摘要
本文针对国产卫片多时相、长时序、全天候、多源海量等特点,提出了一种高效、准确的卫片图斑提取方法.该方法在深度学习理论基础上构建了地物目标语义分割模型和图斑提取智能算法群,实现了国产卫片图斑的特征、规律及属性的自动识别,完成了卫片图斑提取的智能化和自动化.试验结果表明,该方法在国产卫片图斑提取中具有较高的准确率,为后续图像处理、分析和应用提供了重要支持.
Abstract
This paper addresses the characteristics of domestic satellite imagery,such as multi-temporal,long-time series,massive,and massive multi-source data,proposes an efficient and accurate method for the extraction of satellite imagery patches.Based on the principles of deep learning,this method constructs a semantic segmentation model for ground objects and a group of intelligent algorithms for patch extraction based on deep learning theory,enabling the automatic recognition of the features,patterns,and attributes of satellite imagery patches,which leads to the intelligent and automated extraction of these patches.Experimental results demonstrate that this method achieves a high level of accuracy in the extraction of patches from domestic satellite imagery,provides important support for subsequent image processing,analysis,and applications.
基金项目
山西省重点研发计划(202202010101005)
维普
万方数据