福建省龙岩市人为扰动图斑AI遥感识别模型

doi:10.13474/j.cnki.11-2246.2026.0520

摘要/Abstract

摘要： [目的]针对福建省龙岩市人为扰动地块监测需求,本文开展了人为扰动地块AI遥感识别模型架构优选及应用。[方法]基于福建龙岩2 m高分影像构建人为扰动样本,采用DeepLabV3+、PAN、SegFormer、U-Net++、SCSE-UNet及TransUNet 6种模型架构构建AI识别模型,开展遥感智能识别与分析。[结果]结果显示,TransUNet综合性能最优,交并比为0.75,F1分数为0.84;SegFormer次之。在2024年龙岩市识别应用中,TransUNet总体精度(OA)为0.99,Kappa系数为0.87,生产者精度(Q_pa)为0.95,用户精度(Q_ua)为0.81,识别精度与轮廓细节刻画更优。[结论]TransUNet识别结果与实际人为活动空间契合,扰动集中于城镇化建设、采矿及道路建设沿线,可为龙岩市水土保持动态监测提供高效精准的技术支撑。

关键词: AI, 高分辨率, 人为扰动, 模型识别

Abstract: [Purposes]In response to the monitoring needs of artificially disturbed land in Longyan,Fujian province,optimize the architecture of the AI remote sensing recognition model for artificially disturbed land and apply it to promote sustainable development of the ecological environment.[Methods]This research constructs a artificially disturbed sample based on 2 m high-resolution imagery from Longyan,Fujian,using six model architectures:DeepLabV3+,PAN,SegFormer,U-Net++,SCSE-UNet,and TransUNet,to develop AI recognition models and conduct remote sensing intelligent recognition and analysis.[Findings]TransUNet has the best overall performance,with an intersection over union of 0.75 and an F1-Score of 0.84,SegFormer follows closely.In the recognition application in Longyan city in 2024,the overall accuracy (OA)of TransUNet is 0.99,the Kappa coefficient is 0.87,the producer's accuracy (Q_pa) is 0.95,and the user's accuracy (Q_ua) is 0.81,achieving superior recognition accuracy and contour detail depiction.[Conclusions]The identification results of TransUNet align with the actual human activity space,with disturbances concentrated in urbanization construction,mining,and along road construction lines,providing efficient and accurate technical support for dynamic monitoring of soil and water conservation in Longyan city.

Key words: AI, high resolution, artificial disturbance, model recognition

中图分类号:

顾祝军, 刘佳, 麦贤智, 吴家晟, 岳辉, 林根根, 贺燕子, 曹正金, 廖广慧. 福建省龙岩市人为扰动图斑AI遥感识别模型[J]. 测绘通报, 2026, 0(5): 122-127,142.

GU Zhujun, LIU Jia, MAI Xianzhi, WU Jiasheng, YUE Hui, LIN Gengen, HE Yanzi, CAO Zhengjin, LIAO Guanghui. AI remote sensing identification model for artificially disturbed patches in Longyan,Fujian province[J]. Bulletin of Surveying and Mapping, 2026, 0(5): 122-127,142.

参考文献

[1] 史志华,杨洁,李忠武,等.南方红壤低山丘陵区水土流失综合治理[J].水土保持学报,2018,32(1):6-9.
[2] 应凌霄,王军,周妍.闽江流域生态安全格局及其生态保护修复措施[J].生态学报,2019,39(23):8857-8866.
[3] 王松伦,马晓楠,潘子轩.基于遥感影像的生产建设项目扰动图斑人机协同智能提取方法[J].测绘通报,2023(11):132-138.
[4] SHIVAKUMAR B R,RAJASHEKARARADHYA S V.Investigation on land cover mapping capability of maximum likelihood classifier:a case study on north Canara,India[J].Procedia Computer Science,2018,143:579-586.
[5] ALMEIDA AFFONSO A,MANDAI S S,PORTELLA T P,et al.Tracking land use and land cover changes in the Volta grande do Xingu(PARÁ-Brazil)between 2000 and 2017 through three pixel-based classification methods[C]//Proceedings of 2022 IEEE International Geoscience and Remote Sensing Symposium.Kuala Lumpur:IEEE,2022:5630-5633.
[6] 胡玉福,邓良基,匡先辉,等.基于纹理特征的高分辨率遥感图像土地利用分类研究[J].地理与地理信息科学,2011,27(5):42-45.
[7] 熊羽,左小清,黄亮,等.基于多特征组合的彩色遥感图像分类研究[J].激光技术,2014,38(2):165-171.
[8] SVETNIK V,LIAW A,TONG C,et al.Random forest:a classification and regression tool for compound classification and QSAR modeling[J].Journal of Chemical Information and Computer Sciences,2003,43(6):1947-1958.
[9] 苗水清,闫文耀,吴梦蝶.基于并行权重自适应k-邻域算法的图像分类方法[J].贵州师范大学学报(自然科学版),2023,41(2):113-120.
[10] FRIEDL M A,BRODLEY C E.Decision tree classification of land cover from remotely sensed data[J].Remote Sensing of Environment,1997,61(3):399-409.
[11] LIU Guangxin,WANG Liguo,LIU Danfeng,et al.Hyperspectral image classification based on non-parallel support vector machine[J].Remote Sensing,2022,14(10):2447.
[12] BELGIU M,DRǍGUŢ L.Random forest in remote sensing:a review of applications and future directions[J].ISPRS Journal of Photogrammetry and Remote Sensing,2016,114:24-31.
[13] MA Lei,LIU Yu,ZHANG Xueliang,et al.Deep learning in remote sensing applications:a meta-analysis and review[J].ISPRS Journal of Photogrammetry and Remote Sensing,2019,152:166-177.
[14] KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[J].Communications of the ACM,2017,60(6):84-90.
[15] 李雪涛,王盼成,曾永年.基于高分辨率遥感影像深度特征的城市非渗透表面集成学习提取[J].测绘学报,2024,53(4):700-711.
[16] SHELHAMER E,LONG J,DARRELL T.Fully convolutional networks for semantic segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(4):640-651.
[17] ZHOU Zongwei,RAHMAN SIDDIQUEE M M,TAJBAKHSH N,et al.UNet++: a nested U-Net architecture for medical image segmentation[C]//Proceedings of 2018 Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support.Cham:Springer,2018:3-11.
[18] CHEN L C,ZHU Yukun,PAPANDREOU G,et al.Encoder-decoder with atrous separable convolution for semantic image segmentation[C]//Proceedings of 2018 Computer Vision-ECCV.Cham:Springer,2018:833-851.
[19] LI Hanchao,XIONG Pengfei,AN Jie,et al.Pyramid attention network for semantic segmentation[C]//Proceedings of 2018 IEEE Conference on Computer Vision and Pattern Recognition.Salt Lake City:IEEE,2018:2778-2787.
[20] DOSOVITSKIY A,BEYER L,KOLESNIKOV A,et al.An image is worth 16x16 words:transformers for image recognition at scale[EB/OL].[2025-07-20].https://arxiv.org/abs/2010.11929.
[21] BAZI Y,BASHMAL L,AL RAHHAL M M,et al.Vision transformers for remote sensing image classification[J].Remote Sensing,2021,13(3):516.
[22] YAN Chuan,FAN Xiangsuo,FAN Jinlong,et al.Improved U-Net remote sensing classification algorithm based on multi-feature fusion perception[J].Remote Sensing,2022,14(5):1118.
[23] 赵凤,耿苗苗,刘汉强,等.卷积神经网络与视觉Transformer联合驱动的跨层多尺度融合网络高光谱图像分类方法[J].电子与信息学报,2024,46(5):2237-2248.
[24] 张继贤,顾海燕,倪欢,等.遥感智能变化检测的深度学习方法:演变与发展趋势[J].测绘学报,2025,54(8):1347-1370.
[25] 王宁涛,谭建民,闫举生,等.矿区地下水监测与预警系统研究:以福建省龙岩市马坑铁矿为例[J].安全与环境工程,2011,18(1):95-100.
[26] 王克勤,姜德文,高天天,等.基于层次分析法的生产建设项目水土流失影响综合指数研究[J].水土保持通报,2015,35(3):136-142.
[27] CHEN Hui,QIN Yuanshou,LIU Xinyuan,et al.An improved DeepLabV3+ lightweight network for remote-sensing image semantic segmentation[J].Complex ＆ Intelligent Systems,2024,10(2):2839-2849.
[28] LUO Qian,SHAO Jie,DANG Wanli,et al.An efficient feature pyramid attention network for person re-identification[J].Image and Vision Computing,2024,145:104963.
[29] LIU Siyu,ZHOU Zhiyuan,LIN Jian,et al.Global seamount semantic segmentation on bathymetric maps based on improved deep learning model Segformer[J].Marine Geophysical Research,2025,46(3):23.
[30] TANG Meng,HE Yuelin,ASLAM M,et al.Enhanced U-Net++ for improved semantic segmentation in landslide detection[J].Sensors,2025,25(9):2670.
[31] LIU Xiuli,WU Xiangqiong,CHEN Yang,et al.SCSE-UNet:a spatial channel squeeze-and-excitation UNet for medical image segmentation[C]//Proceedings of the 12th International Conference on Information Systems and Computing Technology.Xi'an:IEEE,2025:1-7.
[32] CHEN Jieneng,MEI Jieru,LI Xianhang,et al.TransUNet:rethinking the U-Net architecture design for medical image segmentation through the lens of transformers[J].Medical Image Analysis,2024,97:103280.
[33] GARCÍA-BALBOA J L,ALBA-FERNÁNDEZ M V,ARIZA-LÓPEZ F J,et al.Analysis of thematic similarity using confusion matrices[J].ISPRS International Journal of Geo-Information,2018,7(6):233.
[34] 王子彦,任超,梁月吉,等.基于高分六号红边特征的面向对象桉树人工林信息提取[J].测绘通报,2021(6):6-11.