摘要
针对高分辨率遥感影像建筑物精度较低问题,提出了多特征自适应融合的高分辨遥感影像建筑物自动提取算法。首先,提取纹理特征与形态学建筑物指数;然后,采用混合核函数对多种特征进行自适应融合,采用单分类器提取建筑物范围;最后,结合面向对象思想,采用建筑物形状特征对提取的建筑物结果进行优化处理已获得的最终结果。试验结果表明,该方法能有效避免漏洞和“椒盐噪声”现象,提高建筑物检测精度。
Aiming at the low accuracy of buildings in high-resolution remote sensing images,an automatic building extraction algorithm based on multi-feature adaptive fusion is proposed.Firstly,texture features and morphological building index are extracted;Then,the mixed kernel function is used to adaptively fuse multiple features,and a single classifier is used to extract the building range;Finally,combined with the object-oriented idea,the extracted building results are optimized by using building shape features,and the final results are obtained.The experimental results show that this method can effectively avoid loopholes and"salt and pepper noise"and improve the detection accuracy of buildings.
作者
槐燕萍
HUAI Yanping(Gansu Institute of Surveying and Mapping Engineering,Lanzhou 730000,China;Gansu Emergency Surveying and Mapping Engineering Research Center,Lanzhou 730000,China)
出处
《测绘与空间地理信息》
2023年第6期151-154,共4页
Geomatics & Spatial Information Technology
关键词
多特征
自适应融合
建筑物提取
单分类器
multi-feature
adaptive fusion
building extraction
single classifier