The Hyperspectral Inversion Method of Heavy Metal Contents in Cultivated Soils Based on GA-SVM
-
摘要: 为提高耕地土壤重金属含量高光谱反演模型精度,以岳阳县某地区耕地土壤重金属铁(Fe)、砷(As)、铬(Cr)为例,提出了一种遗传算法(GA)优化支持向量机(SVM)的重金属含量反演模型。在对光谱进行SG平滑和10 nm重采样后,利用一阶/二阶微分、倒数对数和连续统去除光谱变换方法增强光谱特征,通过相关性分析筛选最优变换光谱,使用皮尔森相关系数与主成分分析提取各重金属光谱特征变量,分别建立SVM和GA-SVM土壤重金属高光谱反演模型并进行精度验证。结果表明,二阶微分变换光谱与各重金属含量相关性整体最突出;三种重金属在可见光波段490 nm、500 nm、510 nm和530 nm具有共同敏感特征;经GA算法优化SVM参数后,对比SVM回归模型,预测精度有明显提高,其重金属Fe、As和Cr的验证集R2分别为0.968、0.821和0.976;研究结果可为应用遥感技术反演耕地土壤重金属含量提供新的参考。Abstract: In order to improve the accuracy of hyperspectral inversion model for heavy metal content in cultivated soil, a genetic algorithm (GA) optimized support vector machine (SVM) was proposed to retrieve the heavy metal content of cultivated soil in a certain area of Yueyang County. After SG smoothing and 10 nm resampling, the first-order / second-order differential, reciprocal logarithm and continuum removal spectral transformation methods were used to enhance the spectral characteristics. The optimal transform spectra were selected by correlation analysis. Pearson correlation coefficient and principal component analysis were used to extract the spectral characteristic variables of heavy metals. SVM and GA-SVM were used to establish soil heavy metal hyperspectral inversion models and their accuracies were verified. The results showed that the correlation between the second-order differential transform spectra and the contents of heavy metals was the most prominent. The visible light bands of 490 nm, 500 nm, 510 nm and 530 nm were the most prominent compared with the SVM regression model. The prediction accuracy was significantly improved, and the verification set R2 values of Fe, As and Cr were 0.968, 0.821 and 0.976, respectively. The research results could provide a new reference for the application of remote sensing technology to retrieve the content of heavy metals in cultivated soil.
-
表 1 土壤重金属含量描述性统计
Table 1. Descriptive statistics of soil heavy metal contents
重金属
Heavy metal最小值
Minimum最大值
Maximum均值
Average标准差
Standard deviation变异系数
Coefficient of variationFe(g kg−1) 4.76 21.31 13.957 2.933 0.210 As(mg kg−1) 4.72 12.96 9.053 1.627 0.180 Cr(mg kg−1) 5.96 24.10 14.298 4.060 0.284 表 2 皮尔森相关系数特征波段提取
Table 2. Pearson correlation coefficient feature band extraction
重金属
Heavy metal特征波段数
Feature band
number特征波段(nm)
Characteristic bandFe 17 460、490、500、510、530、570、580、590、740、750、770、790、890、1700、2200、2220、2270 As 8 440、490、500、510、530、540、1010、2120 Cr 9 460、490、500、510、530、540、570、580、590 表 3 SD变换光谱主成分特征值、贡献率与累计贡献率
Table 3. Principal component characteristic value, contribution rate and cumulative contribution rate
成分
IngredientPC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 PC11 PC12 特征值 44.871 16.754 11.059 8.885 5.257 3.225 2.842 2.551 2.187 1.755 1.289 1.086 贡献率 (%) 41.547 15.513 10.240 8.227 4.867 2.986 2.632 2.362 2.025 1.625 1.193 1.005 累计贡献率 (%) 41.547 57.06 67.299 75.526 80.393 83.379 86.011 88.373 90.398 92.023 93.216 94.221 注:SD为二阶微分光谱变换数据;PCi(i = 1,2,$3\cdots12 $)表示主成分数。 表 4 SVM回归模型估测
Table 4. SVM regression model estimation
重金属
Heavy metalPCA-SVM PCC-SVM 建模集R2
Calibration set R2RMSE 验证集R2
Validation set R2RMSE 建模集R2
Calibration set R2RMSE 验证集R2
Validation set R2RMSE Fe 0.766 0.090 0.737 0.096 0.864 0.070 0.831 0.079 As 0.716 0.117 0.447 0.128 0.668 0.127 0.469 0.128 Cr 0.764 0.123 0.713 0.113 0.691 0.133 0.533 0.135 表 5 GA-SVM回归模型估测
Table 5. GA-SVM regression model estimation
重金属
Heavy metalPCA-GA-SVM PCC-GA-SVM 建模集R2
Calibration set R2RMSE 验证集R2
Validation set R2RMSE 建模集R2
Calibration set R2RMSE 验证集R2
Validation set R2RMSE Fe 0.951 0.057 0.963 0.049 0.964 0.045 0.968 0.046 As 0.884 0.092 0.821 0.084 0.772 0.102 0.638 0.105 Cr 0.982 0.047 0.976 0.049 0.824 0.102 0.6 0.122 -
[1] 郭云开, 周烽松, 丁美青, 等. 水稻冠层与土壤高光谱反演土壤重金属对比研究[J]. 遥感信息, 2017, 32(2): 173 − 179. doi: 10.3969/j.issn.1000-3177.2017.02.026 [2] 张东辉, 赵英俊, 陆冬华, 等. 高光谱在土壤重金属信息提取中的应用与实现[J]. 土壤通报, 2018, 49(1): 31 − 37. [3] 李媛媛, 李 微, 刘 远, 等. 基于高光谱遥感土壤有机质含量预测研究[J]. 土壤通报, 2014, 45(6): 1313 − 1318. [4] 黄长平, 刘 波, 张 霞, 等. 土壤重金属Cu含量遥感反演的波段选择与最佳光谱分辨率研究[J]. 遥感技术与应用, 2010, 25(3): 353 − 357 + 341. doi: 10.11873/j.issn.1004-0323.2010.3.353 [5] Choe E, Meerf V D, Ruiteenbeek F V, et al. Mapping of heavy metal pollution in stream sediments using combined geochemistry, field spectroscopy, and hyperspectral remote sensing: a case study of the Rodalquilar mining area[J]. Remote Sensing of Environment, 2008, 112(7): 3222 − 3233. doi: 10.1016/j.rse.2008.03.017 [6] 郭云开, 钱 佳, 蒋 明, 等. 基于核偏最小二乘的矿区土壤Cu含量高光谱反演[J]. 土壤通报, 2019, 50(1): 52 − 56. [7] 龚绍琦, 王 鑫, 沈润平, 等. 滨海盐土重金属含量高光谱遥感研究[J]. 遥感技术与应用, 2010, 25(2): 169 − 177. doi: 10.11873/j.issn.1004-0323.2010.2.169 [8] 涂宇龙, 邹 滨, 姜晓璐, 等. 矿区土壤Cu含量高光谱反演建模[J]. 光谱学与光谱分析, 2018, 38(2): 575 − 581. [9] Vapnik V. The nature of statistical learning theory[M]. New York: Springer-Verlag, 1995: 138 - 145. [10] 王晓玲, 杜培军, 谭 琨. 高光谱遥感影像SVM分类中训练样本选择的研究[J]. 测绘科学, 2011, 36(3): 127 − 129. [11] 鄢文苗, 任 东, 黄应平, 等. 基于SVM土壤重金属污染评价的训练数据集构建[J]. 武汉大学学报(理学版), 2019, 65(3): 316 − 322. [12] 沈 强, 张世文, 夏沙沙, 等. 基于支持向量机的土壤有机质高光谱反演[J]. 安徽理工大学学报(自然科学版), 2019, 39(4): 39 − 45. [13] 吕 杰, 刘湘南. 利用支持向量机构建水稻镉含量高光谱预测模型[J]. 应用科学学报, 2012, 30(1): 105 − 110. doi: 10.3969/j.issn.0255-8297.2012.01.016 [14] Wang X J, Zhang Z P. A comparison of conditional simulation, Kriging and trend surface analysis for soil heavy metal pollution pattern analysis[J]. Journal of Environmental Science & Health Part A, 1999, 34(1): 73 − 89. [15] Xu M X, Wu S H, Zhou S L, et al. Hyperspectral reflectance models for retrieving heavy metal content: application in the archaeological soil[J]. Journal of Infrared and Millimeter Waves, 2011, 30(2): 109 − 114. [16] 王 璐, 蔺启忠, 贾 东, 等. 基于反射光谱预测土壤重金属元素含量的研究[J]. 遥感学报, 2007, 25(6): 906 − 913. doi: 10.11834/jrs.200706122 [17] 陈红艳, 赵庚星, 张晓辉, 等. 去除水分影响提高土壤有机质含量高光谱估测精度[J]. 农业工程学报, 2014, 30(8): 91 − 100. doi: 10.3969/j.issn.1002-6819.2014.08.011 [18] 徐明星, 吴绍华, 周生路, 等. 重金属含量的高光谱建模反演: 考古土壤中的应用[J]. 红外与毫米波学报, 2011, 30(2): 109 − 114. [19] 滕 靖, 何政伟, 倪忠云, 等. 西范坪矿区土壤铜元素的高光谱响应与反演模型研究[J]. 光谱学与光谱分析, 2016, 36(11): 3637 − 3642. [20] 古力扎提·艾买提, 阿不都拉·阿不力孜, 茹克亚·沙吾提, 等. 准东煤田土壤铅含量高光谱估算[J]. 土壤通报, 2018, 49(5): 1233 − 1239. [21] 章 琼, 郭云开, 钱 佳, 等. 像元二分模型的土壤重金属Cr的遥感反演[J]. 测绘科学, 2020, 45(4): 58 − 63. -