Lightweight Building Detection Model Based on YOLOv4 Optimization for Remote Sensing Images

doi:10.3778/j.issn.1002-8331.2202-0287

Abstract

Abstract: Aiming at the problems of low detection accuracy and large model size of existing building detection models, which lead to unbalanced speed and accuracy of remote sensing image detection and unfavorable to later deployment, a lightweight building detection model based on YOLOv4 optimization for remote sensing images is proposed. Firstly, GhostNet, a lightweight network, is used to replace CSP DarkNet53 for feature extraction. Secondly, the Dense-PANet feature fusion module is proposed by drawing on the idea of dense connection. Finally, the ECA attention mechanism is introduced into the Ghost module to replace the traditional convolution of the neck network. The experimental results show that the model proposed in this paper, compared with YOLOv4, sacrifices a small amount of detection speed, but increases the average precision by 0.96 percentage points, the recall by 1.08 percentage points, and decreases the model volume by 71.39%, the floating point of operations by 76.60%, which can effectively meet the demand of remote sensing image building detection.

Key words: building detection, YOLOv4, lightweight, feature fusion, ECA attention mechanism

摘要： 针对现有建筑物检测模型检测精度低下，模型体积较大，导致遥感图像检测速度和精度无法平衡，不利于后期部署等问题，提出一种基于YOLOv4优化的轻量级遥感图像建筑物检测模型。利用轻量化网络GhostNet替换CSP DarkNet53进行特征提取；借鉴稠密连接思想，提出了Dense-PANet特征融合模块；将ECA注意力机制引入Ghost模块，替换特征融合颈部网络的传统卷积。实验结果表明，提出的模型与YOLOv4相比，牺牲少量检测速度，但是平均精度提高了0.96个百分点，召回率提升了1.08个百分点，模型体积降低了71.39%，浮点计算量降低了76.60%，能有效满足遥感图像建筑物检测的需求。

关键词: 建筑物检测, YOLOv4, 轻量级, 特征融合, ECA注意力机制

DING Fei, SHI Jie, WU Hongjie. Lightweight Building Detection Model Based on YOLOv4 Optimization for Remote Sensing Images[J]. Computer Engineering and Applications, 2023, 59(10): 213-220.

丁飞, 石颉, 吴宏杰. 改进YOLOv4的轻量级遥感图像建筑物检测模型[J]. 计算机工程与应用, 2023, 59(10): 213-220.

References

[1] JIANG K，WANG Z，YI P，et al.Edge-enhanced GAN for remote sensing image superresolution[J].IEEE Transactions on Geoscience and Remote Sensing，2019，57（8）：5799-5812.
[2] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the 2014 IEEE Conference on Computer Vision and Pattern Recognition，2014：580-587.
[3] GIRSHICK R.Fast R-CNN[C]//Proceedings of the 2015 IEEE International Conference on Computer Vision，2015：1440-1448.
[4] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[C]//Advances in Neural Information Processing Systems，2015：91-99.
[5] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//Proceedings of the 14th European Conference on Computer Vision，2016：21-37.
[6] REDMON J，FARHADI A.YOLOv3：an incremental improvement[J].arXiv：1804.02767，2018.
[7] 李东子，范大昭，苏亚龙.结合Faster R-CNN模型的遥感影像建筑物检测[J].测绘科学技术学报，2018，35（4）：389-394.
LI D Z，FAN D Z，SU Y L.Building detection in remote sensing image based on Faster R-CNN[J].Journal of Geomatics Science and Technology，2018，35（4）：389-394.
[8] 董彪，熊风光，韩燮，等.基于改进Yolo v3算法的遥感建筑物检测研究[J].计算机工程与应用，2020，56（18）：209-213.
DONG B，XIONG F G，HAN X，et al.Research on remote sensing building detection based on improved Yolo v3 algorithm[J].Computer Engineering and Applications，2020，56（18）：209-213.
[9] 张通.基于深度学习和直线检测的高分辨率遥感影像建筑物提取[D].武汉：武汉大学，2018.
ZHANG T.Building extraction from high resolution remote sensing images based on deep learning and line detection[D].Wuhan：Wuhan University，2018.
[10] 刘文涛，李世华，覃驭楚.基于全卷积神经网络的建筑物屋顶自动提取[J].地球信息科学学报，2018，20（11）：1562-1570.
LIU W T，LI S H，TAN Y C.Automatic building roof extraction with fully convolutional neural network[J].Journal of Geo-Information Science，2018，20（11）：1562-1570.
[11] DING J，ZHANG J，ZHAN Z，et al.A precision efficient method for collapsed building detection in post-earthquake UAV images based on the improved NMS algorithm and Faster R-CNN[J].Remote Sensing，2022，14（3）：663.
[12] BAI T，PANG Y，WANG J，et al.An optimized faster R-CNN method based on DRNet and RoI align for building detection in remote sensing images[J].Remote Sensing，2020，12（5）：762.
[13] 孟晓龙，任正非.基于改进YOLOv3的震后遥感图像倒塌建筑物检测[J/OL].激光与光电子学进展（2022-02-14）[2022-03-23].http：//kns.cnki.net/kcms/detail/31.1690.TN.
20220211.1846.024.html.
MENG X L，REN Z F，Detection of collapsed buildings in post-earthquake remote sensing images based on improved YOLOv3[J/OL].Laser & Optoelectronics Progress（2022-02-14）[2022-03-23].http：//kns.cnki.net/kcms/detail/31.1690.TN.20220211.1846.024.html.
[14] 赵若辰，王敬东，林思玉，等.基于卷积神经网络的小型建筑物检测算法[J].系统工程与电子技术，2021，43（11）：3098-3106.
ZHAO R C，WANG J D，LIN S Y，et al.Small building detection algorithm based on convolutional neural network[J].Systems Engineering and Electronics，2021，43（11）：3098-3106.
[15] LIU Z，LI J，SHEN Z，et al.Learning efficient convolutional networks through network slimming[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision，2017：2736-2744.
[16] CHEN G，CHOI W，YU X，et al.Learning efficient object detection models with knowledge distillation[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems，2017：742-751．
[17] ZHANG D，YANG J，YE D，et al.LQ-nets：learned quantization for highly accurate and compact deep neural networks[C]//Proceedings of the 15th European Conference on Computer Vision，2018：365-382.
[18] IANDOLA F N，HAN S，MOSKEWICZ M W，et al.SqueezeNet：AlexNet-level accuracy with 50x fewer parameters and < 0.5 MB model size[J].arXiv：1602.07360，2016.
[19] HOWARD A G，ZHU M，CHEN B，et al.MobileNets：efficient convolutional neural networks for mobile vision applications[J].arXiv：1704.04861，2017.
[20] SANDLER M，HOWARD A，ZHU M，et al.MobileNetv2：inverted residuals and linear bottlenecks[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition，2018：4510-4520.
[21] HOWARD A，SANDLER M，CHU G，et al.Searching for MobileNetv3[C]//Proceedings of the 2019 IEEE International Conference on Computer Vision，2019：1314-1324.
[22] TAN M，PANG R，LE Q V.EfficientDet：scalable and efficient object detection[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：10781-10790.
[23] HAN K，WANG Y，TIAN Q，et al.GhostNet：more features from cheap operations[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：1580-1589.
[24] BOCHKOVSKIY A，WANG C Y，LIAO H Y M.YOLOv4：optimal speed and accuracy of object detection[J].ArXiv：2004.10934，2020.
[25] HUANG G，LIU Z，VAN D M L，et al.Densely connected convolutional networks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition，2017：4700-4708.
[26] 金雨芳，吴祥，董辉，等.基于改进YOLOv4的安全帽佩戴检测算法[J].计算机科学，2021，48（11）：268-275.
JIN Y F，WU X，DONG H，et al.Improved YOLOv4 algorithm for safety helmet wearing detection[J].Computer Science，2021，48（11）：268-275.
[27] WANG Q，WU B，ZHU P，et al.ECA-net efficient channel attention for deep convolutional neural networks[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：11534-11542.
[28] JI S，WEI S Q，LU M.Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set[J].IEEE Transactions on Geoscience and Remote Sensing，2018，57（1）：574-586.
[29] 张晓勐，朱德利，余茂生.无人机遥感图像中的玉米雄穗轻量化检测模型[J].江西农业大学学报，2022，44（2）：461-472.
ZHANG X M，ZHU D L，YU M S.Lightweight detection model of maize tassel in UAV remote sensing image[J].Acta Agriculturae Universitatis Jiangxiensis，2022，44（2）：461-472.
[30] 黄丽明，王懿祥，徐琪，等.采用YOLO算法和无人机影像的松材线虫病异常变色木识别[J].农业工程学报，2021，37（14）：197-203.
HUANG L M，WANG Y X，XU Q，et al.Recognition of abnormally discolored trees caused by pine wilt disease using YOLO algorithm and UAV images[J].Transactions of the Chinese Society of Agricultural Engineering，2021，37（14）：197-203.