ass日本风韵熟妇pics男人扒开女人屁屁桶到爽|扒开胸露出奶头亲吻视频|邻居少妇的诱惑|人人妻在线播放|日日摸夜夜摸狠狠摸婷婷|制服 丝袜 人妻|激情熟妇中文字幕|看黄色欧美特一级|日本av人妻系列|高潮对白av,丰满岳妇乱熟妇之荡,日本丰满熟妇乱又伦,日韩欧美一区二区三区在线

基于決策樹和混合像元分解的玉米種植面積提取方法
CSTR:
作者:
作者單位:

作者簡介:

通訊作者:

中圖分類號:

基金項目:

國家自然科學(xué)基金資助項目(41371327)、“十二五”國家科技支撐計劃資助項目(2012BAD20B0103)和北京高等學(xué)校青年英才計劃資助項目(YETP0316)


Extraction of Maize Planting Area Based on Decision Tree and Mixed-pixel Unmixing Methods
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 圖/表
  • |
  • 訪問統(tǒng)計
  • |
  • 參考文獻
  • |
  • 相似文獻
  • |
  • 引證文獻
  • |
  • 資源附件
  • |
  • 文章評論
    摘要:

    Landsat 8影像具有較高空間分辨率和時間分辨率,,長時間序列Landsat 8-NDVI曲線反映農(nóng)作物的物候歷、種植模式和種植結(jié)構(gòu)信息,,是精確提取玉米種植面積的理想數(shù)據(jù)源?;跁r序Landsat 8-NDVI影像提取玉米種植面積的方法中,,決策樹方法快速、高效,,可通過多閾值限定進行分類,,但由于混合像元問題,如果閾值設(shè)置過寬,,提取面積偏大,;閾值設(shè)置過窄,提取面積偏??;混合像元分解通過計算端元組分豐度可以排除異質(zhì)地類干擾。因此,,以時序NDVI為數(shù)據(jù)源,、耦合使用2種算法是精確提取作物種植面積的有效方法。本研究基于時序Landsat 8-NDVI,,提取河北省保定市大田玉米的種植面積,。首先,分析典型作物區(qū)的NDVI曲線特征,,并構(gòu)建決策樹從而初步提取早播夏玉米,、小麥夏玉米和春玉米的分布范圍。然后,,根據(jù)端元平均NDVI波譜曲線,,進行3種玉米混合度分解,進而根據(jù)玉米豐度比例精確提取玉米種植面積,。精度評價結(jié)果表明:利用本方法提取的玉米種植區(qū)總分類精度在98%以上,,Kappa系數(shù)在0.97以上;所提取的玉米種植類型主要是夏玉米,,春玉米種植主要集中在涿州市中部,,這與實地調(diào)查結(jié)果一致。上述定量和定性的評價結(jié)果表明該方法可用于快速,、精確提取玉米種植面積,。

    Abstract:

    Landsat 8 remote sensingimages possess higher spatial resolution and higher temporal resolution.The timeseries Landsat 8-NDVI metrics could reflect the phenology calendar, planting pattern, planting structure and planting area information due to its high spatial resolution and high temporal resolution, thus it is an ideal data source for accurate extraction of maize planting area. In most extraction methods, the decision tree classification method is considered to be rapid and efficient, which could extract maize planting area using multithreshold. However, because of the mixedpixel, both the larger and smaller threshold will lead to errors. This problem could be resolved by mixedpixel unmixing method using endmember abundance calculation to eliminate the disturbance of heterogeneous classes. Therefore, taking timeseries Landsat 8-NDVI metrics as data source and using the combined method of decision tree and mixedpixel unmixing methods are effective way to extract crop planting area. The maize planting area in Hebei Province was extracted in this paper based ontimeseries Landsat 8-NDVI. Firstly, the features of timeseries Landsat 8-NDVI curves were analyzed and the decision tree was built to get the distribution of early sowing maize, interplanted summer maize and spring maize. Secondly, mixedness decomposition was calculated among three kinds of maize based on mean NDVI spectral curve of endmember, so maize planting area could be extracted accurately by using computed maize endmember abundance. The accuracy assessment results indicated that the overall classification accuracy of maize planting area was higher than 98% and Kappa coefficient was higher than 0.97. Generally speaking, the main planting crop was summer maize, and spring maize was mostly planted in the south part of Zhuozhou City. These results were accordant with field work data. The above quantitative and qualitative accuracy assessment results indicated that this method can be used to extract maize planted area quickly and accurately.

    參考文獻
    相似文獻
    引證文獻
引用本文

蘇 偉,姜方方,朱德海,展郡鴿,馬鴻元,張曉東.基于決策樹和混合像元分解的玉米種植面積提取方法[J].農(nóng)業(yè)機械學(xué)報,2015,46(9):289-295. Su Wei, Jiang Fangfang, Zhu Dehai, Zhan Junge, Ma Hongyuan, Zhang Xiaodong. Extraction of Maize Planting Area Based on Decision Tree and Mixed-pixel Unmixing Methods[J]. Transactions of the Chinese Society for Agricultural Machinery,2015,46(9):289-295.

復(fù)制
分享
文章指標
  • 點擊次數(shù):
  • 下載次數(shù):
  • HTML閱讀次數(shù):
  • 引用次數(shù):
歷史
  • 收稿日期:2014-12-11
  • 最后修改日期:
  • 錄用日期:
  • 在線發(fā)布日期: 2015-09-10
  • 出版日期:
文章二維碼