摘要
在智能化交通管理中,对道路上监控视频采集的行人和车辆数据进行识别是至关重要的,为此提出基于图像实例分割的行人车辆检测识别方法。图像实例分割能分割出不同对象,建立基于Mask-Rcnn模型的图像分割方法,测试结果表明该方法有效实现了道路上目标实例分割,并且有很高的精确度和鲁棒性。
In intelligent traffic management,it is important to identify pedestrian and vehicle data collected from surveillance video on the road.For this reason,a method of pedestrian vehicle detection and recognition based on image instance segmentation is proposed.Image instance segmentation can segment different objects,and an image segmentation method based on Mask-Rcnn model is established.Test results show that the method effectively implements target instance segmentation on the road,and has high accuracy and robustness.
作者
程远航
吴锐
Cheng Yuanhang;Wu Rui(The College of Science and Technolegy of Guizhou University,Guiyang Guizhou 550003,China)
出处
《信息与电脑》
2020年第6期130-132,共3页
Information & Computer
基金
贵州省科技合作计划项目(项目编号:黔科合LH字[2017]7227号)
贵州大学2017年度学术新苗培养及创新探索专项(项目编号:黔科合平台人才[2017]5788号)。
