连续形变图像动态相似信息的自适应识别仿真  

Adaptive Recognition Simulation of Dynamic Similar Information of Continuous Deformation Image

作　　者：李丽[1] 庄庆华[1] LI Li;ZHUANG Qing-hua(College of Humanities&Information,Changchun University of Technology,Changchun Jilin 130000,China)

机构地区：[1]长春工业大学人文信息学院,古林长春130000

出　　处：《计算机仿真》2021年第10期414-418,共5页Computer Simulation

摘　　要：针对连续图像相似信息识别过程适应度较差的问题,提出基于特征矩阵的动态相似信息自适应识别方法。通过图像预处理凸显出图像最原始特质,剔除冗余信息。在连续形变图像中,考虑其被遮盖的范围不能使用当前帧的像素动态信息进行更新,采取参照原本像素信息与灰度值融合方式更新像素,实现图像可变光强背景区域的更新。计算连续形变图像特征矩阵相似度,完成基于特征矩阵的动态相似信息自适应识别。仿真结果证明,上述方法能够同时满足图像相似信息的完整、高效识别,真实监控图像的实验结果验证了论文方法能准确地识别连续形变图像相似信息,说明研究成果具有更优的适应性。In this paper,an adaptive recognition method of dynamic similarity information based on a feature matrix was proposed.The image preprocessing highlighted the most original features of images and eliminated redundant information.In the continuous deformation image,the covered area could not be updated by the dynamic information of the current frame,so original pixel information was integrated with gray value to update the pixel and thus to realize the update of the variable optical background area of an image.The similarity between feature matrices was calculated.Thus,the adaptive recognition for dynamic similarity information based on the feature matrix was completed.Following conclusions can be drawn from simulation results:the proposed method can simultaneously satisfy the requirements of complete and efficient recognition of the image similar information.The experimental results of real monitoring images prove that the proposed method can accurately identify the similarity of continuous deformational images.The research results have better adaptability.

关 键 词：连续形变图像 动态相似 自适应识别 融合差分 平均灰度值 

分 类 号：TP310[自动化与计算机技术—计算机软件与理论]