摘要
提出了运用支持向量机进行目标检测的方法。通过对航空影像中的军事目标和自然背景两类样本进行学习,支持向量机检测方法建立了针对目标和非目标有效区分的识别模型,该模型能够对航空影像中所有的区域进行快速的检测和识别,检测到所有感兴趣的人造军事目标。试验表明,该方法快速、高效且具备一定的鲁棒性。
In the light of the problem that the man-made targets are difficult to detect in the aviation images with complex background, a new method using support vector machines is put forward in this paper. Through training the target and non-target samples, support vector machines method can set up an effective recognition model, which can detect all possible interested man-made targets through scanning all areas in images. Experiments show that this method is rapid, effective and robust.
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2004年第10期912-915,932,共5页
Geomatics and Information Science of Wuhan University
基金
测绘遥感信息工程国家重点实验室开放研究基金资助项目(020101)。
关键词
支持向量机
模式识别
目标检测
support vector machines
pattern recognition
target detection
