Research And Application On Aircraft Remnant Ice Detection Technology Based On NIR

In the cold winter with deteriorative weather, to ensure aircraft flight safety,remnant ice detection is a necessary process before the plane take-off or after thedeicing operation. Incomplete deicing operations and refreezing stuff are significanthazards for flight safety, which has caused several air disasters. Therefore,improvement on aircraft remnant ice detection technology and method is an importanttask to guarantee the safe operation of aircraft in winter. With the rapid developmentof air transport, conventional artificial visual and tactile inspection methods cannotmeet the need of safe efficiency operation of civil aviation in winter.To overcome the disadvantages of traditional remnant ice detection, such as highfalse detecting rate and missing rate, slow speed, the spectral characteristics of ice andwater on aircraft are analyzed; a near infrared multispectral solution of aircraftremnant ice detection and measurement is proposed. The main contents are asfollows:(1) This paper researches on remnant ice reflection and scattering mechanismand near infrared multispectral detection theory. The reflection and scattering aredifferent when the infrared light form infrared light source reaches the differentsurface. The performance of ice surface is significant different from aircraft skin, sothat the aircraft remnant ice is detected and located.(2) Infrared multispectral detection system is constructed with infraredmultispectral ice detection unit and visual detecting area location unit. The remnantice is accurately positioning by visual detecting area location unit based on attitudeestimation. The specific characteristics of aircraft structure and CCD camera are usedto position the detection area. Three band filters drove by the disk are in front of thelens separately when the image is obtained. The result is displayed in real time underpseudo color.(3) This paper presents an improved RANSAC robust matching algorithm, withapplication of pre-inspection procedures based on the disparity gradient constraint andthe effective area constraints, we can effective reduce the count of calculation errorhomograph, thus eliminating the corresponding homograph test session. This paperpresents a matching framework aimed at the four images matching algorithm. Andincrease the matching efficiency to a higher degree, shorten the matching time. (4) The strategy of influence elimination is researched. The aircraft skin material;the coating layer and the infrared radiation are concerned. The compensation andcorrection of system is completed. The performance of the system can be improved bythe compensation and correction according to different spectral characteristics ofaircraft skin materials. A compensation method based on integrating sphere isproposed to overcome the instability of infrared light source and interference ofbackground infrared light. Different wave bands intensities of light source aredetected and feedback in real time, and the system get compensated and corrected.(5) This method is based on infrared spectral region detection and attitudeestimation positioning technologies. Infrared camera achieves detection of localremnant ice. Positioning camera achieves the conversion of the scanning area. Thusthis method can complete the positioning of remnant ice. The attitude estimationalgorithm is used to get the external position. CCD camera extracts the feature pointsof the specific parts of the aircraft. According to the relationship between thesefeature point's space coordinates and image coordinates to determine the relativeposition of the device. It is possible to get the camera's position to obtain the positionof infrared camera relative to aircraft. Remnant ice eventually gets the positionrelative to the aircraft.