SURF算法和RANSAC算法相结合的台标检测与识别  

TV Logo Detection and Recognition Based on SURF Algorithm and RANSAC Algorithm

作　　者：吴月凤[1] 何小海[1] 张峰[1] 滕奇志[1] 

机构地区：[1]四川大学电子信息学院,四川成都610064

出　　处：《电视技术》2014年第13期191-195,共5页Video Engineering

基　　金：国家自然科学基金委员会和中国工程物理研究院联合基金项目(11176018)

摘　　要：通过分析电视台台标的特点,首先对视频中的某一帧进行感兴趣区域检测,提出了对感兴趣区域进行基于SURF特征提取、并与台标库中的台标特征点进行预匹配、用RANSAC算法去除错误匹配点对的方法。结合了SURF算法与RANSAC算法的优点,提高台标识别的正确率。实验结果表明,该方法具有较高的识别精度,较强的抗干扰能力,识别速度也在一定程度上有所提高。A novel method is proposed in this paper by analyzing the features of TV logos. Firstly, the ROI of a certain frame in the video is detected, then SURF（ speeded up robust feature） algorithm is introduced to extract features of the ROI ,the result of which is used to do prematching with the TV logos in the database and mismatching is removed by RANSAC（random sample consensus） algorithm. Combining the advantages of SURF algorithm and RANSAC algorithm,the recognition accuracy is improved. The results show that this method has higher recognition accuracy and stronger anti - disturb- ance capability,the recognition speed is improved to some extent as well.

关 键 词：电视台标识别 感兴趣区域 SURF RANSAC 

分 类 号：TN957[电子电信—信号与信息处理]