论文部分内容阅读
潮滩土壤含水量具有变化频率快、空间变化大的特征,是影响潮滩地表反射率的重要因素。潮滩土壤含水量的精确提取,可为潮滩特征地物信息遥感反演提供基础。本文利用江苏大丰王港潮滩4种典型沉积物、449组不同含水量对应的实测光谱曲线数据进行特征分析,构建高光谱预测模型,实现了潮滩沉积物含水量的遥感反演。研究结果表明:(1)在短波红外波段,沉积物含水量与反射率之间存在良好的分段线性相关关系,分段点对应的含水量分别为42%和62%;(2)1165nm、1336nm、1568nm和1780nm特征波段反射率,对含水量变化具有良好响应,由特征波段组合计算得到的差值水指数DWI、比值水指数RWI和归一化水指数NDWI与含水量呈显著线性相关,可有效改善单波段反射率与含水量之间的分段线性关系;(3)3个水指数中,DWI反演的含水量精度优于RWI和NDWI,且对不同含水量大小均有良好适应性,而RWI和NDWI更适合含水量变化范围中等的情况;(4)对于粉砂、砂质粉砂、粉砂质砂和砂4种沉积物类型,DWI1336,1780验证组模拟含水量与实测含水量的相关系数,分别为0.891、0.915、0.920和0.905,均方根误差分别为9.87%、3.56%、4.24%和2.98%,表明由DWI构建的高光谱遥感反演模型,可有效实现潮滩表层含水量的时空变化预测。
The moisture content of tidal flat has the characteristics of fast changing frequency and large spatial variation, which is an important factor affecting the surface reflectance of tidal flat. Accurately extracting soil moisture content of tidal flat can provide a basis for remote sensing inversion of tidal flat feature data. In this paper, the characteristics of four typical sediments and 449 groups of corresponding spectral curves of Daping Wanggang tidal flat in Jiangsu Province were analyzed, and the hyperspectral prediction model was constructed to realize the remote sensing inversion of sediment moisture content. The results show that: (1) There is a good piecewise linear correlation between moisture content and reflectance in the shortwave infrared band, with water content corresponding to 42% and 62% respectively; (2) 1165nm, 1336nm, 1568nm and 1780nm, respectively, which have good response to the change of water content. The DWI, RWI and NDWI calculated by the combination of characteristic bands show a significant linear correlation with water content, Which can effectively improve the piecewise linear relationship between single-band reflectance and water content. (3) Among the three water indices, the accuracy of DWI inversion is better than that of RWI and NDWI, and it is well adapted to different water content While RWI and NDWI are more suitable for medium water content variation. (4) For the sediment types of silt, sandy silt, silty sand and sand, DWI1336 and 1780 validated the simulated water content and measured The correlation coefficients of water content were 0.891,0.915,0.920 and 0.905 respectively, and the root mean square errors were 9.87%, 3.56%, 4.24% and 2.98%, respectively. This indicated that the hyperspectral remote sensing inversion model constructed by DWI can effectively achieve tide Temporal and Spatial Prediction of Surface Water Content in Beach.