摘要
通过分析室内室外图像的内容,发现其差异性有一定规律性,通常情况下,室内图像含有较高比例的具有一定规则的几何形状的人造物体,而室外图像含有一定比例的具有分形结构的自然物体。本文的方法是利用这个差异性和图像的颜色作为出发点,使用支持向量机(SVM)作为分类器依据图像的边缘和颜色矩特征对图像进行分类。实验结果表明此方法对室内、室外图像分类可以获得较高的准确率。
By analyzing the content difference between indoor images and outdoor images,usually,indoor images contain a certain proportion of straight edges,whereas outdoor images usually have a certain proportion of natural objects with abnormal edges.According to the difference and color,the method is that the classification feature is edge and color feature and the classification tool is the support vector machine.The result shows that high accuracy could be obtained.
出处
《广东轻工职业技术学院学报》
2010年第3期1-5,共5页
Journal of Guangdong Industry Polytechnic
关键词
支持向量机
特征
直线度
圆形度
support vector machine
feature
straightness
roundness
