文献类型：期刊文献

中文题名：风蚀地表颗粒含量提取技术——基于决策树图像分类

英文题名：Extraction of the Wind Erosion Surface Particle Content——Based on Decision Tree Classification Algorithms

作者：王淮亮[1] 李玉宝[2] 高君亮[3] 高永[1] 虞毅[4] 王鹿振[5]

第一作者：王淮亮

机构：[1]内蒙古农业大学生态环境学院;[2]温州大学生命与环境科学学院;[3]中国林业科学研究院沙漠林业实验中心;[4]国际竹藤网络中心;[5]青海省水利水电勘测设计研究院

年份：2013

卷号：35

期号：2

起止页码：134-137

中文期刊名：农机化研究

外文期刊名：Journal of Agricultural Mechanization Research

收录：北大核心:【北大核心2011】；

基金：林业公益性行业科研专项项目(201204205);内蒙古农业大学防沙治沙科研创新团队资助计划项目(NDTD2010-11)

语种：中文

中文关键词：地表颗粒;图像分类;决策树

外文关键词：particles of the wind erosion surface; digital image processing; decision tree classification

分类号：S152.3;TP391.41

摘要：地表粗颗粒含量是反映土壤侵蚀状况的重要指标。为此,利用数码相机获取地表数字图像,并根据地表颗粒与土壤基底空隙的光谱差异,建立决策树判别函数来进行图像分类,最后利用面积补偿方程计算出地表不同粒级颗粒的百分含量。试验结果表明,该技术的误差在3%以内,精度可以满足实际工作的需要。这种非接触式的地表颗粒含量提取技术省时省力,可为开展风蚀定点监测与评价提供一种行之有效的手段。
The content of wind erosion surface＇ s coarse particles is an important indicator to reflect the soil erosion condition. In this study, portable digital camera was used to obtain surface images, and the decision tree discrimination function was built to make image classification based on the spectral differences between the earth surface particles and the soil substrate interspace. Finally, area compensation equation was established to calculate the content of the wind erosion surface＇ s coarse particles. The results showed that the error of this technology was reduced to less than 3%. That provides an non-contact extraction technology to calculate wind erosion surface＇ coarse particles, which can not only save cumbersome soil samples collecting work, but also provides an effective means to carry out the fixed-point monitoring and evaluation of wind erosion.