欢迎访问林业科学,今天是

林业科学 ›› 2002, Vol. 38 ›› Issue (2): 61-67.doi: 10.11707/j.1001-7488.20020211

• 论文及研究报告 • 上一篇    下一篇

荒漠化程度评价高光谱遥感信息模型

范文义   

  1. 东北林业大学,哈尔滨150040
  • 收稿日期:2001-07-12 修回日期:1900-01-01 出版日期:2002-03-25 发布日期:2002-03-25

HYPERSPECTRAL REMOTE SENSING INFORMATION MODEL FOR DISERTIFICATION DEGREE ASSESSMENT

Fan Wenyi   

  1. Northeast Forestry University Harbin150040
  • Received:2001-07-12 Revised:1900-01-01 Online:2002-03-25 Published:2002-03-25

摘要:

本文用国产高光谱分辨率成像光谱仪系统数据对荒漠化评价建立定量化遥感信息模型。对荒漠化评价因子中的主要定量因子(植被盖度、生物量和土壤含水率)进行了定量反演;对难于进行定量计算的评价指标,先通过目视解译获得各因子的编码图,分别进行影像化后参加荒漠化程度评价遥感信息模型计算。通过每个像元都可获取全部评价因子的指标值,在现有的荒漠化评价方法的基础上,建立以像元为单位的荒漠化程度评价的定量化遥感信息模型并输出荒漠化程度分布图。结果表明,用高光谱数据定量反演荒漠化地区植被生物量、盖度和土壤含水率是比较可靠的。当反演区域内灌木和草地同时存在时多项式模型的精度要明显高于线性模型;当植被类型单一时,模型即为较高精度的线性模型,但模型的应用地域范围受到限制,只能分块进行计算。因此,在只有灌木和草地的区域用多项式模型反演会提高效率。土壤含水量的反演方法适合于地形平坦、植被比较稀疏的条件。但研究发现,基于土壤热惯量的含水量模型具有一定的抗植被干扰能力。荒漠化程度评价的遥感信息模型的精度主要取决于现有荒漠化评价的方法(即评价指标是否科学合理、专家给定的权重和等级标准是否客观)以及各指标数据的获取精度。

关键词: 高光谱分辨率, 荒漠化评价, 遥感信息模型

Abstract:

Quantitative RSIM (Remote Sensing Information Model)was improved to evaluate the desertification degrees by using the data of state produced hyperspectral resolution imagining spectrometer, and the mainly quantitative factors on desertification assessment were retrieved which including the vegetation cover, the biomass and the soil water content. For the indexes which are difficult to count, the recoding maps based on the visual interpretation are obtained and then imaged respectively to be used in the RSIM. Every pixel in image can be used to acquire the indexes of all the evaluative factors. Based on the current methods on desertification assessment, quantitative RSIM on the basis of the pixel is developed and the distributing map of desertification degree is plotted in this paper. The result shows that it is relied to retrieve quantitatively the vegetation cover, the biomass and the soil water content of the desert area by the data of hyperspectral resolution imagining spectrometer, when there are both the shrub and the grassland in the retrieved region, the precision of the polynomial model is obvious higher than that of the linear model, contrastingly when the type of the vegetations is simplification, the linear model has the higher precision but limited applied range and can only be applied in block. So the retrieval efficiency can be improved by using the polynomial model in the region having only the shrub and the grassland. The method retrieving the soil water content is suitable to the flat area with sparse vegetations, at the same time the research shows that the model based on the soil thermal inertia stands against the interference of vegetations. On the whole, the precision of the desertification assessment RSIM lies on the data acquired precision and the current methods on desertification assessment that is to say whether the appraisable indexes are rationale and scientific or not and whether the weigh and the grade criteria that the experts give are objective or not.

Key words: Hyperspectral resolution, Desertification assessment, Remote Sensing Information Model