基于SVM的无人直升机着舰标志识别方法研究

Identification Method of Landing Mark for Unmanned Helicopter Based on SVM

无人直升机在海战场有着非常广泛的用途,但自主着舰一直是其难题。针对基于视觉的无人直升机自主着舰问题,通过着舰标志检测得到较为理想的图像特征;然后采用基于支持向量机(supportvectormachine,SVM)的机器学习方法对着舰标志进行识别,实现整个着舰标志识别流程,准确率满足要求;并针对2种不同核函数对应的SVM训练算法进行了对比分析,指出了各自适用的条件,具有一定的现实意义和工程实践价值。

Unmanned helicopters have a wide range of application in naval warfare, but autonomous landing has always been a problem. In view of the identification problem of landing mark based on vision, a universal landing mark has been designed, ideal image characteristics has been obtained through the landing mark detection, and then the SVM(support vector machine)-based machine learning method has been used to identify the landing mark. The whole process of the landing mark identification has been implemented, and the identification accuracy can meet the actual demand. The SVM training algorithms based on the two different kinds of kernel functions have been analyzed to point out their applicable conditions, and they have important practical significance and practical engineering values.