《基于高空间分辩率遥感影像的森林资源调查小班区划》由会员分享，可在线阅读，更多相关《基于高空间分辩率遥感影像的森林资源调查小班区划（58页珍藏版）》请在金锄头文库上搜索。
1、西安科技大学 硕士学位论文 基于高空间分辩率遥感影像的森林资源调查小班区划 姓名：张艮龙 申请学位级别：硕士 专业：地图学与地理信息系统 指导教师：贾建华 论文题目：基于高空间分辩率遥感影像的森林资源调查小班区划 专业：地图学与地理信息系统 硕 士 生：张艮龙（签名） 指导教师：贾建华（签名） 摘要 我国森林资源调查体系中用途最广的调查是森林资源规划设计调查，其最重要的内 容之一就是小班区划。这是因为森林空间分布定位的准确性、面积测量的精确性和最终 资源成果的可靠性，都直接受到小班区划结果的影响。 当前森林资源调查小班区划主要是在现地进行对坡勾绘，这种方法存在人为主观 性，不够科学客观。本文首
2、先分析了 SPOT-5 遥感数据的特点以及现行小班区划中所存 在因人的视野小而产生误判的缺点，然后采用了面向对象的多尺度分割策略，针对不同 尺度的地物信息采用不同的分割尺度，在此基础上构建多尺度分割等级网，实现了地物 信息的分层提取。 其具体流程为在小班边界控制条件下，对SPOT-5 影像进行初始分割， 并输出矢量化结果，建立“对象”的拓扑关系（邻接关系和连通关系），为建立多尺度 分割数据间的联系奠定基础。此时，对影像分割结果进行地块类型识别，形成初步的小 班区划结果。然后从已矢量化的地块类型图中提取出有林地，在有林地斑块控制范围内 的影像进行分割，并进行矢量化。最后在有林地边界控制范围内的矢
3、量化结果与小班区 划初步结果叠加，形成小班区划第二次结果，如果对第二次小班区划结果不满意，使用 小班区划模块对边界进行手工调整，形成最后的小班边界。 此外，所构建的森林资源规划设计调查数据内业生成系统基础平台主要实现了：文 件管理及视图、图像基本操作、图像预处理、图像边界提取、图像分类模块、小班区划 模块、帮助等模块。为小班区划技术路线的实施提供了一个很好的平台。 基于该系统对吉林省汪清县金沟岭林场的森林资源进行小班区划，实验结果表明该 技术路线取得了较好的预期效果，是对计算机自动识别和小班区划的一个新的尝试。 关 键 词：SPOT-5 遥感数据；森林资源调查；面向对象；多尺度分割；小班区划
4、研究类型：应用研究 SubjectSubjectSubjectSubject：BasedBasedBasedBased onononon H H H Highighighigh S S S Spatialpatialpatialpatial R R R Resolutionesolutionesolutionesolution R R R Remoteemoteemoteemote S S S Sensingensingensingensing I I I Imagesmagesmagesmages ofof ofof F F F Forestorestorestorest R R R Res
5、ourcesesourcesesourcesesources I I I Investigationnvestigationnvestigationnvestigation S S S Subububub- - - -compartmentcompartmentcompartmentcompartment D D D Divisionivisionivisionivision SpecialtySpecialtySpecialtySpecialty：CartographyCartographyCartographyCartography andandandand GeographicGeogr
6、aphicGeographicGeographic InformationInformationInformationInformation SystemSystemSystemSystem NameNameNameName：ZhangZhangZhangZhang Gen-longGen-longGen-longGen-long（SignatureSignatureSignatureSignature） InstructorInstructorInstructorInstructor ：JiaJiaJiaJia JianJianJianJian- - - -huahuahuahua（Sign
7、atureSignatureSignatureSignature） ABSTRACT The forest resources planning and design surveys was widely used in the survey system of forest resources in our Country, one of the most important is the sub-compartment division. The positioning accuracy of the spatial distribution of forest, the accuracy
8、 of area measurement and the reliability of the final resource, were directly affected by the sub-compartment division. The main method of current forest resources sub-compartment division is adopted drawing division boundary according to the condition of opposite hillside in the field, which existe
9、nce of human subjectivity and lacks of scientific objectivity in this approach. In this paper, firstly, we analyzed the characteristics of Spot-5 remote sensing data and the shortcomings of the sub-compartment division arising from small field of vision due to miscalculation in current forest. Secon
10、dly, this paper brings forward the strategy of object- oriented multi-scale segmentation, which takes different segmentation scales according to the feature information of different scales. Based on this, the multi-scale segmentation network is established and the layered extraction of the feature i
11、nformation is realized. The specific process, under the conditions of the control in sub-compartment border for the initial segmentation of the SPOT-5 images, and output the results of vector, a object of the topological relations (adjacency and connectivity relationships), for the establishment of
12、multi- scale segmentation of data links between the foundation. At this point, identification the results of the segmentation block type, and format the initial results of sub-compartment division. Then, extract forest land from the vector of the block type diagram, in a forest patch within the cont
13、rol of the image segmentation and vectorization. Finally, within the forest land border controls of Vector-based results and the preliminary results of superposition of sub- compartment to form the second division results in sub-compartment,ifnot satisfied with the second division results in sub-com
14、partment, using sub-compartment division modules to manually adjust the boundary to form the final sub-compartment border. In addition, the forest resourceinventory data base platform production within the industry system mainly achieved: the document management and view images of basic operations,
15、image preprocessing, edge extraction, image classification, sub-compartment division, help models. The system provides a good platform for the implementation of sub- compartment division. Based this system carry on forest resources sub-compartment division for Wangqing county Jingoulin farm, experim
16、ental results show that the method achieved fairly good results. This is a new attempt of theautomatic identification and sub-compartment division. KeyKeyKeyKeywordswordswordswords：SPOT-5 Remote Sensing DataForestResource InventoryObject-oriented Multi-scale SegmentationSub-compartment Division ThesisThesisThesisThesis：ApplicationResearch 1 绪论 1 1 绪论 1.1 论文的研究意义 近几年高分辨率遥感卫星以其独特的优越性(信息准确、现势性强、制图周期短、 信息来源不受国境线以及环境、地面条件等的制约)为森林资源调查，环境监测和规划 管理等行业提供了快速、可靠的数
