摘要
分析了应用小波矩特征进行地面复杂背景下装甲车辆识别的理论依据,实地采集了某型坦克和某型步兵战车的灰度图像,提取其小波矩特征,采用支持向量机进行分类识别,进行了性能测试实验。结果表明：归一化后的图像的小波矩特征具有良好的不变性;小波矩特征对噪声和局部遮挡有较强的适应性,识别率比较稳定;支持向量机方法具有良好的分类识别能力。
The theory that the wavelet moment could be applied to solve the problem of recognizing ar- mored vehicles under complex ground background is analyzed. True gray images of certain tank and infantry fighting vehicle are acquired, the features of wavelet moment are extracted, and classification and recognition are carried out by using Support Vector Machine （SVM）. Performance test experiments are made, the results show that： the features of wavelet moments of the normalization images are stable; the characteristics of wavelet moment have better adaptation to noise and partial-cover, and the recognition rate is more stable ; the method of SVM has better capabilities of classification and recognition.
出处
《装甲兵工程学院学报》
2012年第3期61-64,共4页
Journal of Academy of Armored Force Engineering
基金
军队科研计划项目
关键词
小波矩
HU矩
支持向量机
识别
wavelet moment
Hu moment
Support Vector Machine （SVM）
recognition
