基于3S技术的都兰县土地利用变化及枸杞地光谱识别研究

发布时间：2018-03-08 02:29

本文选题：都兰县　切入点：土地利用　出处：《青海大学》2017年硕士论文　论文类型：学位论文

【摘要】：土地利用是人类对土地长期、动态的利用、改造。土地利用变化不仅直接影响社会经济的可持续发展,同时还对当今生态资源环境产生重要的影响。研究土地利用变化对土地,特别是农业用地的综合利用以及协调经济生产和土地生态之间的矛盾有很大的作用。利用遥感和地理信息系统技术对土地利用进行综合研究仍是当下土地学科研究的热点之一。所以,本研究在都兰县2000~2015年的时间尺度范围内,根据土地利用分类标准,利用动态度、转移矩阵分析该区土地利用变化,在此基础上进一步尝试利用基于光谱信息的建模方法识别枸杞,辨别出枸杞地与其他相近地类的不同,从而得出都兰县枸杞的种植区域。结果表明:都兰县六种土地利用类型中,未利用地的面积最大,约占全县总面积的一半;未利用地和草地面积之和约占全县总面积的80%。2000~2015年间未利用地、农田面积逐年减少;草地、林地、建设用地面积增长明显;水域面积保持稳定;其中面积变化较大的是未利用地(减少506.97km2)、草地(增长316.85km2)以及林地(增长208.22km2)。在各类型转移变化矩阵中,林地转入率最大;未利用地、草地的转入及转出面积相对其他类型是比较大的,而水域与大部分类型均无转移关系。此外,利用全县植被覆盖情况对比全县土地利用现状,验证了植被覆盖区与草地、林地及农田分布大致相同。采集Landsat遥感影像以及野外实地测量的枸杞光谱信息,在ENVI软件下,在最小噪声分离(Minimum Noise Fraction)和纯净像元指数(Pixel Purity Index)方法的测算下,目标像元多为非纯净像元,这与实际上枸杞种植不能完全覆盖是一致的。再根据两者间光谱信息的关系并调用光谱信息散度(Spectral Information Divergence)和混合调谐匹配滤波(Mixture Tuned Matched Filtering)方法,建立两种模型。其中以实测数据与对应的影像数据建立方程模型,两者在数值上的关系是显著相关的,但在反演视图上表现可能受同谱异物及非纯净像元的影响。此外,基于SID和MTMF方法建立的模型在反演视图上可以较为清晰的反映出主要的枸杞种植区域,并能与未利用地、林地及农田等土地利用类型区别开来。
[Abstract]:Land use is the long-term, dynamic use and transformation of land. Land use change not only directly affects the sustainable development of social economy, but also has an important impact on the ecological resources and environment. In particular, the comprehensive utilization of agricultural land and the coordination of the contradiction between economic production and land ecology play a very important role. The comprehensive study of land use using remote sensing and GIS technology is still the present scientific research of land science. One of the hot spots. So, According to the land use classification standard, the paper analyzes the land use change in Dulan County from 2000 to 2015 by using the dynamic degree and the transfer matrix. On this basis, we further try to use the modeling method based on spectral information to identify Lycium barbarum and distinguish the difference between Lycium barbarum and other similar ground. The results show that: among the six types of land use in Dulan County, the area of unused land is the largest, accounting for about half of the total area of the whole county; The sum of unused land and grassland area accounted for 80.2000m of the total area of the county from 2015 to 2015, the farmland area decreased year by year, the area of grassland, woodland and construction land increased obviously, the area of water area remained stable; The most significant changes in the area were in unused land (by 506.97km2), grassland (increase by 316.85km2) and forestland (increased by 208.22km2). Compared with other types, the area of grassland was larger than other types, and there was no relation between the water area and most of the types. In addition, compared with the present situation of land use in the whole county, the vegetation cover area and grassland were verified. The distribution of forest land and farmland is about the same. The spectral information of Lycium barbarum collected from Landsat remote sensing image and field measurement is calculated by the method of minimum noise separating minimum Noise Fractionation and pure pixel index (Pixel Purity Index) under ENVI software. Most of the target pixels are non-pure pixels, which is consistent with the fact that Lycium barbarum planting can not be completely covered. Then according to the relationship between the spectral information and spectral information divergence, the spectral information divergence and mixed tuned matched filter mix Tuned Matched filtering method are called. Two kinds of models are established, in which the equation model is established from the measured data and the corresponding image data. The relationship between the two is significant in numerical terms, but it may be influenced by the isospectral foreign body and the impure pixel in the inversion view. The model based on SID and MTMF method can clearly reflect the main planting areas of Lycium barbarum in the inversion view and can be distinguished from the land use types such as unused land, forest land and farmland.
【学位授予单位】：青海大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F301.2;S567.19

【相似文献】