基于多源异构数据的粮食作物种植面积提取

Extraction of Grain Crop Planting Area Based on Multi-source Heterogeneous Data

摘要:

利用遥感影像进行农作物面积提取已经普遍应用到农作物种植面积提取工作中，但是普遍存在着提取效率低，精度达不到生产要求。随着高分辨率传感器的相继问世,融合多平台遥感数据有利于提高分类和动态监测的精度。本文以日照市莒县粮食种植面积提取为例，融合多时相和多源异构数据，利用深度学习算法提取粮食种植区域，深入挖掘多源数据的互补效应，提出了融合多时段、多源地理信息提取农作物种植面积的方法，为实现农作物种植面积快速高效的提取奠定了基础。

Abstract:

Crop area extraction by using remote sensing images has been widely used in the extraction of crop planting area, but the extraction efficiency is low and the accuracy can not meet the production requirements. Accompanying with the advent of high-resolution sensors, the fusion of multi platform remote sensing data is conducive to improve the accuracy of classification and dynamic monitoring. In this paper, taking the extraction of grain planting area in Juxian county, Rizhao city as an example, multi temporal and multi-source heterogeneous data have been integrated, grain planting area by using deep learning algorithm has been extracted, the complementary effect of multi-source data has been deeply excavated, and a method of extracting crop planting area by integrating multi temporal and multi-source geographic information has been puts forward. It will lay a foundation for the rapid and efficient extraction of crop planting area.