面向路面抛洒物检测的多任务学习算法研究
Research on Multi-task Learning Algorithms for Pavement Spills Detection
井晶 1赵广明 2赵作鹏2
作者信息
- 1. 江苏联合职业技术学院 徐州财经分院,江苏 徐州 221008;中国矿业大学 计算机科学与技术学院,江苏 徐州 221116
- 2. 中国矿业大学 计算机科学与技术学院,江苏 徐州 221116
- 折叠
摘要
针对现有检测算法在定点抛洒物检测方面的不足,提出一种基于多任务学习和图像处理的路面抛洒物检测算法.首先,为YOLOv6s设计一种融合多层语义特征的语义分割头,并为目标检测分支和语义分割分支设计损失函数,进行多任务学习;其次,结合语义分割分支分割连续性的优势以及图像处理的方法,提取完整路面区域;最后,引入结合IoU和质心位置的抛洒物区域识别与排除方法,实现非抛洒物的排除和最终抛洒物区域提取.实验结果表明,在BDD100k数据集上,改进算法在车辆目标检测和可行驶区域分割上分别达到 77.8%的平均精度值(mAP)和91.5%的平均交并比(mIoU).
Abstract
To address the shortcomings of existing detection algorithms in detecting road-spill objects,a no-vel approach based on multi-task learning and image processing for road-spill object detection is proposed.Firstly,a semantic segmentation head that integrates multi-layer semantic features is designed for YOLOv6s,and loss functions are devised for both the object detection and semantic segmentation branches to enable multi-task learn-ing.Secondly,by leveraging the advantage of continuity in the semantic segmentation branch,coupled with im-age processing techniques,the complete road area is extracted.Finally,a method for identifying and excluding spilled object regions is introduced,which combines IoU and centroid position,enabling the exclusion of non-spilled objects and the final extraction of spilled object regions.Experimental results demonstrate that on the BDD100k dataset,the improved algorithm achieves a mean Average Precision(mAP)of 77.8%for vehicle ob-ject detection and a mean Intersection over Union(mIoU)of 91.5%for drivable area segmentation.Additionally,spilled objects of various scales can be accurately extracted.The algorithm can be directly deployed on vehicle-mounted equipment for existing road damage inspection tasks.
关键词
多任务学习/抛洒物检测/YOLOv6s/图像处理/语义分割Key words
multi-task learning/debris detection/YOLOv6s/image processing/semantic segmentation引用本文复制引用
出版年
2024
NETL
NSTL
万方数据