基于 Sentinel--2 影像的吉林省中部黑土区 保护性耕作范围的遥感识别

沈峰, 石璞, 井逢瑞

世界地质 ›› 2022, Vol. 41 ›› Issue (4) : 873-881.

PDF(975 KB)
PDF(975 KB)
世界地质 ›› 2022, Vol. 41 ›› Issue (4) : 873-881.
基础地质

 基于 Sentinel--2 影像的吉林省中部黑土区 保护性耕作范围的遥感识别

  •  沈峰,石璞,井逢瑞
作者信息 +

 Remote sensing identification of conservation tillage extent in black soil region in central Jilin Province based on Sentinel--2 image

  •  SHEN Feng,SHI Pu,JING Fengrui
Author information +
文章历史 +

摘要

 针对黑土区保护性耕作范围识别和监测方法的缺失问题,笔者选择位于吉林省中部的公主岭市 和梨树县为研究区,通过耕地地表图像采集,构建了一种基于纠缠随机森林模型的秸秆覆盖率图像识 别方法,获取研究区秸秆覆盖率数据集,并据此建立基于 Sentinel--2 遥感光谱指数的保护性耕作范围 识别模型 ( 支持向量机和逻辑回归) ,实现了黑土区保护性耕作范围的遥感反演与空间表征。结果表 明: ①由于 “梨树模式”的高成效推广,该县免耕地块占比高达 72. 86% ,而公主岭市免耕地块仅占 36. 03% ; ②免耕方式的推行带来了地表秸秆覆盖率的提升,免耕地块的秸秆覆盖率均值较非免耕地 块高 2 倍以上,但仍然有 40% 左右的免耕地块存在秸秆覆盖率 < 4% 的情况; ③利用逻辑回归模型获 得了耕作方式的高精度分类,其预测生成的保护性耕作范围空间分布图显示梨树县保护性耕作的绝对 面积 ( 1. 95 × 105 hm2 ) 和保护性耕作占耕地总面积的比例 ( 77. 79% ) 均高于公主岭市 ( 1. 02 × 105 hm2 ,30. 01% ) 。

Abstract

 To address the lack of methods for identifying and monitoring the extent of conservation tillage in black soil region,the authors selected the Gongzhuling City and Lishu County in central Jilin Province as the study area. Based on the entangled random forest model through surface image collections of cropland a straw cover rate image identification method is constructed and the straw cover rate dataset is obtained in the study area. A conservation tillage extent identification model ( Support Vector Machine and Logistic Regression) is established on the basis of the Sentinel--2 remote sensing spectral index accordingly. Finally,the remote sensing inversion and spatial representation of the conservation tillage extent in black soil region are realized. The results show that ① the proportion of no-tillage plots in the county is as high as 72. 86% due to the highly effective promotion of “Lishu model”,while that in Gongzhuling City accounts for only 36. 03% ; ② the implementation of the no-tillage method has brought about an increase in straw cover rate on the surface of cropland,with the average value of straw cover rate on no-tillage plots being more than twice as high as on non-no-tillage plots,but there are still about 40% of notillage plots with straw cover rate < 4% ; ③ the logistic regression model was used to obtain a high-precision classification of tillage practices,and the spatial distribution map of conservation tillage extent generated by its prediction shows that the absolute area of conservation tillage and the proportion of conservation tillage to total cropland area in Lishu County ( 1. 95 × 105 hm2 and 77. 79% ) were larger and higher than those in Gongzhuling City ( 1. 02 × 105 hm2 and 30. 01% ) .

关键词

 免耕 / 秸秆覆盖率 / Sentinel--2 / 保护性耕作 / 遥感分类 / 吉林省中部

Key words

 no-tillage / straw cover rate / Sentinel--2 / conservation tillage / remote sensing classification / central Jilin Province

引用本文

导出引用
沈峰, 石璞, 井逢瑞.  基于 Sentinel--2 影像的吉林省中部黑土区 保护性耕作范围的遥感识别[J]. 世界地质, 2022, 41(4): 873-881
SHEN Feng, SHI Pu, JING Fengrui.  Remote sensing identification of conservation tillage extent in black soil region in central Jilin Province based on Sentinel--2 image[J]. World Geology, 2022, 41(4): 873-881
中图分类号:     

参考文献

 

基金

 

PDF(975 KB)

34

Accesses

0

Citation

Detail

段落导航
相关文章

/