An Improved Algorithm to Near Infrared Multispectral Ice Detection

Abstract

In order to improve the accuracy of near infrared multispectral ice detection on aircraft wings, we present an improved unit detection algorithm to reduce the noise influenced by the ice detection system itself. The improved algorithm, first select matrices being 3*3 as a detection unit, then calculate the contrast value of the five pixel points in the cross region of the unit, and finally count the number of pixel points, whose contrast value are larger than the threshold value of ice on the surface. The experimental results show that the unit detection method has high accuracy and efficiency compared with conventional point-by-point method.