基于多源遥感影像的农业资源分类勘查技术研究

Research on agricultural resources classification and exploration technology based on multi source remote sensing images

针对传统基于像元的农田遥感技术,存在着对于农作物的识别精度较差、处理效率较低的问题。文中提出了一种遥感数据压缩分级传输系统的设计方案,对遥感数据自适应扫描,码流的比特分配,最终实现了遥感数据的可视化自定义传输。通过多组基于机器学习分类的农作物分类实验,确定了遥感数据传输的最优参数。测试与实验结果表明,在不损失传输信息的情况下,使用机器学习算法的面向对象分类设计总体精度高于95%,Kappa系数大于0.91。

Aiming at the traditional pixel based remote sensing technology for farmland,there are some problems,such as low recognition accuracy and low processing efficiency for crops.In this paper,a design scheme of remote sensing data compression and hierarchical transmission system is proposed,which adaptively scans remote sensing data and allocates bit streams,and finally realizes the visual and customized transmission of remote sensing data.The optimal parameters of remote sensing data transmission are determined through a series of experiments of crop classification based on machine learning classification.The test and experimental results show that the overall accuracy of object oriented classification design using machine learning algorithm is higher than 95%and the Kappa coefficient is greater than 0.91 without losing the transmission information.