基于克隆选择聚类改进的植被变化回归趋势分析

Improved Linear Trend Analysis in Vegetation Cover Change Based on Clonal Selection Clustering Algorithm

摘要:

基于克隆选择聚类算法对植被变化趋势回归分析中线性方程的斜率k进行聚类，并按聚类中心大小分类，解决了对斜率k的划分问题。实验结果表明：所提算法的DBI值0.65明显小于与之对比3种算法的DBI值0.89、0.91和0.89。北京市城区植被变化趋势显著增强的结论与1998~2011年北京市城区绿化率和人均绿化面积稳步增长的统计数据一致，验证了算法的正确性，因此该方法有一定的实用性，可应用于长时序植被变化趋势分析。

Abstract:

A clonal selection clustering algorithm was proposed to solve classification problem of the slope produced by the linear trend analysis. The experimental results showed that the DBI value of 065 was significantly less than the compared three algorithms. The result of vegetation change trend was in accord with the statistic data of greenbelt for everyone and green coverage percentage in Beijing from 1998 to 2011. So the proposed method had some application value and could widely apply in long time series of vegetation change trend analysis.