基于改进DeepLabV3+的栖霞市苹果园遥感识别Remote sensing identification of apple orchards in Qixia City based on improved DeepLabV3+
杜欣苑,张小咏
摘要(Abstract)：
苹果园遥感识别是苹果种植精细化管理的重要基础，但在复杂地物背景下易出现错检、漏检和边界模糊等问题。为提升识别精度，基于高分二号影像与实地采样数据，构建了高分辨率苹果园数据集，并提出一种改进DeepLabV3+多层次特征融合模型，对栖霞市苹果园进行识别。模型采用轻量级MobileNetV2作为主干特征提取网络，将坐标注意力（coordinate attention, CA）机制和条形池化（strip pooling, SP）引入空洞空间金字塔池化（atrous spatial pyramid pooling,ASPP），构建CASP-ASPP模块以融合多尺度特征，并在解码阶段加入边界细化模块优化边界识别。实验结果表明，改进模型的平均交并比较原始模型提升1.9百分点，整体识别精度优于多种主流深度学习网络。该方法可有效提升苹果园遥感识别精度，为果园监测与精细农业管理提供可靠技术支撑。
关键词(KeyWords)： 苹果园识别;深度学习;语义分割;DeepLabV3+;注意力机制
基金项目(Foundation): 遥感大数据智能分析系统开发算法研究（9152335903）
作者(Author): 杜欣苑,张小咏
DOI: 10.16508/j.cnki.11-5866/n.2026.01.002
参考文献(References)：
- [1]ZHU Y H,YANG G J,YANG H,et al. Identification of apple orchard planting year based on spatiotemporally fused satellite images and clustering analysis of foliage phenophase[J]. Remote Sensing,2020,12(7):1199.
- [2]王壹.中国苹果产量蝉联世界第一[N].农民日报，2024-11-09(007).WANG Y. China′s apple production has ranked first in the world for consecutive years[N]. Farmers′Daily,2024-11-09(007).
- [3]唐晓颖.基于多时相遥感和GEE的苹果园识别与面积提取[D].泰安：山东农业大学，2023.TANG X Y. Apple orchard recognition and area extraction based on multi-temporal remote sensing and GEE platform[D]. Taian:Shandong Agricultural University,2023.(in Chinese)
- [4]KOTSIANTIS S B. Decision trees:a recent overview[J].Artificial Intelligence Review,2013,39(4):261-283.
- [5]HEARST M A,DUMAIS S T,OSUNA E,et al. Support vector machines[J]. IEEE Intelligent Systems and Their Applications,1998,13(4):18-28.
- [6]BREIMAN L. Random forests[J]. Machine Learning,2001,45(1):5-32.
- [7]杨彦荣，宋荣杰，胡国强，等.基于随机森林和纹理特征的苹果园遥感提取[J].现代电子技术，2020,43(3):40-44.YANG Y R,SONG R J,HU G Q,et al. Apple orchard remote sensing extraction based on random forest and texture features[J]. Modern Electronics Technique,2020,43(3):40-44.(in Chinese)
- [8]陈印，刚成诚，刘欢欢，等.基于GEE的苹果园地遥感信息提取研究：以陕西省渭北旱塬区为例[J].西北林学院学报，2024,39(1):36-43.CHEN Y,GANG C C,LIU H H,et al. Remote sensing extraction of apple orchard based on GEE:a case study of Weibei plateau region in Shaanxi Province[J]. Journal of Northwest Forestry University,2024,39(1):36-43.(in Chinese)
- [9]ZHAO H L,MENG J H,JI Y P. Application status and prospect of remote sensing technology in precise planting management of apple orchards[J]. Remote Sensing for Natural Resources,2023,35(2):1-15.
- [10]ZHANG T T,HU D N,WU C X,et al. Large-scale apple orchard mapping from multi-source data using the semantic segmentation model with image-to-image translation and transfer learning[J]. Computers and Electronics in Agriculture,2023,213:108204.
- [11]CHENG J Y,ZHU Y H,ZHAO Y Y,et al. Application of an improved U-net with image-to-image translation and transfer learning in peach orchard segmentation[J]. International Journal of Applied Earth Observation and Geoinformation,2024,130:103871.
- [12]LI J B,YANG G J,YANG H,et al. Orchard classification based on super-pixels and deep learning with sparse optical images[J]. Computers and Electronics in Agriculture,2023,215:108379.
- [13]HOU Q B,ZHOU D Q,FENG J S. Coordinate attention for efficient mobile network design[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). New York, USA:IEEE,2021:13708-13717.
- [14]KIPF T N,WELLING M. Semi-supervised classification with graph convolutional networks[EB/OL].(2016-09-09)[2024-06-02]. https：//arxiv. org/abs/1609. 02907.
- [15]赵胜亭，宋秀英.基于DEM的栖霞山区日照时数模拟研究[J].农业网络信息，2011(5):15-18.ZHAO S T,SONG X Y. Research on sunlight hours simulation in Qixia based on DEM[J]. Agriculture Network Information,2011(5):15-18.(in Chinese)
- [16]潘仕梅，史淑一，衣淑玉，等.山东省栖霞市降水特征分析[J].安徽农业科学，2011,39(4):2038-2039.PAN S M,SHI S Y,YI S Y,et al. Analysis on precipitation characteristics about Qixia City of Shandong Province[J].Journal of Anhui Agricultural Sciences,2011,39(4):2038-2039.(in Chinese)
- [17]赵玉山.山东：栖霞市持续做大做强大苹果产业[J].中国果业信息，2023,40(10):53-54.ZHAO Y S. Shandong:Qixia City continues to expand and strengthen its apple industry[J]. China Fruit News,2023,40(10):53-54.(in Chinese)
- [18]CHEN L C,ZHU Y K,PAPANDREOU G,et al. Encoderdecoder with atrous separable convolution for semantic image segmentation[C]//Computer Vision-ECCV 2018. Berlin,Germany:Springer-Verlag Berlin,2018:833-851.
- [19]CHEN L C,PAPANDREOU G,SCHROFF F,et al. Rethinking atrous convolution for semantic image segmentation[EB/OL].(2017-06-17)[2024-06-02]. https：//arxiv. org/abs/1706. 05587.
- [20]SANDLER M,HOWARD A,ZHU M L,et al. MobileNetV2:inverted residuals and linear bottlenecks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York, USA:IEEE,2018:4510-4520.
- [21]YANG M K,YU K,ZHANG C,et al. DenseASPP for semantic segmentation in street scenes[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York, USA:IEEE,2018:3684-3692.