...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Journal of intelligent & fuzzy systems: Applications in Engineering and Technology >Crop region extraction of remote sensing images based on fuzzy ARTMAP and adaptive boost
【24h】

Crop region extraction of remote sensing images based on fuzzy ARTMAP and adaptive boost

机译：基于模糊ARTMAP和自适应Boost的遥感影像作物区域提取。

获取原文
获取原文并翻译 | 示例
   

获取外文期刊封面封底 >>

       
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Crop area statistics and yield prediction will affect adjustment of agricultural policy, to a certain extent. With the development of computer automatic classification techniques, the performance of classifiers are influenced by feature preprocessing and sample selection. Remote sensing classification according to spectral information is affected by false negatives and miscalculation in the complex spectrum area. Corn planting areas and other land-cover objects contain different surface structures and smoothness; other vegetation and villages have coarse textures. This paper introduces texture information based on a Gabor filter group to enrich land-cover information and establish a spectrum-texture feature set. With more samples, the algorithm efficiency is greatly affected. This paper proposes an improved fuzzy ARTMAP (FAM) with an adaptive boost strategy, namely Adaboost FAM. Weak classifiers are trained to construct strong classifiers so as to improve operation efficiency. Meanwhile, classification accuracy will not be greatly improved. Experimental results indicate that the proposed method improves extraction accuracy when compared to classical algorithms, and improves efficiency when compared to algorithms which contain a great number of samples.
机译：作物面积统计和单产预测将在一定程度上影响农业政策的调整。随着计算机自动分类技术的发展，分类器的性能受到特征预处理和样本选择的影响。在复杂频谱区域中，根据频谱信息进行的遥感分类会受到假阴性和错误计算的影响。玉米种植区和其他土地覆盖物具有不同的表面结构和光滑度;其他植被和村庄质地粗糙。本文介绍了基于Gabor滤波器组的纹理信息，以丰富土地覆盖信息并建立频谱纹理特征集。随着样本的增加，算法效率受到很大影响。本文提出了一种具有自适应提升策略的改进的模糊ARTMAP（FAM），即Adaboost FAM。训练弱分类器以构造强分类器，以提高操作效率。同时，分类精度不会得到很大提高。实验结果表明，与经典算法相比，该方法提高了提取精度，与包含大量样本的算法相比，提高了效率。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
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