AM-1DCNN算法在土地覆盖分类中的应用  

Application of the 1DCNN Algorithm with Attention Mechanism in Land Cover Classification

作　　者：孟萌萌 冯贝贝 MENG Mengmeng;FENG Beibei(School of Information and Electronic Engineering,Shangqiu Institute of Technology,Shangqiu,China,476000)

机构地区：[1]商丘工学院信息与电子工程学院,河南商丘476000

出　　处：《福建电脑》2025年第3期36-39,共4页Journal of Fujian Computer

摘　　要：为有效提升土地覆盖分类的效果与效率,本文构建了土地分类模型AM-1DCNN。通过引入注意力机制,使1DCNN算法在处理土地覆盖分类相关数据时,更加聚焦关键特征信息,抑制无关信息干扰,增强特征表达能力。实验结果表明,AM-1DCNN能够进一步提高土地覆盖分类的准确率,能够有效应对复杂地貌特征和多样化的土地覆盖类型,并提高分类准确率。To effectively improve the effectiveness and efficiency of land cover classification,this paper constructs a land classification model AM-1DCNN.By introducing attention mechanism,the 1DCNN algorithm can focus more on key feature information when processing land cover classification related data,suppress irrelevant information interference,and enhance feature expression ability.The experimental results show that AM-1DCNN can further improve the accuracy of land cover classification,effectively cope with complex terrain features and diverse land cover types,and improve classification accuracy.

关 键 词：注意力机制 土地覆盖分类 特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]