【作者】 武巧英；

【导师】 陈丽华；

【作者基本信息】 北京林业大学 ， 农业生物环境与能源工程， 2011， 硕士

【摘要】 保护森林就是保护人类的生存环境和社会的持续发展,北京作为全国的政治、经济和文化中心,对首都的森林生态的保护有着深远的意义。本文以森林健康理论、预警理论和粗糙集理论为研究基础,并将三种理论有机结合,以森林健康评价为基础,森林健康预警为主体,粗糙集理论为途径构建森林健康预警系统：结合森林健康理论和预警理论,论述了森林健康预警研究的必要性和重要性；分析影响森林健康的因素及其作用机理,基于PSR(压力-状态-响应)框架深入分析了森林健康预警的过程；按照森林健康预警系统的构建思路,构建森林健康预警系统；基于粗糙集理论构建了森林健康预警计算模型,并进行指标属性的约简和权重的计算；通过建立预警调控措施专家库和各预警管理系统,实现了预警调控子系统的功能。森林健康预警研究是森林健康理论与预警理论相结合的产物,本研究首次将粗糙集理论应用于森林健康预警中,以粗糙集理论方法作为森林健康预警计算模型,从指标的约简到指标权重的确定,都为森林健康预警提供了新的方法。通过将粗糙集约简理论和BP(误差反向传递)神经网络相集合,扩展了粗糙集理论在预警预测领域的应用。将粗糙集方法作为BP神经网络的前置预处理系统,可以简化神经网络的构造,减小神经网络的训练集并缩短网络的训练学习时间。本研究将这一混合模型应用于鹫峰国家森林公园的健康预测中,分析并验证了这一模型的合理性和可靠性。利用建立的森林生态系统健康预警系统,结合北京山区森林健康发展状况,对北京山区的森林生态系统健康进行预警。预警诊断子系统输出结果表明,北京市山区15块研究样地中,处于无警状态的有4块,轻警4块,中警3块,重警3块,巨警1块。整体警情较严重,森林管理者通过利用预警调控子系统分析森林“病变”的根源,结合专家库,提出相应的森林健康恢复措施,可及时、有效的排除警患。更多还原

【Abstract】 Protection of forests is to protect the human living environment and sustainable development of society. Protection of forest ecosystems is very necessary for Beijing, due to the center of political, economic and cultural of China. In this thesis, an early warning system are built, based on theory of forest health, early warning and rough set theory, as well as evaluation of forest health and forest health warning. We study the need for forest health, the importance of early warning research, analysis of factors affecting forest health and the role of incentives systematically. And also the process of forest health warning is analyzed in depth based on the PSR framework. According to the idea of constructing of forest health warning system, a forest health early warning systems was built; based on rough set model, a model of forest health warning was constructed and calculate the indicator properties and weight reduction calculation; Through the establishment of early warning and the warning expert database management system we achieve the Early Warning Control Subsystem function.Research on early warning of forest health is combination of forest health and early warning theory. In this paper, we first apply the rough set method to early warning of forest health. The rough set theory method is used as early warning model of forest health from reducing to weighting the indicators. Both of them provide a new approach to the early warning of forest health.By integrating the reduction of rough set and BP neural network, we expand the rough set theory in the field of forest health early warning and forecast. Taking the rough set as a preconditioning system of BP neural network, we can simplify the neural network’s structure and reduce the training set of neural network so as to save a lot of training time to learn. The hybrid model is applied to the health forecast of Jiufeng National Forest Park, which proves that the model is reasonable and reliable.We use the established forest ecosystem and combine the development status of Beijing forest health for early warning on Beijing forest ecosystem health. The outputs of early warning diagnostic subsystem showed that there are four no warning blocks, four light warning blocks, three heavy warning blocks and one huge warning block in the fifteen forest areas of Beijing. The whole warning of Beijing forest areas is more serious. Forest mangers analyzed the source of forest disease through combining warning controlled subsystem with expert databases. At last, the corresponding forest health restoration measures can be timely and effective provided in order to suffering from exclusions.更多还原