The Analysis Of Influence Factors About The Near Infrared Multispectral Ice Detection On The Aircraft Wing

With the high-speed development of the national economy, the civil aviation industry in China has entered a period of rapid development. However, the deficiencies in the ability of running support have greatly affected the smooth flow and the safe operation of the flight of many airports because of the snow and ice in winter. Aircraft de-icing and residual ice detection job are particularly important, yet the current artificial residual ice detection does not meet the operational requirements. Automatic residual ice detection technology has become one of the winter running aviation security problem need to be solved badly.Currently, the near-infrared multispectral detection technology for domestic uses to detect residual ice is based on the differences of absorption and reflection characteristics of different substances on near infrared spectroscopy. The wing skin material contained aluminum or carbon fiber composite materials, the skin surface coated with multi-colored paint, temperature and density of ice itself, camera exposure, camera shooting angles, distances from infrared camera to measured surface, etc. Under the combined effects of these factors, it is extremely complex for data calibration, which may result in miss or false detection of residual ice. The detection accuracy of the ice thickness cannot be guaranteed either.Firstly, the principle of near-infrared multispectral detection of residual ice is analyzed. On this basis, these factors which affects residual ice detection, such as active infrared light source fluctuations, changes in ambient light, aircraft skin material and surface coating of different colors, image acquisition angle and distance, are researched on the thesis. Integrating sphere compensation system is proposed to compensate for the light source and ambient light; according to the law of grayscale and contrast C values obtained by experiment, a method which different color paint set different thresholds compensation was put forward and validated by experiments; on the basis of compensation, the linear relationship between ice thickness and contrast C values was determined through experiments. Ice thickness experiments verified that the error of the linear relationship used to detect ice thickness is less than 15%, and also validated that the residual ice detection system can be accurately achieved to detect a thickness of 0.5mm or more of residual ice.