详细信息
基于Faster RCNN的智能车道路前方车辆检测方法
Forward Vehicle Detection Method of Intelligent Vehicle in Road Based on Faster RCNN
文献类型:期刊文献
中文题名:基于Faster RCNN的智能车道路前方车辆检测方法
英文题名:Forward Vehicle Detection Method of Intelligent Vehicle in Road Based on Faster RCNN
作者:史凯静[1];鲍泓[1];徐冰心[1];潘卫国[1];郑颖[1]
第一作者:史凯静
机构:[1]北京联合大学北京市信息服务工程重点实验室
第一机构:北京联合大学北京市信息服务工程重点实验室
年份:2018
卷号:44
期号:7
起止页码:36-41
中文期刊名:计算机工程
外文期刊名:Computer Engineering
收录:CSTPCD;;Scopus;北大核心:【北大核心2017】;CSCD:【CSCD_E2017_2018】;
基金:国家自然科学基金"视听觉信息的认知计算"重大研究计划重点支持项目"智能车驾驶脑认知技术;平台与转化研究"(91420202);北京市教委科研计划项目(KM201811417006)
语种:中文
中文关键词:智能车;前方车辆;深度卷积神经网络;训练模型;准确率
外文关键词:deep Convolution Neural Network ( CNN );training model;[Keywords ] intelligent vehicle;forward vehicle;accuracy rate
摘要:使用Fast RCNN方法进行特征提取存在耗时较长且检测准确率较低的问题。为此,结合Faster RCNN前方车辆检测模型与3种不同大小的卷积神经网络,提出一种改进的前方车辆检测方法,研究对比各方法在3种交通场景数据库上的前方车辆检测能力。实验结果表明,与深度卷积神经网络方法相比,该方法提高了车辆检测的准确性和鲁棒性,具有一定的泛化能力。
Using Fast RCNN method for feature extraction takes a long time and the detection accuracy is low,this paper proposes an improved forward vehicle detection method that combines the forward vehicle detection model based on Faster RCNN with Convolution Neural Network( CNN) of three different sizes. The forward vehicle detection ability of different methods on three traffic scene databases is researched and compared. Experimental results show that compared with the deep CNN,this method improves the accuracy and robustness of vehicle detection,and has a generalization ability.
参考文献:
正在载入数据...