Temperature Measurement Based on Image Processing & Neural Network

Temperature is the most frequently measured physical quantity, second only to time. In any industrial processes, the temperature measurement is an important requirement. Presently thermocouples, pyrometers and contact type sensors etc. devices and techniques are used for temperature measurement of several visible heat sources. Temperature measurement of furnace, boiler, burner flame etc. is a very difficult and tedious task also breakage of sensors occurs due to such high temperature so it requires daily maintenance. In this proposed system image processing and neural network technique is used to estimate the temperature of visible heat sources like furnace, burner, candle flame etc. System uses camera to take the images of heat source. Thermocouple will be used to measure the actual temperature. Various analytical techniques can be applied to estimate the color temperature correlation. This project experiments the model of Artificial neural network (ANN) to create the database of captured images and measured temperature so that approximate temperature is estimated.