玉米叶片中铜离子含量的TSVA和LSVA遥感预测模型

发布时间：2018-06-22 14:16

  本文选题：光谱分析 + 植被重金属污染　； 参考：《江西农业大学学报》2017年06期

【摘要】：植被重金属污染的植物体内重金属元素含量反演方法一直是高光谱遥感研究热点之一。设置不同浓度铜离子(Cu~(2+))胁迫梯度下的玉米盆栽实验,并测量不同浓度Cu~(2+)胁迫下玉米叶片的光谱数据及其叶片中富集的Cu~(2+)含量。由于健康的与受Cu~(2+)胁迫污染的玉米叶片光谱在曲线形态上相似度仍很高,且传统的光谱相似性测度方法难以区分污染光谱的变异性弱差信息,因而采用离散小波变换多层分解、奇异值分解、光谱角度量等理论方法对光谱形态及变异信息进行转换处理,再通过正切与对数函数扩大光谱转换后的变异信息的方式,构建了正切奇异向量角(tangent singular vector angle,TSVA)和对数奇异向量角(logarithmic singular vector angle,LSVA)的玉米叶片中Cu~(2+)含量遥感预测模型。结果表明,TSVA和LSVA模型预测玉米叶片中的Cu~(2+)含量较为理想,也能很好地区分不同浓度Cu~(2+)胁迫梯度下的光谱变异信息。通过预测值与实测值的结果比较与相关性分析(其相关系数均大于0.91),验证了TSVA和LSVA模型预测玉米叶片中Cu~(2+)含量的有效性和可行性。
[Abstract]:The inversion method of heavy metal content in vegetation contaminated by heavy metals has been one of the hotspots in hyperspectral remote sensing. The pot experiment of maize under different concentration of Cu ~ (2) stress was carried out. The spectral data of maize leaves and the contents of Cu ~ (2) enriched in maize leaves were measured under different concentrations of Cu ~ (2) stress. Because the spectral similarity between healthy maize leaves and maize leaves contaminated by Cu2 stress is still very high, and the traditional spectral similarity measurement method is difficult to distinguish the weak difference information of pollution spectra. Therefore, the discrete wavelet transform multilayer decomposition, singular value decomposition, spectral angle quantity and other theoretical methods are used to transform and process the spectral morphology and variation information, and then the variation information after spectral transformation is expanded by tangent and logarithmic functions. A remote sensing prediction model of Cu2 content in maize leaves was established based on tangent singular vector angle (tangent singular vector anglev (tangent singular vector) and logarithmic singular vector angle (logarithmic singular vector anglev LSVA). The results showed that TSVA and LSVA models were ideal for predicting Cu2 content in maize leaves, and could also distinguish the spectral variation information under different Cu ~ (2) stress gradients. The validity and feasibility of TSVA and LSVA models in predicting Cu2 content in maize leaves were verified by comparing the predicted values with the measured values and the correlation analysis (the correlation coefficients were all greater than 0.91).
【作者单位】： 中国矿业大学(北京)地球科学与测绘工程学院;
【基金】：国家自然科学基金项目(41271436) 中央高校基本科研业务费专项资金(2009QD02)~~
【分类号】：S127;S513;X87
，

本文编号：2053097