基于YOLOv5的违章建筑检测方法

doi:10.3778/j.issn.1002-8331.2106-0178

摘要/Abstract

摘要：

针对无人机图像中违章建筑多为小目标且存在部分遮挡目标导致的检测速率慢、误检率高的问题，提出一种基于YOLOv5网络的违章建筑检测方法。在原来的批量标准化模块开始和结束处分别添加中心和缩放校准增强有效特征并形成更稳定的特征分布，加强网络模型的特征提取能力。用平滑处理后的KL（Kullback-Leibler）散度损失函数替换原损失函数置信度中的交叉熵，进一步提高模型的泛化性能。对YOLOv5的主干特征提取网络进行改进，将残差模块替换为LSandGlass模块减少信息损失并剔除低分辨率的特征层以减少语义丢失。实验结果表明，与原版的YOLOv5相比，改进后模型的训练更容易使得网络收敛，检测违章建筑的速度有了较大提升，同时提高了检测的精确度。

关键词: 神经网络, YOLOv5, 违章建筑检测, 批量标准化, KL散度

Abstract:

Aiming at solving the problem of slow detection rate and high false detection rate caused by the illegal buildings in the UAV images, which are mostly small targets and partially occluded targets, a detection method of illegal buildings based on YOLOv5 network is proposed. Firstly, at the beginning and end of original batch normalization module, centering and scaling calibration are added to enhance the effective features and form a more stable feature distribution, which strengthens the feature extraction ability of network model. Then, the smoothed KL（Kullback-Leibler） divergence loss function is adopted to replace the cross entropy in the confidence of original loss function, which further improves the generalization performance of model. Finally, the backbone feature extraction network of YOLOv5 is improved, and the residual module is replaced with the LSandGlass module to reduce information loss and eliminate low-resolution feature layers to reduce semantic loss. Experimental results show that the training of the proposed improved model is easier to make network converge in comparison with original YOLOv5, and the speed of detecting illegal buildings has been greatly improved, and then detection accuracy has been improved.

Key words: neural network, YOLOv5, illegal buildings detection, batch normalization, KL divergence

于娟，罗舜. 基于YOLOv5的违章建筑检测方法[J]. 计算机工程与应用, 2021, 57(20): 236-244.

YU Juan，LUO Shun. Detection Method of Illegal Building Based on YOLOv5[J]. Computer Engineering and Applications, 2021, 57(20): 236-244.

参考文献

[1] 朱虎明，李佩，焦李成，等.深度神经网络并行化研究综述[J].计算机学报，2018，41（8）：171-191.
ZHU Huming，LI Pei，JIAO Licheng，et al.Review on parallelization of deep neural networks[J].Chinese Journal of Computer，2018，41（8）：171-191.
[2] 李晏婷.基于遥感技术的河道违法构筑物和违章建筑监测的实践[J].城市建筑，2019，16（36）：140-141.
LI Yanting.Practice of monitoring river illegal structures and illegal buildings based on remote sensing technology[J].City Building，2019，16（36）：140-141.
[3] 胡卫，范龙龙，龙洋洲.基于高分辨率卫星影像的违章建筑监测及时空分析[J].国土资源导刊，2019，16（1）：6-10.
HU Wei，FAN Longlong，LONG Yangzhou.Supervision of illegal constructions and analysis of the density change by time and space[J].Land & Resources Herald，2019，16（1）：6-10.
[4] 林剑远，马凌飞，周运杭.基于改进形态学标记分水岭算法的高分辨率遥感影像违章建筑识别研究[J].建设科技，2016（4）：45-47.
LIN Jianyuan，MA Lingfei，ZHOU Yunhang.Recognition of illegal buildings in high resolution remote sensing images based on improved morphological marker watershed algorithm[J].Construction Science and Technology，2016（4）：45-47.
[5] KRIZHEVSKY A，SUTSKEVER I，HINTON G E.Imagenet classfication with deep convolutional neural networks[C]//Advances in Neural Information Processing Systems，2012：1097-1105.
[6] NAIR V，HINTON G E.Rectified linear units improve restricted Boltzmann machines[C]//International Conference on Machine Learning，2010：807-814.
[7] HINTON G E，SRIVASTAVA N，KRIZHEVSKY A，et al.Improving neural networks by preventing co-adaptation of feature detectors[J].arXiv.1207.0580v3，2012.
[8] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2014：580-587.
[9] GIRSHICK R.Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision，2015：1440-1448.
[10] 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.
[11] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：779-788.
[12] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//European Conference on Computer Vision.Cham：Springer，2016：21-37.
[13] BOCHKOVSKIY A，WANG C Y，LIAO H Y M.YOLOv4：optimal speed and accuracy of object detection[C]//IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2020.
[14] 薛芳芳，王月明，李琦.基于特征部位空间关系的牛日常行为识别[J/OL].激光与光电子学进展：1-14[2021-06-09].http：//kns.cnki.net/kcms/detail/31.1690.TN.20210301.1018.
020.html.
XUE Fangfang，WANG Yueming，LI Qi.Recognition of cattle daily behavior based on spatial relationship of feature parts[J/OL].Laser & Optoelectronics Progress：1-14[2021-06-09].http：//kns.cnki.net/kcms/detail/31.1690.TN.
20210301.1018.020.html.
[15] 葛俏，梁桥康，邹坤霖，等.基于轻量化网络与嵌入式的喷码质量检测方法[J/OL].控制工程：1-9[2021-06-09].https：//doi.org/10.14107/j.cnki.kzgc.20200903.
GE Qiao，LIANG Qiaokang，ZOU Kunlin，et al.Detection method of inkjet code quality based on lightweight network and embedded[J/OL].Control Engineering of China：1-9[2021-06-09].https：//doi.org/10.14107/j.cnki.kzgc.
20200903.
[16] 董彪，熊风光，韩夑，等.基于改进Yolov3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 Biao，XIONG Fengguang，HAN Xie，et al.Research on remote sensing building detection based on improved Yolov3 algorithm[J].Computer Engineering and Applications，2020，56（18）：209-213.
[17] 杨焰飞，曹阳.改进YOLOv3的无人机拍摄图玻璃绝缘子检测[J/OL].计算机工程与应用：1-11[2021-09-14].http：//kns.cnki.net/kcms/detail/11.2127.TP.20201224.1102.010.html.
YANG Yanfei，CAO Yang.Detection of glass insulators in images taken by drones based on improved YOLOv3[J/OL].Computer Engineering and Applications：1-11[2021-09-14].http：//kns.cnki.net/kcms/detail/11.2127.TP.20201224.
1102.010.html.
[18] WU D，LV S，JIANG M，et al.Using channel pruning-based YOLOv4 deep learning algorithm for the real-time and accurate detection fapple flowers in natural environments[J].Computers and Electronics in Agriculture，2020，178（5）：174-178.
[19] 谈世磊，别雄波，卢功林，等.基于YOLOv5网络模型的人员口罩佩戴实时检测[J].激光杂志，2021，42（2）：147-150.
TAN Shilei，BIE Xiongbo，LU Gonglin，et al.Real-time detection for mask-wearing of personnel based on YOLOv5 network model[J].Laser Journal，2021，42（2）：147-150.
[20] GAO Shanghua，HAN Qi，LI Duo，et al.Representative batch normalization with feature calibration[C]//IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2021.
[21] 赵振兵，李延旭，甄珍，等.结合KL散度和形状约束的Faster R-CNN典型金具检测方法[J].高电压技术，2020，46（9）：3018-3026.
ZHAO Zhenbing，LI Yanxu，ZHEN Zhen，et al.Typical fittings detection method with faster R-CNN combining KL divergence and shape constraints[J].High Voltage Engineering，2020，46（9）：3018-3026.
[22] ZHOU Daquan，HOU Qibin，CHEN Yunpeng，et al.Rethinking bottleneck structure for efficient mobile network design feature parts[C]//European Conference on Computer Vision，2020：680-697.
[23] 侯涛，蒋瑜.改进YOLOv4在遥感飞机目标检测中的应用研究[J].计算机工程与应用，2021，57（12）：224-230.
HOU Tao，JIANG Yu.Application research of improved-YOLOv4 in remote sensing aircraft target detection[J].Computer Engineering and Applications，2021，57（12）：224-230.
[24] 郭晓静，隋昊达.改进YOLOv3在机场跑道异物目标检测中的应用[J].计算机工程与应用，2021，57（8）：249-255.
GAO Xiaojing，SUI Haoda.Application of improved YOLOv3 in foreign object debris target detection on airfield pavement[J].Computer Engineering and Applications，2021，57（8）：249-255.
[25] 宋艳艳，谭励，马子豪，等.改进YOLOV3算法的视频目标检测[J].计算机科学与探索，2021，15（1）：163-172.
SONG Yanyan，TAN Li，MA Zihao，et al.Video target detection based on improved YOLOV3 algorithm[J].Journal of Frontiers of Computer Science and Technology，2021，15（1）：163-172.