摘要
传统视频图像帧特征识别方法主要基于三维模型来完成,但由于图像识别的复杂性,导致很难构建准确的三维模型,所以基于模型的方法只能在理想条件下具有较高识别率。为了提升视频图像动态识别的实际应用,提出了基于视觉传达的视频图像帧特征动态识别方法。通过灰度变换和阈值化处理方法对视频图像进行预处理,根据小波变换分析预处理后的灰度视频图像中灰度与细节的特点,利用灰度投影法对视频图像进行帧特征提取,在此基础上,采用以K-L(Karhuncn-Loeve)变换为核心的局部识别方法,最大限度上保留了完整的图像信息,完成了视频图像帧特征的动态识别。仿真结果表明,该方法具有较高的识别率和精准度,且识别时间较短,满足了实际应用需求。
The traditional frame feature recognition method is mainly based on the three-dimensional model,but it is difficult to construct an accurate three-dimensional model due to the complexity of image recognition.In order to improve the real application of video image dynamic recognition,this article presented a dynamic recognition method for video image frame feature based on visual communication.Firstly,the video image was pre-processed by the gray level transformation and threshold method.According to the wavelet transform,the features of gray level and detail in pre-processing gray video image were analyzed.After that,the gray projection method was used to extract the frame feature of video image.The local recognition method based on K-L(Karhuncn-Loeve)transform was adopted to preserve the complete image information to the maximum extent.Finally,the dynamic recognition for video image frame features was finished.Simulation results show that the proposed method has high recognition rate and precision.Meanwhile,the recognition time is short.Therefore,the actual application requirement can be satisfied by the proposed method.
作者
王卓
WANG Zhuo(College Of Optical and Electronical Information Changchun University of Science and Technology,Jilin Changchun 130000,China)
出处
《计算机仿真》
北大核心
2020年第7期455-458,479,共5页
Computer Simulation
关键词
三维模型
灰度投影法
光学图像特征
加权映射法
Three-dimensional model
K-L transform
Gray projection method
Optical image feature
Weighted mapping method
