An object detection algorithm for complex urban scenarios based on YOLOv11

Wenshuai Yuan, Changming Zhu
Proceedings of 2025 2nd International Conference on Machine Learning and Intelligent Computing, PMLR 278:331-340, 2025.

Abstract

In recent years, with the rapid development of fields such as autonomous driving and intelligent transportation, the detection of pedestrians and vehicles in complex urban scenarios has become a hot topic in the field of object detection. However, these complex urban scenarios pose significant challenges to object detection. This paper proposes an improved algorithm based on YOLOv11, namely the YOLOv11 - APAS - MDC algorithm for object detection in complex urban scenarios on the Urban Environment Detection dataset. The aim is to enhance the accuracy and robustness of detecting pedestrians and vehicles under conditions such as occlusion and multi - scale targets in complex urban environments. This paper proposes a multi - scale edge information enhancement module called APAS based on the YOLOv11 basic model. This module highlights important edge feature information and can improve the model’s perception ability of multi - scale features. Secondly, this paper presents the MDC module. By using convolutional layers with different dilation rates, this module can extract features at different scales. In addition, this paper introduces the RepGFPN feature network. This network re - parameterizes the structure and reduces redundant operations. Through a more complex cross - layer connection mechanism, it enhances the interaction of features at different levels, thereby improving the performance and efficiency of the detection model. The experimental verification results on the Urban Environment Detection dataset show that the improved algorithm in this paper outperforms the traditional YOLOv11 algorithm in terms of detection accuracy and robustness under conditions such as occlusion and multi - scale targets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v278-yuan25a, title = {An object detection algorithm for complex urban scenarios based on YOLOv11}, author = {Yuan, Wenshuai and Zhu, Changming}, booktitle = {Proceedings of 2025 2nd International Conference on Machine Learning and Intelligent Computing}, pages = {331--340}, year = {2025}, editor = {Zeng, Nianyin and Pachori, Ram Bilas and Wang, Dongshu}, volume = {278}, series = {Proceedings of Machine Learning Research}, month = {25--27 Apr}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v278/main/assets/yuan25a/yuan25a.pdf}, url = {https://proceedings.mlr.press/v278/yuan25a.html}, abstract = {In recent years, with the rapid development of fields such as autonomous driving and intelligent transportation, the detection of pedestrians and vehicles in complex urban scenarios has become a hot topic in the field of object detection. However, these complex urban scenarios pose significant challenges to object detection. This paper proposes an improved algorithm based on YOLOv11, namely the YOLOv11 - APAS - MDC algorithm for object detection in complex urban scenarios on the Urban Environment Detection dataset. The aim is to enhance the accuracy and robustness of detecting pedestrians and vehicles under conditions such as occlusion and multi - scale targets in complex urban environments. This paper proposes a multi - scale edge information enhancement module called APAS based on the YOLOv11 basic model. This module highlights important edge feature information and can improve the model’s perception ability of multi - scale features. Secondly, this paper presents the MDC module. By using convolutional layers with different dilation rates, this module can extract features at different scales. In addition, this paper introduces the RepGFPN feature network. This network re - parameterizes the structure and reduces redundant operations. Through a more complex cross - layer connection mechanism, it enhances the interaction of features at different levels, thereby improving the performance and efficiency of the detection model. The experimental verification results on the Urban Environment Detection dataset show that the improved algorithm in this paper outperforms the traditional YOLOv11 algorithm in terms of detection accuracy and robustness under conditions such as occlusion and multi - scale targets. } }
Endnote
%0 Conference Paper %T An object detection algorithm for complex urban scenarios based on YOLOv11 %A Wenshuai Yuan %A Changming Zhu %B Proceedings of 2025 2nd International Conference on Machine Learning and Intelligent Computing %C Proceedings of Machine Learning Research %D 2025 %E Nianyin Zeng %E Ram Bilas Pachori %E Dongshu Wang %F pmlr-v278-yuan25a %I PMLR %P 331--340 %U https://proceedings.mlr.press/v278/yuan25a.html %V 278 %X In recent years, with the rapid development of fields such as autonomous driving and intelligent transportation, the detection of pedestrians and vehicles in complex urban scenarios has become a hot topic in the field of object detection. However, these complex urban scenarios pose significant challenges to object detection. This paper proposes an improved algorithm based on YOLOv11, namely the YOLOv11 - APAS - MDC algorithm for object detection in complex urban scenarios on the Urban Environment Detection dataset. The aim is to enhance the accuracy and robustness of detecting pedestrians and vehicles under conditions such as occlusion and multi - scale targets in complex urban environments. This paper proposes a multi - scale edge information enhancement module called APAS based on the YOLOv11 basic model. This module highlights important edge feature information and can improve the model’s perception ability of multi - scale features. Secondly, this paper presents the MDC module. By using convolutional layers with different dilation rates, this module can extract features at different scales. In addition, this paper introduces the RepGFPN feature network. This network re - parameterizes the structure and reduces redundant operations. Through a more complex cross - layer connection mechanism, it enhances the interaction of features at different levels, thereby improving the performance and efficiency of the detection model. The experimental verification results on the Urban Environment Detection dataset show that the improved algorithm in this paper outperforms the traditional YOLOv11 algorithm in terms of detection accuracy and robustness under conditions such as occlusion and multi - scale targets.
APA
Yuan, W. & Zhu, C.. (2025). An object detection algorithm for complex urban scenarios based on YOLOv11. Proceedings of 2025 2nd International Conference on Machine Learning and Intelligent Computing, in Proceedings of Machine Learning Research 278:331-340 Available from https://proceedings.mlr.press/v278/yuan25a.html.

Related Material