A major concern in the petrochemical industry is both safety and efficiency of process heaters operation. As the tubes temperature and the symmetry of heating belong to the critical components, infrared imaging and measuring methods have been used worldwide for many years. Monitoring these high temperature objects has frequently been relying on highly subjective analyses, particularly due to fluctuations of flame and heating medium and/or sometimes inaccurate or not well-fit thermocouple data. Recent developments in infrared camera technology and data processing possibilities have brought significant progress for high resolution spatial and temporal analysis of infrared radiation distributions. This paper presents an innovative method which deals with the flickering and spectrally selective features of the heating mediums, analysed basing on capturing and elaboration of long sequence of images instead of the snapshot method. Thereupon, digital image processing algorithms enable automatic search of a few chosen statistic values for every pixel of the every frame, with the aim to form substitute images, which consist only from pixels of min., max, or mean values and their standard deviation distributions. By applying this new methodology, it is possible to separate extremes of fluctuating signals and, in result, to obtain deeper and more reliable knowledge about temperature distributions or about heating phenomena inside process furnaces. These data can be utilised to significantly increase heater throughput while helping to ensure safe operation of the heater. Many other applications could take advantage of presented idea, algorithm and tools.