基于样本增量学习的遥感影像分类  

Remote sensing image classification based on sample incremental learning

作　　者：李雪 姚光乐[1,2] 王洪辉 李军[2] 周皓然[3,4] 叶绍泽[4] LI Xue;YAO Guangle;WANG Honghui;LI Jun;ZHOU Haoran;YE Shaoze(State Key Laboratory of Geohazard Prevention and Geoenvironment Protection(Chengdu University of Technology),Chengdu Sichuan 610059,China;College of Computer Science and Cyber Security,Chengdu University of Technology,Chengdu Sichuan 610059,China;Key Laboratory of Natural Disaster Monitoring,Early Warning and Assessment of Jiangxi Province(Jiangxi Normal University),Nanchang Jiangxi 330022,China;Shenzhen Sensing Data Technology Company Limited,Shenzhen Guangdong 518052,China)

机构地区：[1]地质灾害防治与地质环境保护国家重点实验室(成都理工大学),成都610059 [2]成都理工大学计算机与网络安全学院,成都610059 [3]江西省自然灾害监测预警与评估重点实验室(江西师范大学),南昌330022 [4]深圳市森歌数据技术有限公司,广东深圳518052

出　　处：《计算机应用》2024年第3期732-736,共5页journal of Computer Applications

基　　金：四川省重点研发项目(2021YFG0298);四川省基础研究项目(2021YJ0086);数学地质四川省重点实验室开放基金资助项目(SCSXDZ2020YB04)。

摘　　要：深度学习模型在遥感影像分类中取得了显著的成绩。随着新的遥感数据不断被采集,基于深度学习的遥感影像分类模型在训练新数据、学习新知识时,对旧数据的识别性能会下降,即旧知识遗忘。为帮助遥感影像分类模型巩固旧知识和学习新知识,提出一种基于样本增量学习的遥感影像分类模型——增量协同学习知识模型(ICLKM)。该模型由两个知识网络组成,第一个网络通过知识蒸馏保留旧模型的输出,缓解知识遗忘问题;第二个网络将新数据的输出作为第一个网络的学习目标,通过维护双网络模型的一致性有效地学习新知识。最后两个网络共同学习,通过知识协同策略生成更精确的模型。在两个遥感数据集NWPU-RESISC45和AID上的实验结果表明,相较于微调训练(FT)方法,ICLKM的准确率分别提升了3.53和6.70个百分点。可见ICLKM能够有效解决遥感影像分类的知识遗忘问题,不断提高对已知遥感影像的识别准确率。Deep learning models have achieved remarkable results in remote sensing image classification.With the continuous collection of new remote sensing images,when the remote sensing image classification models based on deep learning train new data to learn new knowledge,their recognition performance of old data will decline,that is,old knowledge forgetting.In order to help remote sensing image classification model consolidate old knowledge and learn new knowledge,a remote sensing image classification model based on sample incremental learning,namely ICLKM(Incremental Collaborative Learning Knowledge Model)was proposed.The model consisted of two knowledge networks.The first network mitigated knowledge forgetting by retaining the output of the old model through knowledge distillation.The second network took the output of new data as the learning objective of the first network and effectively learned new knowledge by maintaining the consistency of the dual network models.Finally,two networks learned together to generate more accurate model through knowledge collaboration strategy.Experimental results on two remote sensing datasets NWPU-RESISC45 and AID show that,ICLKM has the accuracy improved by 3.53 and 6.70 percentage points respectively compared with FT(Fine-Tuning)method.It can be seen that ICLKM can effectively solve the knowledge forgetting problem of remote sensing image classification and continuously improve the recognition accuracy of known remote sensing images.

关 键 词：遥感影像分类 增量学习 知识蒸馏 协同学习 微调 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]