采用混合策略联合优化的模糊C-均值聚类信息熵点云简化算法  

Fuzzy C -means Clustering Information Entropy Point Cloud SimplificationUsing Mixed Strategy Joint Optimization

作　　者：黄鹤[1,2] 黄佳慧 刘国权 王会峰[2] 高涛[3] HUANG He;HUANG Jiahui;LIU Guoquan;Wang Huifeng;GAO Tao(Xi’an Key Laboratory of Intelligent Expressway Information Fusion and Control,Chang’an University,Xi’an 710064,China;School of Electronic and Control Engineering,Chang’an University,Xi’an 710064,China;Institute of Data Science and Artificial Intelligence,Chang’an University,Xi’an 710064,China)

机构地区：[1]长安大学西安市智慧高速公路信息融合与控制重点实验室,西安710064 [2]长安大学电子与控制工程学院,西安710064 [3]长安大学数据科学与人工智能研究院,西安710064

出　　处：《西安交通大学学报》2024年第7期214-226,共13页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(62341301);西安市智慧高速公路信息融合与控制重点实验室(长安大学)开放基金资助项目(300102323502);中央高校基本科研业务费资助项目(300102324501)。

摘　　要：针对传统聚类算法处理点云简化问题时精度低、耗时长且易丢失特征信息等问题,提出了一种基于动态精英自适应混合策略的鹈鹕算法(DEAMPOA)与加权熵法联合优化的模糊C-均值聚类(FCM)信息熵点云简化算法。采用动态自适应种群混合策略,同时融合了精英反向化思路,显著提升了鹈鹕优化算法(POA)的收敛趋势和全局寻优能力,提高了寻找FCM最优聚类中心的成功率;利用DEAMPOA结合加权熵法对FCM进行优化,提高鲁棒性的同时增强了搜索精度,得到较好的聚类结果;在8种UCI标准数据集上与4种算法对比进行聚类性能评估实验,验证了所提方法综合性能优越;将所提方法与信息熵融合,并应用在三维点云KITTI数据集简化中。实验结果表明:与包围框简化法、随机采样简化法和特征选择简化法对比,所提方法全局误差简化前后点集之间平均欧式距离(MED)指标分别降低了2.25%、6.93%、5.74%,点云简化效果最优且运行速度满足要求。To solve the problems of low accuracy,long time consumption,and easy loss of feature information of the traditional clustering algorithm in conducting point cloud simplification,a point cloud simplification method for FCM information entropy jointly optimized by DEAMPOA and weighted entropy method is proposed.Firstly,a dynamic adaptive population mixing strategy is proposed,which integrates the elite reverse idea.This strategy significantly improves the convergence trend and global optimization ability of POA and increases the success rate of finding the optimal clustering center of FCM.Secondly,the optimization of FCM using DEAMPOA combined with weighted entropy method improves the robustness while enhancing the search accuracy,yielding better clustering results.Thirdly,clustering performance evaluation experiments are carried out through comparison with four comparison algorithms in eight UCI standard datasets,which verifies that the proposed method has superior comprehensive performance.Finally,the proposed method is fused with information entropy and applied to the KITTI point cloud dataset simplification.The experimental results show that the global error MED index of the proposed method is reduced by 2.25%,6.93%,and 5.74%respectively compared with that of the bounding frame simplification method,random sampling simplification method,and feature selection simplification method.Furthermore,this method generates the optimal point cloud simplification effect and the running speed meets the requirements.

关 键 词：C-均值聚类 鹈鹕优化算法 点云简化 信息熵 

分 类 号：U666.1[交通运输工程—船舶及航道工程]