摘要
为了能够快速、准确地识别飞机目标,文章给出了一种基于支持向量机的飞机目标自动识别方法;采用Touzi边缘提取,得到目标形状参数的几何特征,Hu不变矩等16个特征矢量作为SVM的训练样本,通过SVM训练得到飞机目标识别模型,从而完成飞机目标的自动识别;试验结果显示.该算法对不同尺度和模糊程度的飞机目标的识别度可达99%;该算法减少了样本训练时间,在提高识别准确率的同时降低了算法的复杂度,具有识别度高、识别速度快的特点,可用于飞机目标的快速识别。
In order to identify the aircraft target in images quickly and accurately, this paper proposes a method of automatid identification of the plane target based on support vector machines （SVM）. Using Touzi edge detection, geometric characteristics of the target shape parameters is generated. Applying Hu invariant moments and other feature vectors as the SVM training samples, through SVM training, we get the plane target recognition model and complete the automatic identification of the plane target. The results on images with different scale and clarity targets show that 99% targets are identified using this algorithm. ALL in all, the algorithm reduces the time of the sample training and the algorithm complexity, and at the same time it improves the recognition accuracy.
出处
《计算机测量与控制》
北大核心
2014年第9期2851-2852,2889,共3页
Computer Measurement &Control
关键词
支持向量机
自动识别
飞机目标
support vector machine
automatic identification
plane target
