Smart sensing of systems thermal behavior using low cost infrared cameras

Abstract

Diagnosis through the analysis of thermal images is widely used and is one of the favorite tools of predictive maintenance. Due to the high cost of infrared cameras, this analysis is carried out through periodic manual inspections reusing the sensor element for several applications. The recent appearance of low-cost infrared cameras has enabled the development of static applications that observe and analyze the thermal behavior of critical systems over time. This work presents a proposal of an intelligent thermal sensor that acquires and processes low resolution images in real time with energy restrictions. The proposed thermal image processing is based on the analysis of local maxima and spatial and temporal gradients centered on these maxima. The standard IEEE 21451-001-2017 was applied used for signal processing. The image acquisition and processing are executed into a 16-bit microcontroller resulting in a compact thermal sensor with low cost and low power consumption and with the ability to understand and classify the thermal evolution of the observed system.