Temperature Detection Algorithm Based On Digital Image Processing Improvements

A real time and high precision temperature measurement method is expected because of its positively significant in improving the quality of products, saving energy resource, protecting environment and production safety. The temperature measurement method based on the digital image processing is developed in recent years due to its simpleness and its high efficiency. In this thesis, we analyze large numbers of experimental data and improve the processing algorithm.The main works are as follows:1. We analyze the radiative properties of scattering medium in the boiler in the visible range and deduce the expression between the temperature of the radiation object and the luminance of pixel in the image produced by CCD camera;2. We have designed a temperature measurement system based on the CCD image and validate this system on the black-body oven;3. We adopt an algorithm based on the structure of adaptive center weighted modified trimmed mean filter to eliminate the mixed noise in the CCD image;4. We apply the federated Kalman filter to temperature measurement data fusion and improve the temperature measurement precision.