基于Google图像的农作物地块信息仿真提取技术  

Simulating Method of Getting Crops Land Information Based on Google Images

作　　者：齐玉斌[1] 李国昉[1] 崔维[1] 张文元[1] 丁亚周[1] 

机构地区：[1]河北科技师范学院,河北秦皇岛066004

出　　处：《农机化研究》2008年第10期125-127,共3页Journal of Agricultural Mechanization Research

基　　金：河北省科技厅科技攻关项目(7221004)

摘　　要：高分辨率遥感影像是农作物地块提取的主要数据源。为此,以由Google图像得到的多尺度影像分割与面向对象影像分析方法为主要技术,利用样本多边形对象的成员函数建立训练区,自动提取农作物的覆盖信息,可以直接应用于农业调查和生长趋势预测等方面。该方法具有信息获取周期短、精度高和成本低的特点,能够实现农作物地块信息的精确获取与快速更新。High resolution remote sense portrait is the main data source that crops land, the portrait carves up the main body of a book to much dimension gets from Google images and he geared to the needs of marriage partner portrait analysis method is main technology, make use of the member of sample book polygon marriage partner function to build the area training, the automation extracts the crops coverage information, can he direct apply to aspect such as agricultural investigation, growth trend forecast. The method has information gaining a period shortly, accuracy height, cost hang down, the precise being able to realize crops land information gains and the fleetness refresh.

关 键 词：Google图像 多尺度量化 面向对象 农作物地块 仿真 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] S126[自动化与计算机技术—计算机科学与技术]