摘要
针对高速公路环境下的违章停车事件,提出了一种基于视频图像处理技术的检测方法。通过混合高斯模型对采集的视频序列实现背景建模,利用背景差分法获得前景目标,同时使用背景更新方法减小环境因素对背景模型的影响。通过提出的目标质心检测法实现划定区域内的运动车辆检测,对违章停车目标进行标定识别并报警。最后,通过实验对某高速公路交通视频图像进行算法验证,结果表明该算法场景适应性强、实时性好、检测准确率高。
In order to solve the problem of illegal parking on highway, a detection method based on video image processing technology was proposed. Ganssian mixture model was used to collect the video sequence to realize background modeling. The back-ground difference method was used to obtain the target foreground object and the background update method was also used to reduce the influence of environmental factors on the background model Through the proposed target centroid method, not only the moving vehicle detection in designated area can be realized but also the illegal parking targets can be identified and alerted. Finally, the results of the experiment which have been done on a certain highway to verify the video sequence of the highway traffic scenes show that the proposed algorithm has the merits of strong scene adaptability, good efficiency and high veracity.
出处
《重庆交通大学学报(自然科学版)》
CAS
北大核心
2014年第4期138-141,共4页
Journal of Chongqing Jiaotong University(Natural Science)
基金
陕西省火炬计划项目(2010HJ-22)
关键词
交通工程
智能交通
目标质心检测法
混合高斯模型
停车检测
高速公路
traffic engineering
intelligent transportation
target centroid detection method
Ganssian mixture model
parking detection
highway