北方农业学报 ›› 2023, Vol. 51 ›› Issue (2): 116-125.doi: 10.12190/j.issn.2096-1197.2023.02.14

• 水产·农业信息技术 • 上一篇    下一篇

基于时序高光谱遥感数据的地表覆盖物多维光谱特征分析——以新疆阿拉尔市垦区为例

卢吉瑞1,2, 陈万基1,2, 赵阳1,2, 张仁凤1   

  1. 1.伊犁师范大学 资源与生态研究所/伊犁师范大学 生物与地理科学学院,新疆 伊宁 835000;
    2.新疆伊犁河谷资源植物保护与利用重点实验室,新疆 伊宁 835000
  • 收稿日期:2023-03-06 出版日期:2023-04-20 发布日期:2023-07-05
  • 通讯作者: 赵 阳(1988―),男,讲师,硕士,主要从事资源与环境遥感方面的研究工作。
  • 作者简介:卢吉瑞(2001—),男,本科生,专业为地理科学。
  • 基金资助:
    伊犁师范大学资源与生态研究所开放课题(YLNURE202208); 伊犁师范大学大学生创新创业训练计划项目(X20210764011)

Multidimensional spectral characteristics analysis of land cover based on time series hyperspectral remote sensing data—Taking the reclamation area of Alar City,Xinjiang as an example

LU Jirui1,2, CHEN Wanji1,2, ZHAO Yang1,2, ZHANG Renfeng1   

  1. 1. Institute of Resources and Ecology/College of Biology and Geographical Sciences,Yili Normal University,Yining 835000,China;
    2. Yili Valley Key Laboratory of Plant Protection and Utilization of Resources,Xijiang,Yining 835000,China
  • Received:2023-03-06 Online:2023-04-20 Published:2023-07-05

摘要: 【目的】分析新疆南部主要农作物光谱特征,为该地区大范围农作物的识别监测、精准提取等提供便捷。【方法】基于珠海一号高光谱卫星遥感数据对新疆南部阿拉尔市垦区地表覆盖物的光谱特征进行分析。通过端元光谱特征分析提取信息,构建二维(波长-反射率)光谱空间模型和三维(时相-波长-反射率)光谱空间模型,分析垦区4—10月棉花、玉米等农作物的光谱特征。【结果】根据不同时相相同区域的光谱曲线特征判断出了垦区农田所种植农作物的缓苗期、旺盛期、结果期等生长周期;通过计算地表覆盖物光谱曲线平均值,构建了平均光谱曲线,反映出其在特定时间段内总体反射率特点,棉花整体反射率高于玉米和枣树,红外波段区间水稻的吸收能力明显较强,沙地、水体光谱曲线趋于平滑稳定;以时间为基础,最大值光谱反射率曲线可突出农作物的特性,棉花反射率较高,玉米反射率偏低,水稻反射率相对稳定。在植被生长周期内单个时间点的光谱反射率切面曲线均遵循植被光谱曲线特征,可见光区间为波谷,波峰出现在近红外区间,棉花、枣树、玉米和水稻等主要农作物在三维光谱空间模型中,波谱曲面整体趋势相似;沙地、水体整体反射率不高,多集中于25%以下,其光谱反射曲线各具特点。【结论】新疆南部阿拉尔市垦区的农作物在不同生长期内,不同波长所对应的反射率不同;在植被缓苗期可见光波段对光的吸收能力均偏弱,反射率较高;在植被旺盛期可见光波段对光的吸收能力均偏强,反射率较低。沙地较农作物反射率相对稳定,水体反射率整体变化幅度不大且数值不高。

关键词: 高光谱, 新疆南部, 地表覆盖物, 光谱特征, 反射率, 波谱曲线

Abstract: 【Objective】To provide convenience for the identification,monitoring and precise extraction of large-scale crops in the region by analyzing the spectral characteristics of major crops in southern Xinjiang.【Methods】Based on the remote sensing data of Zhuhai No.1 hyperspectral satellite,the spectral characteristics of land cover in reclamation area of Alar City,southern Xinjiang were analyzed. Two dimensional(wavelength-reflectance) spectral spatial models and three dimensional(phase-wavelength-reflectance) spectral spatial models were constructed by extracting information from analyzed end-member spectrum features,then analyzed the spectral characteristics of crops such as cotton and corn in the reclamation area from April to October.【Results】Based on the spectral curve characteristics of the same region at different times,the growth cycles of crops planted in the farmland of the reclamation area including the slow seedling stage,peak stage,and fruiting stage were determined. By calculating the mean value of land cover spectral curve,the average spectral curve was constructed to reflect its overall reflectance characteristics in a specific period of time. The overall reflectance of cotton was higher than that of corn and jujube trees. The absorption capacity of rice was noticeably stronger in the infrared band range. The spectral curves of sandy land and water body tend to be smooth and stable. Based on time,the maximum spectral reflectance curve could highlight the characteristics of crops,with cotton having higher reflectance,corn having lower reflectance,and rice having relatively stable reflectance. The spectral reflectance profile curves at a single time point during the vegetation growth cycle all followed the characteristics of vegetation spectral curves,with visible light range being valleys and peaks appearing in the near-infrared range. Major crops such as cotton,jujube,corn,and rice had similar overall spectral surface trends in the three dimensional spectral spatial model. The overall reflectance of sandy land and water body was not high,mostly concentrated below 25%,and their spectral reflectance curves had their own characteristics.【Conclusion】In different growth stages of crops in the reclamation area of Alar City in southern Xinjiang,the reflectance of different wavelengths was different. In the slow seedling stage,the light absorption capacity of visible light band was weak,and the reflectance was high. During the peak stage of vegetation,the light absorption capacity in the visible light band was relatively strong,and the reflectance was relatively low. Sandy land had relatively constant reflectance compared with that of crops. The overall variation range of water body reflectance was not large and the value was not high.

Key words: Hyperspectrum, Southern Xinjiang, Land cover, Spectral characteristics, Reflectivity, Spectral curve

中图分类号: 

  • S127