基于视频的道路识别新算法在交通监测系统中的应用  

APPLYING VIDEO-BASED NEW ALGORITHM OF ROAD RECOGNITION IN TRAFFIC SURVEILLANCE SYSTEM

作　　者：郇洪江[1] 宫宁生[1] 胡斌[1] 

机构地区：[1]南京工业大学信息科学与工程学院,江苏南京210009

出　　处：《计算机应用与软件》2010年第6期216-218,共3页Computer Applications and Software

摘　　要：针对传统的道路识别方法过程复杂,易受干扰,识别效果不够理想等缺点;根据道路上会有大量车辆通过,而车辆通过的地方也必为道路这一特点。采用逆向思维,通过识别运动物体达到识别静态物体的目的,提出一种应用累积帧差技术通过识别道路上经过的车辆来识别道路位置的新算法。实验结果表明使用新算法识别出的道路效果比较理想,可以满足实际应用的需要。Traditional road recognition algorithm is complex in procedure,susceptible to disturbance and unsatisfied in recognising result.To resolve the above problems,according to the characteristic of the road that passing vehicles are of great deal and where there vehicles are passing have to be the road,a contrary thoughts,in which the recognition of static objectives is achieved by the recognition of dynamic objectives,was applied in presenting a new algorithm of road location recognition by recognising the vehicles passing on the road with the technology of accumulated frame differences.Experimental results showed that the road recognition result acquired with the new algorithm is of preferable ideal and can meet the requirements in practical application.

关 键 词：道路识别 累积帧差技术 阈值分割 图像平滑 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]