• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>ISPRS vol.36 pt.7/W20; International Symposium on Physical Measurements and Signatures in Remote Sensing pt.2; 20051017-19; Beijing(CN) >Combining spectral library into intelligent expert classifier for Lychee orchard mapping --- A case study on lychee in Shenzhen, South China
【24h】

Combining spectral library into intelligent expert classifier for Lychee orchard mapping --- A case study on lychee in Shenzhen, South China

机译：将谱库结合到智能专家分类器中进行荔枝果园制图-以深圳南部荔枝为例

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Lychee condition monitoring and yield estimation is an important remote sensing application field at all times and it essentially requires a good classification algorithm for featured crops information extraction such lychee, banana and sugarcane etc. However, there are few researches dealing with intelligent expert classifier for their classification by remote sensing in Guangdong province, South China. With the support of China's '863' high-tech research plan and the science and technology project of Guangdong Province, the research develops a module which is used to monitor lychee condition and estimate yield using spectral library of featured crops of South China and here present an approach to combine spectral library with expert system classification methods by inductive learning. Spatial data mining techniques are used to discover knowledge from spectral library for expert system and inductive learning algorithm, which is designed specially for featured crops of South China. It makes use of spectral data, attribute data and spatial data of spectral library, and form intelligent expert classifier for South China's featured crop extraction combining spectral library. The estimation of lychee planting area of Shenzhen city in 2000 is presented as a case study. The following classification rules of lychee are contained by running inductive learning algorithm: for example, 0.06 < reflectance of TM2 < 0.07, 0.48 < NDVI < 0.53, DEM < 100, probability of lychee planting of different region in Guangdong province, 30 < area of lychee orchard (pixel number) < 5000 and perform classification etc. Compared with traditional unsupervised classification, it improves classification accuracy greatly, and classification accuracy is reached at 91.8% with KAPPA coefficient of 0.81. The result indicates that intelligent expert classifier is a good classification method for lychee planting area extraction and be able to content with the need in agriculture application for quick lychee area monitoring in South China, and the method is easy to extend into other countries and regions of planting lychee with some differences of classification rules. But the classification algorithm has still some defects. First, our inductive learning algorithm is comparatively simple, only based on pixel granularity to select rules. At present, there exist some difficulties for the knowledge discovery of complicated temporal-spatial data. Secondly, the prior lychee information of some remote regions in Guangdong Province has not been collected into the spectral library, which affects the obtaining of corresponding classification rules. Meanwhile, the image classification is restricted by only assigning a kind of land cover category to a whole pixel. So, it is still necessary to further study in spatial object granularity and mixed pixel unmixing for improvement of classification precision.
机译：荔枝状态监测和产量估算一直是重要的遥感应用领域，它本质上需要一种良好的分类算法来提取荔枝，香蕉和甘蔗等特色作物信息。但是，针对智能专家分类器的研究很少。中国广东省的遥感分类。在中国“ 863”高技术研究计划和广东省科技计划的支持下，该研究开发了一个模块，该模块可使用华南特色作物的光谱库监测荔枝状况并估算产量，目前一种通过归纳学习将光谱库与专家系统分类方法相结合的方法。空间数据挖掘技术用于从光谱库中发现专家系统和归纳学习算法的知识，该系统是专为华南特色作物设计的。利用光谱库的光谱数据，属性数据和空间数据，结合光谱库，形成华南特色作物提取的智能专家分类器。本文以2000年深圳市荔枝种植面积估算为例。运行归纳学习算法包含以下荔枝分类规则：例如，0.06

著录项

相似文献

  • 外文文献
  • 中文文献
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号