违章停车检测及识别算法.pdfVIP

40 1 吉林 大学 学 报 ( 工学 版) Vol. 40 No. 1 2010 1 Journal of Jilin University (Engineering and Technology Edition) Jan. 2010 王殿海, 胡宏宇, 李志慧, 曲昭伟 ( 吉林大学 交通学院, 长春 130022) : 针对传统违章停车人工检测方式的缺陷, 设计了基于 像处理技术的停车违章监控算 法在禁停路段区域设置视觉传感器采集视频 像序列, 利用自适应的混合高斯模型实现复 杂交通场景下的背景抽取, 提取可能运动前景目标利用像素级时间序列特征检测静止物体, 并根据对象级区域特征实现停驶车辆的辨识, 获取车辆的违章停车信息根据不同禁停区域 的具体违章要求实现自动警报最后, 通过实际交通场景视频序列对算法进行了验证, 结果表 明了本文方法的有效性 : 交通运输系统工程; 智能交通; 视频监控; 违章停车; 背景模型 : U 121, TP391 : A : 2010) Detectionandrecognitionalgorithm of illegalparking WANG Dianhai, HU Hongyu, LI Zhihui, QU Zhaow ei (College of Transp ortation, J ilin University, Changchun 130022, China) Abstract:To overcome the shortcoming of the manual detection of illegal parking cars, a surveillance algorithm w as developed to detect the illegal parking cars based on the image processing technology. A visual sensor w as set at the forbidden parking region of the road to record the video sequence. An adaptive mixed Gaussian model w as used to extract the background in the complicated environment. The still objects w ere detected by the temporal sequence feature of the pixel level. The still vehicles w ere recognized by the regional feature of the object level to get the illegal car parking information. The autoalarm w as realized according to the requirements from the regulation of different forbidden parking regions. The suggested algorithm w as validated by the video sequence of the real traffic scenes to be effective. Keywords:engineering of communications and transportation system; intelligent transportation; video surveillance; illegal parking; background model