戴光明, 张全元, 包建全. 一种车型特征提取的新算法[J]. 武汉大学学报 ( 信息科学版), 2009, 34(10): 1155-1158.
引用本文: 戴光明, 张全元, 包建全. 一种车型特征提取的新算法[J]. 武汉大学学报 ( 信息科学版), 2009, 34(10): 1155-1158.
DAI Guangming, ZHANG Quanyuan, BAO Jianquan. A New Algorithm of Vehicle Features Extraction[J]. Geomatics and Information Science of Wuhan University, 2009, 34(10): 1155-1158.
Citation: DAI Guangming, ZHANG Quanyuan, BAO Jianquan. A New Algorithm of Vehicle Features Extraction[J]. Geomatics and Information Science of Wuhan University, 2009, 34(10): 1155-1158.

一种车型特征提取的新算法

A New Algorithm of Vehicle Features Extraction

  • 摘要: 在对实时视频流分析的基础上,提出了一种新的算法———弹性松弛袋算法来对车辆的最显著的特征———车长进行提取。在利用弹性松弛袋算法提取车长特征的同时,确定了关键帧图像,对车辆存在进行了判定。并为了提高车长提取精度,给出了弹性追尾算法以实现对弹性松弛袋算法的改进。实验表明,弹性松弛袋算法具有良好的统计特性,能够有效地提取视频的关键帧,而且提取车长特征准确、灵敏,能够满足系统的实时性要求,可应用到图像识别的其他领域。

     

    Abstract: A new algorithm-flexible relaxation bag method for vehicles to the most notable features-car length extracted.Flexible use of relaxation algorithm models feature extraction process,but also the key image frames selective extraction process.At the same time,the process is to determine the existence of vehicles.In this paper,the flexible relaxation bag algorithm is updated.A updated algorithm called elastic collision algorithm has further precision in some situation.The experiments show that the flexibility relaxation method has better statistics feature,to extract effectively video key frame.The method extract features accurately,is able to effectively meet the demands of the real-time system.It is also applied in other photograph fields.

     

/

返回文章
返回