development for the AI4IoT Pilot, with the goal of providing a tool to calibrate the measurements of low-cost sensors. The model is pre-trained for the low-cost sensor deployed in Trondheim, Norway.
The Calibration component is part of the AI4IoT pipeline. It implements a machine learning model that calibrates data coming from low-cost sensors, such that the output is as close as possible to reference values. The component is deployed with a pre-trained model and outputs the calibrated values for PM2.5 and PM10 measurements. Inputs are PM measurements from the sensor and meteorological data.
The Calibration module exposes the calibrate_sample service, which receives raw data in the message input and output the values of the calibrated measurements for PM2.5 and PM10.
See the GitHub repository and the AI4IoT Pilot page for more information.