| 李艳茹,杨可明,韩倩倩,高伟,张建红.基于SD-SWT的铜胁迫下玉米光谱奇异性甄别与污染监测[J].农业环境科学学报,2020,39(9):1869-1877. |
| 基于SD-SWT的铜胁迫下玉米光谱奇异性甄别与污染监测 |
| Spectral singularity identification and pollution monitoring of corn under copper stress based on SD-SWT |
| 投稿时间：2020-04-24 |
| DOI：10.11654/jaes.2020-0468 |
| 中文关键词: 玉米 铜污染 光谱奇异分析 离散平稳小波变换 小波奇异指数 逐步多元线性回归 |
| 英文关键词: corn copper pollution spectral singularity analysis discrete stationary wavelet transform wavelet singularity indexes stepwise multiple linear regression |
| 基金项目:国家自然科学基金项目（41971401）；中央高校基本科研业务费专项资金项目（2009QD02） |
|
| 摘要点击次数: 2002 |
| 全文下载次数: 1522 |
| 中文摘要: |
| 为甄别重金属铜（Cu）胁迫下玉米光谱的弱差及奇异信息以监测玉米受Cu污染的程度，于2017年设置多浓度Cu胁迫下玉米培株盆栽实验，测定玉米叶片反射光谱和Cu含量数据，将一阶光谱微分（SD）和离散平稳小波变换（SWT）相结合，定义并提取小波奇异指数（WSI），进行光谱的奇异性甄别，并与常规的光谱特征参数进行对比；结合逐步多元线性回归（SMLR）算法，构建玉米叶片Cu含量的WSI-SMLR反演模型，同时利用不同年份采集的玉米叶片反射光谱和Cu含量数据验证反演模型的可行性及稳定性，并与一些已有的类似研究成果进行对比。结果表明：相比于常规的光谱特征参数，WSI与玉米叶片中的Cu含量有更显著的相关性及线性关系，可用来监测玉米叶片中的Cu含量变化；与一些已有的类似研究成果相比，WSI-SMLR模型反演玉米叶片中Cu含量的精度更高且更稳定。研究验证了小波奇异指数在监测玉米Cu污染方面具有有效性和优越性，为监测农作物重金属污染提供了新的光谱奇异指数与技术方法。 |
| 英文摘要: |
| An initial pot experiment of a corn culture plant under the stress of multi-gradient Cu was set up in 2017 to monitor the degree of corn polluted Cu. The reflectance spectra and Cu content data of corn leaves were measured to identify the weak difference and singular information of corn spectra under the stress of the heavy metal Cu. Wavelet singularity indexes(WSI)were defined and extracted by combining first-order spectral derivative(SD)and discrete stationary wavelet transform(SWT)to identify spectral singularity compared to the conventional spectral characteristic parameters. The model of WSI-stepwise multiple linear regression(SMLR)to retrieve the Cu content in corn leaves was built by combining these algorithms. Meanwhile, the feasibility and stability of the retrieving model were verified using the reflectance spectra and Cu content data of corn leaves collected over the past year and compared with similar research results. The results showed that, compared with the conventional spectral characteristic parameters, WSI had a more significant correlation and linear relationship with the Cu content in corn leaves and could be used to monitor changes in the Cu content in corn leaves. Compared with similar research results, the WSI-SMLR model exhibited higher accuracy and stability in retrieving the Cu content in corn leaves. Therefore, wavelet singular indexes are effective and superior in monitoring the Cu pollution of corn, providing new spectral singular indexes and technical methods to monitor heavy metal pollution in crops. |
| HTML
查看全文
查看/发表评论 下载PDF阅读器 |
