新疆博斯腾湖湖滨绿洲不同土地利用类型土壤电导率高光谱估算  

Hyperspectral estimation of soil conductivity of different land use types in lakeside oasis of Bosten Lake in Xinjiang

作　　者：樊泳灼 李新国 FAN Yongzhuo;LI Xinguo(College of Geographical Sciences and Tourism/Laboratory of Lake Environment and Resources in Arid Regions of Xinjiang,Xinjiang Normal University,Urumqi 830054,China)

机构地区：[1]新疆师范大学地理科学与旅游学院/新疆干旱区湖泊环境与资源实验室,新疆乌鲁木齐830054

出　　处：《中山大学学报（自然科学版）（中英文）》2023年第6期31-39,共9页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：新疆维吾尔自治区自然科学基金(2022D01A214);国家自然科学基金(41661047)。

摘　　要：为了更精准了解博斯腾湖湖滨绿洲不同土地利用类型的土壤含盐量,应用竞争性自适应重加权采样(CARS)、连续投影算法(SPA)、竞争性自适应重加权-连续投影算法(CARS-SPA)3种方法筛选不同土地利用类型土壤电导率高光谱数据的特征波段,基于全波段和特征波段结合BP神经网络分别构建湖滨绿洲的耕地、林地、荒地和整体土地的土壤电导率估算模型,对比不同方式的估算模型精度。研究结果表明:1)耕地、林地、荒地及整体土地的土壤电导率均值分别为0.84、5.43、5.78、3.26 mS/cm。湖滨绿洲整体土地的电导率相比耕地平均值增加了2.42 mS/cm,相比林地和荒地减少了2.17、2.52 mS/cm。2)通过CARS-SPA方法可以降低输入模型的波段数,提高模型的效率,筛选耕地、林地、荒地及整体土地的土壤电导率的特征波段数仅占全波段的0.71%、0.59%、0.06%、1.00%。3)对耕地、林地、荒地的土壤电导率构建单独的估算模型明显提高了研究区土壤电导率的估算精度,在FDR-CARS-BP、FDR-SPA-BP、FDR-CARS-SPA-BP共3种模型中,耕地、林地、荒地土壤电导率建模的平均R2相比整体土地建模分别提高0.12、0.14、0.15,FDR-CARS-SPA-BP模型为研究区土壤电导率高光谱估算最优模型。To more accurately understand the soil salinity of different land use types in the lakeside oasis of Bosten Lake,Xinjiang,three methods of competitive adaptive reweighted sampling(CARS),successive projection algorithm(SPA),and competitive adaptive reweighted-successive projection algorithm(CARS-SPA)were applied to screen the characteristic bands of soil conductivity hyperspectral data of different land use types,based on the full band and characteristic bands.The estimation models of soil conductivity of the lakeshore oasis were constructed based on the full band and characteristic bands combined with BP neural network to compare the accuracy of estimation models in different ways.The results showed that:(1)the mean values of soil conductivity of cropland,forest land,wasteland,and overall land are 0.84,5.43,5.78,and 3.26 mS/cm,respectively;the overall soil conductivity of the lakeshore oasis is 2.42 mS/cm higher than the mean value of cropland,2.17 and 2.52 mS/cm lower than forest land and wasteland.(2)The CARS-SPA method can reduce the number of bands input to the model and improve the efficiency of the model.The number of characteristic bands for screening the electrical conductivity of cropland,forestland,wasteland,and overall land soil only accounts for 0.71%,0.59%,0.06%,and 1.00%of the full bands.(3)Constructing separate estimation models for soil conductivity of cropland,forest land,and wasteland significantly improves the estimation accuracy of soil conductivity in the study area.Among the three models of FDR-CARSBP,FDR-SPA-BP,and FDR-CARS-SPA-BP,the average R^(2) of soil conductivity modeling for cropland,forest land,and wasteland increased by 0.12,0.14,and 0.15,respectively,compared with the overall soil modeling,the FDR-CARS-SPA-BP model is the optimal model for hyperspectral estimation of soil conductivity in the study area.

关 键 词：土壤电导率 土地利用类型 高光谱数据 竞争性自适应重加权-连续投影算法 BP神经网络 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]