基于机器视觉的飞机模型冰形轮廓提取方法研究

Research on ice contour extraction of aircraft model based on machine vision

测量结冰风洞试验中飞机模型结冰冰形的轮廓,对飞机设计和测试飞机安全性能都具有重要的作用和意义。在冰形测量中,非接触式测量具有较大的优势,但是当前的非接触式测量方法流程复杂,且试验成本较高,制约精度的因素较多。提出了一种非接触式的冰形图像处理方法,能实现对采集来的冰形轮廓进行识别,经过试验验证该方法具有较高的鲁棒性,在采集得到的一千张冰形图像上实现了冰形识别和曲线提取,并进一步得到冰形曲线的数字化结果。冰形曲线的数字化结果不仅有利于从二维层面分析不同条件下产生的冰形的几何特征,同时对进一步开展飞机结冰飞行、结冰防护和结冰适航取证等方面的研究具有重要的意义。

It is very important to measure the icing profile of aircraft model in icing wind tunnel test for aircraft design and safety performance test.In ice shape measurement,non-contact measurement has great advantages,but the current non-contact measurement method has complex flow,high test cost and many factors restricting precision.In this paper,a non-contact ice image processing method is proposed,which can recognize the Ice contour.The experiment shows that the method has high robustness,the recognition of ice shape and the extraction of ice curve are realized on the 1000 ice images,and the digital result of ice curve is obtained.The digitizing results of ice shape curves are not only helpful to analyze the geometric features of ice shape produced under different conditions at two-dimensional level,at the same time,it is of great significance for further research on aircraft icing flight,icing protection and icing airworthiness certification.