计算机科学 ›› 2014, Vol. 41 ›› Issue (8): 289-292.doi: 10.11896/j.issn.1002-137X.2014.08.061

• 图形图像与模式识别 • 上一篇    下一篇

基于混合PSO的高斯混合模型地形分类

韩光,孙宁,李晓飞,赵春霞   

  1. 南京邮电大学宽带无线通信技术教育部工程研究中心 南京210003;南京邮电大学宽带无线通信技术教育部工程研究中心 南京210003;南京邮电大学宽带无线通信技术教育部工程研究中心 南京210003;南京理工大学计算机科学与工程学院 南京210094
  • 出版日期:2018-11-14 发布日期:2018-11-14
  • 基金资助:
    本文受国家自然科学基金:支持增量式稀疏编码的在线协同目标跟踪研究(61302156),基于多模型嵌入技术的复杂环境感知研究(61101197),基于稀疏描述的非结构化环境地形识别研究(61272220),江苏省高校自然科学研究面上项目:支持异构协同在线更新的稀疏表示目标跟踪研究(13KJB510021)资助

Gaussian Mixture Model Terrain Classification Based on Hybrid PSO

HAN Guang,SUN Ning,LI Xiao-fei and ZHAO Chun-xia   

  • Online:2018-11-14 Published:2018-11-14

摘要: 提出了一种基于改进的混合粒子群优化(particle swarm optimization,PSO)算法的高斯混合模型地形分类方法。高斯混合模型的求解通常是使用期望最大化算法(expectation maximization,EM),然而EM算法易陷入局部最优,收敛速度不稳定且对初值敏感。因此引入混合PSO算法,并对其进行了一系列改进。实验结果表明:改进后的算法较其它优化算法提高了全局搜索能力和收敛速度,利用该算法求解高斯混合模型可以提高参数估计的精度,并且在户外场景图像的地形分类实验中所提出的地形分类方法也表现优良。

关键词: 混合PSO算法,高斯混合模型,EM算法,地形分类

Abstract: Gaussian mixture model terrain classification based on improved hybrid particle swarm optimization (PSO) algorithm was presented.The expectation-maximization (EM) method is a popular method to solve the Gaussian mixture model,but it is a local optimization method with instable convergence rate and initial value sensitivity.Therefore the hybrid PSO algorithm was introduced,and a series of improvement was conducted.Experimental results show that the improved algorithm can greatly improve the global convergence ability and enhance the rate of convergence.Using the improved algorithm to solve Gaussian mixture model can improve the accuracy of parameter estimation,and the proposed terrain classification method also has excellent performance in the terrain classification experiment of outdoor scene image.

Key words: Hybrid pso algorithm,Gaussian mixture model,EM algorithm,Terrain classification

[1] Christopher A B,Karl I.Self-supervised terrain classification for planetary surface exploration rovers[J].Journal of Field Robo-tics,2012,29(3):445-468
[2] Marcel H,Marc A,Nicolai W,et al.Probabilistic terrain classification in unstructured environments[J].Robotics and Autonomous Systems,2013,61(10):1051-1059
[3] Nguyen D V,Kuhnert L,Jiang T,Kuhnert,et al.A novel approach of terrain classification for outdoor automobile navigation[C]∥Proceedings of the 2011 IEEE International Conference on Computer Science and Automation Engineering.2011,3:609-616
[4] Rana H,Adrian P,Carlos E O,et al.A support vector machine for terrain classification in on-demand deployments of wireless sensor networks[C]∥Proceedings of the 2013-7th Annual IEEE International Systems Conference.2013:841-846
[5] Haselich M,Eggert S,Paulus D.Parallelized Energy Minimiza-tion for Real-Time Markov Random Field Terrain Classification in Natural Environments[C]∥Proceedings of the 2012 IEEE International Conference on Robotics and Biomimetics.2012
[6] Thomas J P,Keith C C,William A M,et al.An improved simplemorphological filter for the terrain classification of airborne LIDAR data[J].ISPRS Journal of Photogrammetry and Remote Sensing,2013,77:21-30
[7] Li W,Prasad S,Fowler J E.Hyperspectral Image Classification Using Gaussian Mixture Models and Markov Random Fields[J].IEEE Geoscience and Remote Sensing Letters,2013,11(1):153-157
[8] Al-Jubouri,H,Du H,Sellahewa H.Applying Gaussian mixture model on discrete cosine features for image segmentation and classification[C]∥Proceedings of the 2012 Computer Science and Electronic Engineering Conference.2012:194-199
[9] Manduchi R.Obstacle detection and terrain classification for autonomous off-road navigation[J].Autonomous Robots,2005,18(1):81-102
[10] 韩光.面向非结构环境图像理解的算法研究[D].南京:南京理工大学,2010
[11] Eberhart R C,Shi Y.Particle swarm optimization:Develop-ments,applications and resources[J].The 2001 Congress on Evolutionary Computation,Seoul,Korea,IEEE Press,2001(1):81-86
[12] 陈允杰,张建伟,韦志辉,等.基于高斯混合模型的活动轮廓模型脑MRI分割[J].计算机研究与发展.2007,44(9):1595-1603
[13] 高尚,韩斌,吴小俊,等.求解旅行商问题的混合粒子群优化算法[J].控制与决策,2004,19(11):1286-1289
[14] Castano R,Manduchi R,Fox J.Classification experiments on real-world texture[C]∥Proceedings of the Third Workshop on Empirical Evaluation Methods in Computer Vision.Pasadena,CA:Jet Propulsion Laboratory,2001:3-20
[15] Fan S,Lin Y.A multi-level thresholding approach using a hybridoptimal estimation algorithm[J].Pattern Recognition Letters,2007,28(5):662-669
[16] Kim S C,Kang T J.Texture classification and segmentationusing wavelet packet frame and Gaussian mixture model[J].Pattern Recognition,2007,40(4):1207-1221

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!