室内场景中密集小目标的人数统计方法

A Population Statistics Method of Dense Small Targets Indoor

计算机视觉任务中,密集小目标的人数统计在人群行为分析、资源优化配置、现代安防等室内场景中具有重要的社会意义。现有的密集小目标统计方法存在着诸如目标相互遮挡造成的漏检、检测目标密集产生的错检以及目标小且人脸特征提取不足等问题。针对室内场景中密集小目标的漏检、错检以及特征不足等问题,提出一种基于YOLOv5框架的人数统计模型STO-YOLO。该方法首先在YOLOv5的主干网络加入针对密集小目标的检测模块以提升特征提取能力,然后在特征融合Neck网络中加入小目标检测模块来增强特征融合能力,从而改善远离监控的密集小目标的错检问题;其次引入OTA机制,将标签分配视作最优传输问题,同时结合上下文信息来减少模糊框的个数,从而有效减少目标遮挡产生的误差。在实际教学场景中自建数据集并验证所提方法。实验结果表明,与SOTA方法YOLOv5相比,STO-YOLO检测结果的precision和recall指标均得到了显著提升;相比最新的YOLOv8,recall和mAP等指标也得到了提升,充分验证了所提STO-YOLO方法的有效性。

The number counting of dense small targets in computer vision tasks is socially important in indoor scenarios such as crowd behavior analysis,optimal resource allocation,and modern security.Existing dense small target counting methods have problems such as omission caused by mutual occlusion of targets,misdetection due to dense detection of targets,and small targets and insufficient extraction of face features.Aiming at the problems of omission,misdetection and insufficient features of dense small targets in indoor scenes,we propose a statistical model STO-YOLO based on the YOLOv5 framework,which firstly adds a detection module for dense small targets to the backbone network of YOLOv5 to improve the feature extraction capability,then adds a small target detection module to the Neck network to enhance the feature extraction capability,and then adds a small target detection module to the Neck network to improve the feature extraction capability.The method firstly adds a small target detection module to the backbone network of YOLOv5 to improve the feature extraction capability,and then adds a small target detection module to the feature fusion network to enhance the feature fusion capability,so as to improve the misdetection problem of dense small targets far away from the surveillance;secondly,it introduces the OTA mechanism,which treats the label assignment as the optimal transmission problem,and at the same time combines with the contextual information to reduce the number of fuzzy frames to reduce the error generated by the target obstruction.Self-constructed dataset and validate the proposed method in a real teaching scenario.The experimental results show that compared with the SOTA method YOLOv5,the precision and recall indexes of STO-YOLO detection results are significantly improved;compared with the latest YOLOv8,the recall and mAP indexes are also improved,which fully verifies the proposed STO-YOLO method.