Vision-based measurement of air temperature using smoke as medium

The near zone of heat diffusing machines in an equipment room is very critical for the normal functioning of the machines. Current knowledge about the near zone of heat diffusing machine is insufficient, causing an increasing need for better measuring methods and representation of the air temperature distribution. We proposed a vision-based measuring technique for visualization of air temperatures and air flow patterns over a large area. As a medium, smoke is generated in the near zone of heat diffusing machine to make the air flow visible. A specialized capturing technique is used to record density distribution of the time-varying smoke. We place a few temperature sensors sparsely over the volume and combine the outputs of the temperature sensors and density information of the smoke to measure the temperature distribute in the space. Experiments validate the effectiveness of the proposed method.