摘要
在用一维Kalman滤波研究森林资源动态基础上,介绍了二维Kalman滤波的原理与方法,并以浙江省丽水市森林资源连续清查样地的地理坐标建立二维坐标系,以样地森林面积和森林蓄积为状态向量,应用二维Kalman滤波研究森林资源动态。结果表明:对森林蓄积动态及样地个数较多、面积较大的用材林和防护林的面积动态预测效果良好,而对于样地个数较少、变异系数较大的样地如特用林和未成林造林林地的面积动态预测误差较大,如要作准确预测,则需要加大这些样地的抽样个数。
Accurately and promptly forecasting the current situation and dynamic for forest resources was one of the important subjects in forest resource management. Most published papers in the past studied the dynamic of forest resources dealt with one-dimenslon Kalman filtering. The principle and method of two-dimensional Kalman filtering was introduced in this paper. Two-dimensional coordinates system was set up by using the geographic coordinate of sample plots of continuous forest inventory in Lishui,Zhejiang province. The dynamic for forest resources was studied using two- dimensional Kalman filtering. The state vectors included the forest area and forest resources. There was no or small difference between the dynamic forecasting and the practical ones especially for timber forest. For the special forest and plots of unestablished stands the results were not good because there were few sample plots and big variation coefficient. To make the accurate forecast,the numbers of the sample plots needed be enlarged.
出处
《南京林业大学学报（自然科学版）》
CAS
CSCD
北大核心
2008年第1期19-23,共5页
Journal of Nanjing Forestry University：Natural Sciences Edition
基金
国家自然科学基金资助项目(30571491)
关键词
森林资源结构
二维Kalman滤波
状态方程
输出方程
动态预测
Forest resource structure
Two dimensional Kalman filtering
State equation
Output equation
Dynamic prediction
