基于改进的YOLO11的天际线检测算法

doi:10.13474/j.cnki.11-2246.2026.0503

摘要/Abstract

摘要： [目的]天际线检测在地理空间定位、飞行器控制、视觉导航等领域具有重要作用。由于受到天气和光照的影响,天空与非天空区域的场景变化较大,给天际线检测带来了挑战。[方法]为了应对这种挑战,本文提出YUNet算法,通过改进YOLO11架构,实现了在复杂场景下对天空区域的分割及天际线提取。将YOLO11网络改进为U型的网络结构,包含编码器网络、颈部网络及解码器网络3部分。编码器网络主要用于从输入图像中提取多尺度特征;颈部网络对编码器输出的多尺度特征进行融合;解码器则利用融合后的多尺度特征完成天空区域分割预测的重建工作。[结果]为了验证该算法的有效性,分别在Skyfinder与CH1数据集上进行天空区域分割及天际线提取试验。结果显示,YUNet的分割 IoU达0.986,天际线检测的平均误差仅为1.36像素。[结论]YUNet具有良好的准确率与实时性,能够在复杂环境下完成天空区域分割与天际线提取任务,具有工程应用价值。

关键词: 天际线检测, YOLO11, 深度学习, 图像处理

Abstract: [Purposes] Skyline detection plays an important role in geolocalization,flight control,visual navigation,etc.The appearance of the sky and non-sky areas are variable,because of different weather or illumination environment,which brings challenges to skyline detection.[Methods]For these challenges,we proposes the YUNet algorithm,which improves the YOLO11 architecture to segment the sky region and extract the skyline in complicated and variable circumstances.In this research,the YOLO11 architecture is extended as an UNet-like architecture,consisting of an encoder,neck and decoder submodule.The encoder extracts the multi-scale features from the given images.The neck makes fusion of these multi-scale features.The decoder applies the fused features to complete the prediction rebuilding.To validate the proposed approach,the YUNet is tested on Skyfinder,CH1 datasets for segmentation and skyline detection,respectively.[Findings] The test shows that the IoU of YUNet segmentation can reach 0.986,and the average error of YUNet skyline detection is just 1.36 pixels.[Conclusions] YUNet has an excellent performance and speed.And it can complete the sky segment and skyline detection task in the complex environment,which is valuable for engineering applications.

Key words: skyline detection, YOLO11, deep learning, image processing

中图分类号:

P237

杨刚, 王淼, 陈思, 周佺, 李江川. 基于改进的YOLO11的天际线检测算法[J]. 测绘通报, 2026, 0(5): 12-16.

YANG Gang, WANG Miao, CHEN Si, ZHOU Quan, LI Jiangchuan. A skyline detection algorithm based on improved YOLO11[J]. Bulletin of Surveying and Mapping, 2026, 0(5): 12-16.

参考文献

[1] CHIU C C,LIU Y J,CHIU S Y,et al.A skyline detection algorithm for use in different weather and environmental conditions[C]//Proceedings of 2016 IEEE International Conference on Electro Information Technology (EIT).Grand Forks:IEEE,2016:680-685.
[2] DONG Yuefang,ZHU Chao,ZHANG Yunfeng,et al.Extracting sea-sky-line based on improved local complexity[C]//Proceedings of 2010 International Conference on Computer,Mechatronics,Control and Electronic Engineering.Changchun:IEEE,2010:82-85.
[3] FEFILATYEV S,SMARODZINAVA V,HALL L O,et al.Horizon detection using machine learning techniques[C]//Proceedings of the 5th International Conference on Machine Learning and Applications.Orlando:IEEE,2006:17-21.
[4] AHMAD T,BEBIS G,NICOLESCU M,et al.An edge-less approach to horizon line detection[C]//Proceedings of 2015 IEEE International Conference on Machine Learning and Applications.Miami:IEEE,2016:1095-1102.
[5] AHMAD T,EMAMI E, AČ ADÍK M,et al.Resource efficient mountainous skyline extraction using shallow learning[C]//Proceedings of 2021 International Joint Conference on Neural Networks.Shenzhen:IEEE,2021:1-9.
[6] PORZI L,ROTA BULÒ S,RICCI E.A deeply-supervised deconvolutional network for horizon line detection[C]//Proceedings of the 24th ACM International Conference on Multimedia.Amsterdam:ACM Press,2016:137-141.
[7] YANG Lujing,ZHANG Peng,HUANG Liang,et al.Sea-sky-line detection based on improved YOLOv5 algorithm[C]//Proceedings of 2021 IEEE International Conference on Information Technology,Big Data and Artificial Intelligence.Chongqing:IEEE,2022:688-694.
[8] FRAJBERG D,FRATERNALI P,TORRES R N.Convolutional neural network for pixel-wise skyline detection[C]//Proceedings of 2017 Artificial Neural Networks and Machine Learning.Cham:Springer,2017:12-20.
[9] HARJOSEPUTRO Y,YUDA I P,DANUKUSUMO K P.MobileNets:efficient convolutional neural network for identification of protected birds[J].International Journal on Advanced Science,Engineering and Information Technology,2020,10(6):2290-2296.
[10] SAURER O,BAATZ G,KÖSER K,et al.Image based geo-localization in the Alps[J].International Journal of Computer Vision,2016,116(3):213-225.
[11] LONG J,SHELHAMER E,DARRELL T.Fully convolutional networks for semantic segmentation[C]//Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition.Boston:IEEE,2015:3431-3440.
[12] BADRINARAYANAN V,KENDALL A,CIPOLLA R.SegNet:a deep convolutional encoder-decoder architecture for image segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(12):2481-2495.
[13] RONNEBERGER O,FISCHER P,BROX T.U-Net:convolutional networks for biomedical image segmentation[M]//Medical Image Computing and Computer-Assisted Intervention.Cham:Springer International Publishing,2015:234-241.
[14] MIHAIL R P,WORKMAN S,BESSINGER Z,et al.Sky segmentation in the wild:an empirical study[C]//Proceedings of 2016 IEEE Winter Conference on Applications of Computer Vision.Lake Placid:IEEE,2016:1-6.
[15] KUANG Boyu,RANA Z A,ZHAO Yifan.Sky and ground segmentation in the navigation visions of the planetary rovers[J].Sensors,2021,21(21):6996.
[16] AYADI M,SUTA L,SCUTURICI M,et al.A parametric algorithm for skyline extraction[C]//Proceedings of 2016 Advanced Concepts for Intelligent Vision Systems.Cham:Springer,2016:604-615.
[17] NICE K A,WIJNANDS J S,MIDDEL A,et al.Sky pixel detection in outdoor imagery using an adaptive algorithm and machine learning[J].Urban Climate,2020,31:100572.
[18] BOCHKOVSKIY A,WANG C Y,LIAO H M.YOLOv4:optimal speed and accuracy of object detection[EB/OL].[2025-12-20].https://arxiv.org/abs/2004.10934.
[19] ZHAO Hengshuang,SHI Jianping,QI Xiaojuan,et al.Pyramid scene parsing network[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Honolulu:IEEE,2017:6230-6239.
[20] GUO Fan,MAI Yuxiang,TANG Jin,et al.Robust and automatic skyline detection algorithm based on MSSDN[J].Journal of Advanced Computational Intelligence and Intelligent Informatics,2020,24(6):750-762.