【作者】 强建华；

【导师】 赵鹏祥； 李卫忠；

【作者基本信息】 西北农林科技大学 ， 森林经理学， 2007， 硕士

【摘要】 本文以黄土高原天然林典型分布地区黄龙山林业局蔡家川林场(20726hm2)为研究对象,以该林场1997年与2004年两期TM遥感影像为主要数据源,以遥感(RS)技术为主要处理手段,辅以地理信息系统(GIS)技术,对该区域森林资源的动态变化情况进行了研究。研究成果主要包括以下内容:(1)依据森林资源二类调查资料,对该地区天然林林地类型信息的提取方法进行了研究。首先对TM影像的6个波段(除热红外波段6)进行主成分变换,以减小特征间的相关性,并将变换后影像的第一、第二和第三主成分分量进行RGB(红、绿、蓝通道)合成,这样得到的影像可以更好地表达地面信息;然后参考先验知识来纯化样本,以建立比较精确的分类器,在此基础上进行计算机监督分类,获取了研究区域各地类的面积。结果表明,最终分类精度可以达到92%,Kappa值达到0.906,说明在该地区的天然林区林地信息提取中,利用TM作为数据源,经过图像处理和监督分类能够达到较高的精度,可以满足林业生产的要求。(2)在遥感数字图像处理系统支持下,研究了该林场天然林资源的时空变化。结果表明,在1997年到2004之间,本林杨优势树种面积增加,其中油松增加了8.4%,杨树增加了6.7%,栎类增加了1.8%;同时,其水平空间分布也发生了一定的变化;森林覆盖率增加了约16.9%;林地类型之间发生了明显的动态变化,其中,55.5%的采伐地区域变成了油松林地,21.1%的采伐地区域变成了杨类林地。这些明显变化都与天然林保护工程的实施有着密切的联系。(3)以研究区油松天然林为研究对象,选择能最佳表达研究区油松林信息的植被指数以及植被指数与盖度之间的回归模型,从遥感影像计算NDVI,并获取研究区油松林的盖度信息,反映油松林的生长状态。结果表明,研究区油松林90%以上的覆盖度超过50%,说明蔡家川林场天然油松林的生长状态良好。并对研究成果经过实地验证,通过植被指数获得的油松林覆盖度信息与实地调查结果吻合度达92%,这说明利用植被指数进行油松天然林生长状态的评价是可行的,为林业部门对于天然林经营措施的制定提供依据。更多还原

【Abstract】 This Thesis studied on dynamic spatial-temproal changes of natural forest resources based on RS .CaiJiachuan,as the typical region of distribution of natural forests was the research object ,the TM images of 2 years of 1997and 2004 as the main data sources,remote sensing (RS) technology as the primary methods,supplemented by geographic information system (GIS) technology,Research achievements include the following elements :(1)Combine survey data to study the natural forest areas of woodland types of information extraction method. In this research,the TM image is performed the PCA transformation in order to reduce the correlation among the features. The new image is synthesized by the first、second、third principal components in the way of RGB after PCA. It obvious that new image is better one in expressing the ground information. With the reference knowledge we can purify the samples and build a more precise classification signature. Then,we perform supervised classification and compute areas of each type. Through evaluating classification accuracy,the results show that the method above is proper and the precise achievement can be obtained using it .The final datum can be used to update the information,In other hand,it can provide precise datum for making the thematic map of the forest.(2)With the survey data and under the support of IMAGE software ,we studied the dynamic spatial-temproal changes of this area,The result indicated,the area of superiority tree has increased from 1997 to 2004,for example ,the robur increased by 1.8%,Chinese pine increased by 8.4% ,poplar increased by 6.7%; Meanwhile,spatial distribution has changed;forest coverage ratio increased by approximately 16.9%;obvious dynamic change has been occurred,such as 55.5% of cutting land was covered by Chinese pine and 21.1% was covered by poplar. These obviously changes has the close relation with the implementation of the natural forest protection project.(3)By Chinese pine natural forest in loess plateau as research object,by the TM image as primarily data pool,choose the proper model and vegetation index that can best express the relation between vegetation index and forest coverage. Calculateing NDVI from the remote sensing image,and gain cover information of Chinese pine that reflection Chinese pine’s growth condition. The result indicated,above 90% of Chinese pine’s coverage in research area surpasses 50%,it explained that natural Chinese pine in Cai Jiachuan growed well. Through testing of the research results,coverage information from the NDVI is properly matching the coverage information from on-the-spot investigation,this indicated that evaluatating growth condition of Chinese pine natural forest with the vegetation index is feasible,and provides the basis for the forestry department regarding the natural forest management measure formulation.更多还原