摘要
在图像分类和识别技术中,针对图像特征的提取这一重要问题,文章实现了用三种不变矩对图像目标提取特征并进行分类的方法,该方法分别提取目标的区域中心矩、区域中心不变矩和径向矩作为特征矢量,并与区域协特征方差算子进行图像分类比较分析。将特征用支持向量机的方法进行分类器的训练,再对目标进行分类。实验结果表明,区域径向矩算子运用在车辆目标分类中具有较好的效果。
An image classification method is presented based on moment invariant and support vertor machine in this paper. Firstly, it extracts the three invariant moments as feature vectors, including regional central moment, regional central moment invariant, and regional radial moment. Then this paper compares the three invariant moments and regional covariance. Finally it uses the support vector machines （SVM） to recognize vehicle targets. The experimental results show that the regional radial moment operator has the best effect in the vehicle classification.
出处
《大众科技》
2014年第9期12-15,共4页
Popular Science & Technology
关键词
区域不变矩
支持向量机
图像分类
目标识别
目标分类
Invariant moment
SVM
image classification
target recognition
object classification
