| 基于土壤反射光谱特性的广东省稻田土壤快速分类 |
| |
| 引用本文： | 李丹,彭智平,韩留生,王重洋,刘尉,黄思宇,陈水森.基于土壤反射光谱特性的广东省稻田土壤快速分类[J].热带地理,2015,35(1):29-34. |
| |
| 作者姓名： | 李丹 彭智平 韩留生 王重洋 刘尉 黄思宇 陈水森 |
| |
| 作者单位： | （1．广州地理研究所，广州 510070；２．广东省农科学院 农业资源与环境研究所，广州 510640） |
| |
| 基金项目： | 2012年度广东省科学院青年科学研究基金(qnjj201205);国家自然科学基金青年基金项目(41301401);广东省科技厅星火计划项目(2012A020602088) |
| |
| 摘 要： | 选择广东省215个村镇稻田的土壤样本，首先利用ASD Field Spec3测量土壤样本在350~2 500 nm的光谱，并采用S-G一阶导数平滑滤波降低样本测量中光照差异的影响，然后将遗传算法（Genetic Algorithm，GA）和支持向量机分类（Support Vector Machine，SVM）分别用于提取分类光谱特征和建立分类模型，分别在土纲、亚纲、土类3个层次进行土壤分类。结果表明：1）在不同的分类层次下，与铁氧化物密切相关的650~710以及900 nm附近光谱，与羟基矿物吸收有关的2 207~2 237和2 377~2 397 nm区间均被作为分类特征变量。2）随着土壤类型的细分，分类所需变量增多。在土类级，对有机质敏感的2 080 nm附近的光谱也被引入分类定标模型中，土纲和亚纲下分类精度＞67%，土类级分类精度为58.67%。利用遗传算法提取光谱特征，进行基于支持向量机的土壤分类具有一定优势。
|
| 关 键 词： | 反射光谱 土壤分类 遗传算法 支持向量机 稻田 广东省 |
Rapid Soil Classification of Paddy Field in Guangdong Province Based on Visible and Near Infrared Reflectance Spectra |
| |
| LI Dan;PENG Zhiping;HAN Liusheng;WANG Chongyang;LIU Wei;HUANG Siyu;CHEN Shuisen.Rapid Soil Classification of Paddy Field in Guangdong Province Based on Visible and Near Infrared Reflectance Spectra[J].Tropical Geography,2015,35(1):29-34. |
| |
| Authors: | LI Dan;PENG Zhiping;HAN Liusheng;WANG Chongyang;LIU Wei;HUANG Siyu;CHEN Shuisen |
| |
| Institution: | （1．Guangzhou Institute of Geography，Guangzhou 510070，China；2．Institute of Agricultural Resources and Environment， Guangdong Academy of Agricultural Sciences，Guangzhou 510640，China） |
| |
| Abstract: | Soil classification is an important work of soil remote sensing. In this paper 215 soil samples collected from paddy field in Guangdong Province were studied. Firstly, the soil spectra of 350-2 500 nm were measured by using ASD Field Spec 3, and the S-G first derivative smoothing was applied to reduce the effects of illumination difference. Secondly, the Genetic Algorithm and Support Vector Machine were used to select the spectral features and construct the soil classification models, respectively. The soil classifications were implemented at 3 different levels (soil order, soil suborder and soil type). The results indicated that in all classification models, the spectra of 650-710 nm and around 900 nm closely related to ferriferous oxide and spectra around 2 207-2 237 and 2 377- 2 397 nm closely related to metal-OH were selected. And for finer soil classification, more spectral variables were used. Even at the soil type level, the spectral data around 2 080 nm sensitive to the variety of soil organic matter were also introduced in the calibration model. The agreement rates at levels of soil order and soil suborder were above 67%, and that at level of soil type was 58.67%. The Genetic Algorithm and Support Vector Machine presented some advantages in soil classification. |
| |
| Keywords: | reflectance spectra soil classification Genetic Algorithm Support Vector Machine paddy field Guangdong Province |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《热带地理》浏览原始摘要信息 |
| 点击此处可从《热带地理》下载免费的PDF全文 |
