摘要
针对喀斯特地区遥感图像中水体和山体阴影信息易混淆使得传统机器学习算法存在分类精度低、计算速度慢的缺点,改进XGBoost算法,通过交叉验证栅格搜索算法对XGBoost算法进行参数优化,构建CVGS-XGBoost分类算法,提取遥感图像中的水体和山体阴影信息。实验结果表明,CVGS-XGBoost算法的总体分类精度达到93.9%,比原始的XGBoost算法、决策树、随机森林和支持向量机算法构建的分类算法的总体分类精度分别提高1.5%、10.0%、6.3%和3.1%,且该算法与分类效果较好的支持向量机相比,运行时间开销少,可有效地识别喀斯特地区遥感图像中水体和山体阴影信息。
Low classification accuracy and slow calculation speed generally occur in extracting water bodies and mountain shadows from remote sensing images with the traditional machine learning algorithms in karst area.After CVGS-XGBoost algorithm was developed by means of optimizing the parameters of XGBoost algorithm with the cross-validation grid search algorithm,the algorithm could classify water bodies and mountain shadows from remote sensing images.The experimental results show that the overall classification accuracy with CVGS-XGBoost algorithm reaches 93.9%,which is 1.5%,10.0%,6.3%and 3.1%higher than that of the traditional XGBoost algorithm,decision tree algorithm,random forest algorithm and support vector machine respectively.Meanwhile the CVGS-XGBoost algorithm run-time is less than that by support vector machine with a better classification.The modified algorithm can be used to recognize water bodies and mountain shadows from remote sensing images effectively.
作者
秦琴
王修信
QIN Qin;WANG Xiu-xin(College of Computer Science and Information Technology,Guangxi Normal University,Guilin 541004,China;Guangxi Key Lab of Multi-Source Information Mining&Security,Guangxi Normal University,Guilin 541004,China;Guangxi Collaborative Innovation Center of Multi-Source Information Integration and Intelligent Processing,Guilin 541004,China)
出处
《桂林理工大学学报》
CAS
北大核心
2020年第4期850-858,共9页
Journal of Guilin University of Technology
基金
国家自然科学基金项目(41561008)
广西自然科学基金项目(2014GXNSFAA118289)。