【作者】 姜春玲；

【导师】 吴泉源；

【作者基本信息】 山东师范大学 ， 自然地理学， 2007， 硕士

【摘要】 森林是陆地生态系统的主体,它是实现环境与发展相统一的关键和纽带,对社会和经济的可持续发展,具有极其重要的战略意义。林业资源是森林资源的重要组成部分,是实现国民经济可持续发展的根本与保障。因此,及时掌握林业资源状况及其消长情况,对于林业资源的合理利用和抚育更新,意义重大。在科技迅速发展的今天,遥感技术和地理信息系统技术为全球变化研究提供了主要数据源和技术手段,也成为林业资源调查研究的一种重要的现代化研究方法。本文以龙口市为研究区,在“龙口市土地利用数据库更新”课题的支撑下,完成了基于遥感技术和GIS技术的林业资源专题信息的提取,并从空间角度对林业资源的分布状况进行了详细的阐述。论文的主要研究内容和研究成果有:1、通过对TM遥感图像地物光谱信息的统计分析,确定了遥感图像合成的最佳技术方案,成为后期决策树分类精度提高的先决条件。2、利用计算机自动识别的方法提取林地类型和林分信息。在分类的过程中,将与林业资源分布密切相关的高程数据、归一化植被指数等知识信息与经过分类预处理的图像光谱信息综合在一起形成知识库,通过提取知识、组织规则、形成决策树,最终建立起专家分类识别系统。然后,利用Erdas软件下的专家分类器实现了各种专题信息的高精度提取。3、在对林木蓄积量进行估算时,借助TM遥感图像采用两种绝然不同的方法分别进行。具体来说,一种是直接利用公式(纯林地的面积=林地总的分布面积×植被覆盖度)生成纯林地的面积,然后采用传统的野外样方调查方法计算出每个样方的林木蓄积量,最终得出研究区总的林木蓄积量。一种是根据森林的反射光谱特征与林木蓄积量之间的相关关系,建立林木材积量的光谱估算模型。通过比较发现,上述两种方法的误差不大,但从推广应用的角度来说,方法一可以不借助其他的影像数据,只要获取几个关键数据,即使通过野外测量也可以进行运算。对于经济不发达的地区,该方法的实用性比较强,可以大范围推广,但是如果不采用图像,该方法是一项很费时、费力的工作。而方法二必须借助图像数据,对于大范围、快速的估算林木蓄积量有很大的推广价值。4、本文从高程、坡度、坡向三个角度借助分布指数对林地类型、林分的空间分布进行了研究。研究结果表明:高程是影响林地类型、林分分布的一个重要因素,随着海拔高度的增加,大气的湿度、温度等会有明显的变化;坡度不同,导致土壤肥力不同,即投入与产出比不同;坡向影响着日照时数(即太阳辐射能量的分配)和土壤水分的再分布,因此上述三种因素共同作用导致林地类型和林分在不同高程、坡度、坡向上的分布不同。更多还原

【Abstract】 Forest is the main terrestrial ecosystems. It is the key and link to realize the unify of environment and development. It has extremely important strategic significance to the sustainable social and economic development. Forestry resources is an important component of forest resources, and it is the basement and safeguard to realize the sustainable development of the national economy. Therefore, mastering forest resources and its situation are of greet significance to the reasonal utilization and raising update. Today, technology develops rapidly; remote-sensing technology provides the main source date and technological means for global changing research. It also becomes an important modern research method of forestry resources investigation. This paper takes Longkou City as a research area, which completed the absorbing of forestry resources information based on and remote-sensing technology with the help of subject“Database updating of land use in Longkou City”. The paper also illustrates the distribution of forestry in detail from a space angle. Main research contents and results in the article are as follows:Through the statistical analysis of TM remote-sensing image. Spectral information, the paper made the best available method of remote-sensing image connection, which becomes the prerequisite for the late improving of Classification Decision Tree accuracy.To obtain the type and method of extracting forest stand information using computers automatically. In the classification process, the elevation data closely related to the distribution of forest resources, and normalized vegetation index together with information after pretreatment classification combined to fore the knowledge base. By extracting the knowledge, organizing rules, creating decision trees, experts classification system will be established finally. Then, high-precision extraction of information on various topics was achieved by using the experts classification device under Erdas software.When does volume estimation of forest, two entirely different ways were carried out with the help of TM remote-sensing images. Specifically, one is the direct use of a formula (Pure forestry area equals the total area of forest land multiply vegetation coverage area) which creates pure forestry area. Then the field was surveyed using the traditional method for calculating the wood reserves every kind to obtain the research area total standing crop quantity finally. The other way is that to establish the forest timber volume spectral estimation model according to the relationship between forest reserves and the reflectance spectral characteristics of forest. By comparison it is found that these two methods are very different form the application point of view. The first method cannot make use of other methods of imaging data, as long as accessing several key data, even of the field measurements will be able to conduct operations. For areas that are not very rich, the method is fairly practical, and it can be widely promoted. But of you are not using image, the method is a very time-consuming and arduous task. Whereas the second must use image data. It has great value for large-scale rapid estimates of wood reserves.With the help of distribution index, the paper makes analysis of forestry types and spatial distribution from three aspects, such as elevation, slope and aspect angle. The research results showed as follows: Elevation is an important factor effecting forest type and distribution of the stand. Along with the altitude increasing, the atmospheric, humidity, temperature will change obviously. Gradient leads to different soil fertility, that is the proportions of input and output are different. Aspect impacts on the hours of sunshine (solar radiation energy distribution) and the sub-soil moisture. Therefore these three factors illustrated above work together and result in different distribution on different stand height, slope and aspect angle.更多还原