公路視頻實時車輛檢測分類與車流量統計算法
2020年信息技術與網絡安全第3期
查偉偉,白天
(中國科學技術大學 軟件學院,安徽 合肥 230000)
摘要： 公路視頻實時車輛檢測分類與車流量統計是計算機視覺領域的一個經典問題。傳統設置檢測帶法，易漏檢復檢，自動化性不好。基于深度網絡的one-stage算法實時性好，但是經常會把變化的背景、運動的非車輛物體納入其中，同時對光照變化敏感，夜間分類效果不好。因此，提出采用one-stage做目標檢測，并不直接獲取分類結果，而是根據標注框將物體切割出來，去除背景，提升抗背景擾動性能和分類效果；再送入一個經過遷移學習的淺層神經網絡；將分類輸出和目標檢測網絡的位置輸出合并送入一個全圖匹配算法，進行車流量統計。該方法在保障實時性的同時降低了漏檢和復檢率。
中圖分類號：TP181
文獻標識碼：A
DOI： 10.19358/j.issn.2096-5133.2020.03.012
引用格式：查偉偉,白天.公路視頻實時車輛檢測分類與車流量統計算法［J］.信息技術與網絡安全，2020,39（3）：6267,72.
文獻標識碼：A
DOI： 10.19358/j.issn.2096-5133.2020.03.012
引用格式：查偉偉,白天.公路視頻實時車輛檢測分類與車流量統計算法［J］.信息技術與網絡安全，2020,39（3）：6267,72.
Highway video real-time vehicle detection classification and traffic flow statistics algorithm
Zha Weiwei, Bai Tian
(Department of Software Engineering,University of Science and Technology of China,Hefei 230000,China)
Abstract： Realtime vehicle detection,classification and traffic statistics based on road video are classic problems in the field of computer vision.The traditional method of setting the detection belt is prone to missed inspection and reinspection,so the automation performance is not good.The realtime performance of onestage algorithm based on deep network can be guaranteed,but the changing background,moving nonvehicle objects are often included,and the change of illumination is sensitive at the same time,so the classification at night is not good.Therefore,an algorithm is proposed to perform target detection by onestage,and the classification result is not directly obtained. Instead, it cuts out the object according to the bounding box,removes the background, and improves resistance to background disturbances and classification accuracy.Then it is sent to a transfer learning shallow neural network. The classified output and the position output of the target detection network are combined and sent to a full map matching algorithm for traffic flow statistics.While ensuring realtime performance,the rate of missed inspections and reinspections is reduced.
Key words : convolutional neural network;target detection and classification;realtime traffic statistics;YOLOv3 network
0 引言
公路視頻的車輛分類與車流量統計是運動物體目標檢測識別與跟蹤問題，可以通過傳統圖像方法和現代深度網絡實現。傳統圖像方法由于計算量較小，因此實時性相對較高?，F代深度網絡在背景分割、目標分類的準確度上有著壓倒性的優勢。
本文詳細內容請下載：http://www.cowatch.cn/resource/share/2000003187
作者信息：
查偉偉,白天
(中國科學技術大學 軟件學院,安徽 合肥 230000)
此內容為AET網站原創，未經授權禁止轉載。