摘要
以三维激光扫描系统获取变电站电气设备的点云数据,采用预处理措施后,通过模糊聚类使得点云数据朝着目标类别靠近,并通过加强似然系数对其进行修正,获得具有最大概率分布的目标关联概率矩阵,最后将其送入基于SVM改进的Adaboost算法分类器中进行分类识别。仿真结果表明训练精度与测试精度均高于98%,具有良好的识别精度。
The point cloud data of electrical equipment in substation is acquired by three-dimensional laser scanning system.After preprocessing,the point cloud data is made close to the target category by fuzzy clustering,and the target correlation probability matrix with maximum probability distribution is obtained by strengthening the likelihood coefficient to modify it.Finally,it is sent to AdaBoost improved based on SVM The algorithm classifiers are used for classification recognition.The simulation results show that the training accuracy and testing accuracy are both higher than 98%,which has a good recognition accuracy.
作者
王菲
王球
任佳依
刘晓波
刘浩
栗志元
WANG Fei;WANG Qiu;REN Jia-yi;LIU Xiao-bo;LIU Hao;LI Zhi-yuan(Economic and Technological Research Institute,State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 210000,China;Beijing Bo Chao Time Software Co.,Ltd.,Beijing 102206,China)
出处
《佳木斯大学学报（自然科学版）》
CAS
2020年第4期124-128,共5页
Journal of Jiamusi University：Natural Science Edition
基金
国家自然科学基金资助(51607106)
国网江苏省电力公司科技项目资助(J2018059)。
