论文部分内容阅读
运用Hyperion数据,以黑龙江省大庆市某一实验区为例,开展对土壤含盐量定量提取的研究,通过对图像预处理、特征提取、建立BP神经网络模型(Back Propagation Network)等研究工作,探讨反演土壤含盐量的方法。研究结果表明:神经网络模型具有极强的线性和非线性拟合能力,模拟遥感影像特征与土壤盐分之间比较复杂的关系上有很大优势。研究结果不但为利用Hyperion数据反演土壤含盐量提供理论依据,而且还为其它地表参数的反演提供参考。
Using Hyperion data, taking a certain experimental area in Daqing City of Heilongjiang Province as an example, this paper carried out a quantitative extraction of soil salt content. Through the research on image preprocessing, feature extraction, and establishment of BP neural network model (Back Propagation Network) Discussion on the Method of Retrieving Soil Salinity. The results show that the neural network model has strong linear and non-linear fitting ability, and it has great advantages to simulate the relationship between remote sensing image features and soil salinity. The results not only provide theoretical basis for the inversion of soil salinity by Hyperion data, but also provide references for the inversion of other surface parameters.