A Research of GPS Height Fitting in Mountainous Terrain by CPSO Optimization FLS-SVM

Abstract:

For the problem of data limited in the mountainous area, a method of FLS-SVM (Fuzzy Least Square Vector Machine) that supporting small sample data and having high noise ability was put forward. The CPSO(chaos particle swarm optimization algorithm) is adopted to optimize the parameters of least squares support vector machine algorithm, and to avoid the uncertainty of artificial parameter selection. Meanwhile, considering the impact of terrain, the terrain correction is introduced to the support vector machine model. The experimental results show that the model can get higher precision fitting effect compared with traditional fitting method such as PSO-LSSVM and GA-LSSVM, and suitable for the SRTM application of getting normal height.

• Periodical: Applied Mechanics and Materials (Volumes 336-338)
• Pages: 2339-2343
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.336-338.2339
• Online since: July 2013
• Authors: Yong Gan, Xin Xin Liu, Yuan Pan Zheng
• Keywords: CPSO, FLS-SVM, Height Anomaly, Terrain Correction

References

[1] Yu Xiaoping, Yang Guodong, Wang Fengyan etc. GPS elevation fitting convert normal height[J]．Science of Surveying and Mapping, 2007, 32(2): 40-42.

[2] Lv Lei, Chen nannan, Dong Xianglin etc. Application of Improved Neural-Ntework Algorithm to GPS HEight Transformation[J]. Journal of Geomatics, 2012, 37 (3): 9-12.

[3] Kong Lingjie, Li Jiebin, CHEN Wei. Application of BP neural network based on GA in height fitting[J] . Geotechnical Investigation & Surveying, 2010, (3): 61-64.

[4] Dong Yu-lei, Gao Fei．The application of support vector machine method in GPS height transformation[J]. Journal of Shandong University of Technology, 2007(11): 64-66.

[5] Xie Bo, Liu Lian-wang. Application of support vector machines in GPS leveling[J]. Science of Surveying and Mapping, 2011, 36 (1): 172-74.

[6] Zhou Lihan. Least Square Support Vector Machine in GPS Height Conversion[J]. Chinese Journal of Engineering Geophysics, 2010, 7 (2): 243-247.

[7] Tang Kaihuang. GPS Height anomaly Surface fitting based on topographic correction [J] . Mine Surveying, 2008(1): 35-37.

[8] Yongqi Chen, Zhangxin Zhou, Qijun Chen. The Research and Application of LS-LVM Based on Particle Swarm Optimization[C]/Proceedings of the IEEE International Conference on Automation and Logistics Jinan, China, 2007: 1115-1120.

DOI: 10.1109/ical.2007.4338735

[9] Wu S, Amari S. Conformal transformation of kernel function, a data department way to improve support vector machine classifiers．Neural Processing Letters(S1370-4621) , 2002(15): 59-67.

[10] Li Ye. Research on Fuzzy Support Vector Machine and ITS Application on Fault Diagnosis[D] . Shenyang Aerospace University, (2011).

[11] Zhang Xingfu, Liu Cheng. The Approach of GPS Height Transformation Based on EGM2008 and SRTM/DTM2006. 0 Residual Terrain Model[J]. Acta Geodaetica et Cartographica Sinica, 2012, 4(1): 25-32.