【作者】 李丹；

【导师】 王霓虹；

【作者基本信息】 东北林业大学 ， 机械设计及理论， 2010， 博士

【摘要】 速生丰产林具有生长周期短、出材量高等特点,对恢复森林资源生态环境,缓解日益紧张的木材资源供求矛盾具有积极而重要的意义。近年来我国加大了对速生丰产林培育的投入力度,但林农和森工企业在培育速生林过程中缺乏有效科学指导,在此背景下,本研究建立了速生丰产林机械化培育过程集成服务平台,提出了面向网络应用的多专题柔性知识管理与推理建模方法,根据此方法分别建立了速生林培育过程专家咨询系统、速生林机械化生产专家系统、速生林生长模型预估与蓄积核算模型组件、速生林虫害防治专家系统,并结合WebGIS技术及门户技术建立了多元化“数字林业”专题应用,为林业信息化建设的网络应用模式进行了有益的探索。本学位论文依托国家“十一五”科技支撑计划项目“东北地区速生丰产林培育技术远程咨询系统”进行了如下几个方面的研究。在分析传统的基于网络应用过程中以面向表结构和表单建模方法来处理多专题信息所存在的缺陷,分别对多专题柔性知识存储方法、柔性知识管理、柔性推理方法进行建模研究,提出了一种柔性结构化的知识管理与推理建模方法并在平台中不同专题系统进行应用,增强了系统的可扩展性和柔性处理能力。速生丰产林培育过程专家咨询系统是针对基层缺乏领域专家指导的现状,以落叶松为研究对象,将培育过程分为割带、整地、刨穴、运苗、植苗、抚育伐方式和采伐方式,通过不同条件的推理过程,为用户提供不同环境下的培育过程咨询服务,此过程采用柔性推理的方法将不同知识结构的7个培育过程进行推理集成。速生丰产林机械化生产专家系统是针对现阶段林业生产过程中机械化程度不高,建立的集成林业机械配型选择、机械保养和机械作业规范等过程的专家系统。其知识提取是在培育过程咨询的基础上将不同培育过程的林业生产机械进行提取,研究了整地机械、造林机械、抚育机械及动力机械选择配型的推理算法,并实现集成推理过程。速生丰产林材积预估与核算模型组件是以黑龙江省林口林业局为示范区,通过对长白落叶松全林分生长模型进行研究,建立了全林分林木生长模型预估组件(横断面积生长模型、平均胸径生长模型、平均树高生长模型、公顷株数生长模型、公顷蓄积生长模型),为用户提供不同参数生长过程的预估模拟；并根据黑龙江省地方材积表建立的以落叶松为代表的4种速生树种的单株蓄积量核算模型组件,并在此基础上开发了模型管理组件,对相应模型组件参数进行管理维护。速生丰产林虫害防治集成专家系统是针对速生林在生长过程中易受到森林虫害的危害,在研究东北地区主要虫害诊断、发生期预测、发生量预测、防治方法、损失预估模型的基础上,以集成化的形式建立的速生丰产林虫害防治集成专家系统,为用户提供系统化的虫害防治服务,在应用中对速生丰产林材积预估与核算模型组件进行调用,可以对虫害的蓄积损失进行预估,并获得虫害防治收益预测。速生丰产林WebGIS及门户集成平台是针对以上4个应用专题软件进行的系统应用集成。WebGIS可将模型组件预估的结果以地图为载体进行可视化显示,以多元化形式反映数据的动态变化规律,从而提高决策过程中的合理性与科学性；门户技术集成是将整个专题软件及平台进行整合,为不同专题软件提供统一访问入口和后台管理功能,为用户提供全方位的服务平台解决方案。本论文的研究成果可为林农和基层森工企业在营造速生丰产林过程中提供理论指导和技术支持,为“数字林业”建设提供了完整范例,同时也对其他行业应用领域(如农业、水利等)服务平台建设具有一定的指导意义和借鉴价值。更多还原

【Abstract】 Rapid-growing and high-yield forest has the characteristics of short growing cycle and high output, and is a favorable way to solve the contradiction between supply and demand in forestry and to protect the ecological environment and biodiversity. In recent years, China has increased the base investment in the construction of rapid-growing and high-yield forest. However, forest farmers and logging enterprises lack effective directions of specialists in the process of forest breeding. In this background, this paper established mechanized breeding production information integration service platform for rapid-growing and high-yield forest and proposed a method of knowledge mangement and inference modeling by network application. Based on the method, we established expert advisory system, mechanized breeding production system, the component of growth-model forecast and accumulation accounting model, insect pest prevention system in breeding rapid-growing forest. This study is supported by The 11th Five Years Key Programs for Science and Technology Development of China. The main contents in the present study were shown as follows:After analyzing weakness and flaw of traditional table structure-oriented modeling in the process of dealing with multi-thematic information on web, the flexible modeling information storage method, flexible knowledge management and flexible reasoning were researched. A new flexible modeling for knowledge management and inference was proposed in this paper. This method can be used in different systems of the platform and enhance the ability of expansibility and flexible capacity.The softwares were constructed by including Breeding Consultation Expert System for Rapid-growing and High-yield Forest (BCES); Mechanization Production Expert System for Rapid-growing and High-yield Forest (MPES); The Stand Growth Prediction and Calculating Component for Rapid-growing and High-yield Forest (SGDPCC); Pest Prevention Expert System for Rapid-growing and High-yield Forest (PPES); WebGIS and Portal Integration Platform for Rapid-growing and High-yield Forest (WPIP).The BECS selected larch as materials, and it can be divided into the process of Strip-Clearing; Soil Preparation; Digging Holes; Delivering Seedlings; Planting Seedlings; Tending Felling Method and Cutting Method. Based on the inference in different processes, the system can provide consultation service for different condition parameters. The flexible inference was used in this process to integrate all the breeding consultation process.The MPES is divided into the process of Digging Holes Machinery, Forestation Machinerym, Tending Felling Machinerym and Motivity Machinery selecting and equipping process. All these reasoning process were integrated with an integration method based on flexible reasoning, and an algorithm for the machinery selecting and equipping was also proposed.Linkou Forestry Bureau, Heilongjiang Province, China was selected as the demonstration plot, and the whole stand model and single entry volume table for Heilongjiang province were researched, SGDPCC was established. It can predict the forestry growth and calculate the stock accumulation for single tree. Moreover, model management module was constructed for maintenance of the entire model modules.The insect pest prevention integrated expert system was established on the basis of prediction models for pest diagnosis, pest occurrence period, infecting quantity, disease prevention and control, damage assessment in the northeast of China. This system can provide services for the users by utilizing the component of cubic measurement and accounting model, and predict the loss of pest caused standing stock and the prevention benefits.WPIP integrated the BECS, MPES, PPES and SGDPCC for the whole information platform. WebGIS showed the visualization of geographic service and predicted results from model modules in visualization maps. It also showed the dynamic regularity for forestry changes with diversification pattern to ensure that decision-makings are scientific and rational. The portal technology was used to combine the all the applications as an entire system, it can provide the unite operator site and back-stage information management.The present research findings provide theory guidance and support of advanced technology for a large number of foresters and logging enterprises in constructing rapid-growing and high-yield forest, and take a good example for digital forest construction. Moreover, the platform constructed in this paper represents a useful effort for other application areas like agriculture or water conservancy.更多还原