摘要
基于深度学习技术设计了交通场景下物体的视觉检测方法。首先基于深度卷积对抗生成网络构建了交通场景数据集,基于faster R-CNN设计了交通场景物体检测模型。上述模型利用卷积神经网络提取图像特征。采用区域推荐网络定位目标物体在图像中的具体位置,并通过卷积层在已定位区域提取的特征识别物体的具体类别。最后在不同光照环境下,测试了所提出方法的物体检测效果。测试结果表明,所设计交通场景物体检测模型获得了较好的检测结果。
In order to apply object detection technology to traffic scene, object detection model of traffic scene was studied in this paper. Firstly, a traffic scene dataset was constructed based on deep convolution generative adversarial network. And then, a traffic scene object detection model based on faster R - CNN was designed. The model the convolution neural network used to extract the feature of the image data, and used the region proposal network to locate the specific position of the object to be detected in the image, and distinguished the specific category of the feature extracted by the convolution layer in the locating area. Finally, under different illumination environment, the object detection effect of traffic scene detection model was tested. The test results show that the traffic scene detection model can achieve better detection results.
作者
毕松
刁奇
孙贵宾
韩存武
BI Song, DIAO Qi, SUN Gui -bin, HAN Cun- wu(Beijing Key Laboratory of Fieldbus Technology and Automation, North China University of Technology, Beijing 100144, China)
出处
《计算机仿真》
北大核心
2018年第10期193-197,共5页
Computer Simulation
基金
北京市教科委立项项目(KM201610009001)
关键词
交通场景
物体检测
卷积神经网络
Traffic scene
Object detection
Convolutional neural network
