基于植被指数和神经网络的热带人工林地上蓄积量遥感估测

热带森林作为陆地生态系统的组成成分之一,研究其蓄积量估测对我们了解其在全球碳循环中的地位和作用有很重要的意义.但遥感估测森林生态参数的精度如何,还是个不确定的问题.利用LANDSAT-TM数据,基于森林清查数据和遥感技术,以尾叶桉和加勒比松为例,对中国南方地区人工林蓄积量估测进行了尝试研究.首先,通过测量样方胸径、树高,建立森林蓄积量估算模型.其次,通过对比分析不同植被指数与森林蓄积量之间的关系,选择合适植被指数组合,建立多元回归和神经网络模型.结果表明:单波段TM数据和大多数植被指数与蓄积量相关性并不好.神经网络比回归分析模拟效果好.而多元回归和神经网络模型大大提高预测精度.本研究方法对大面积的森林蓄积量估测具有一定的参考价值.

Estimating tropical forest stock volume based on vegetation indices and neural network model

Tropics forest is one of a main component of terrestrial ecosystems, it play an important role in terrestrial ecosystem car-bon cycle. But it can only be estimated with great uncertainties .In this paper, based on foest inventory data and remote sensing tech-nology, forest stoke volume was studied. Firstly, according to the status of the tropic Forests and selected Pinu scaribaea and Euca-lyptus urophylla as typical forest vegetation types, the biological characteristics parameters of the vegetation was surveyed, including forest age, diameter at the breast height, height, etc.. Based on field forest inventory data, forest stoke volume of forests in the sample was estimated using the relationship between forest stoke volume and diameter at the breast height, height. Then vegetation index were selected and the multiple regression model and neural network model between Ⅵ and forest stoke volume was set up; This study concludes that single band and many vegetation indices are weakly correlated with selected forest stoke volume due to the saturation of bands, multiple regression models and neural network model improve stoke volume estimation performance and neural network model is better than multiple regression. It provides a good method and theoreticly base for evaluating forest stoke volume by the method can be useful in areas where no other forest information is available.