图像信息检索系统的设计

Design of image information retrieval system

对采用基于图像频率特征匹配和图像识别方法来实现图像信息检索进行了研究。先用小波分析作为特征提起方法,由于小波分析具有多分辨率分析的特点,并且在提取图像频域特征的同时,保留了图像的空间信息,因此用它来提取图像特征;而后根据计算量的大小,分别用改进的二维相关性算法和二维动态规划弹性算法进行图像的匹配和识别;用VC++和Matlab混合编程进行系统构建并进行了测试实验。实验结果表明,系统有良好的检索效果。

Adopting methods that based on the images frequency characteristics matching and pattern recognition to retrieve the relevant information in the images is investigated. First, wavelet analysis is used to extract the image characteristic. Because wavelet analysis has multi-resolution analysis advantages, and it can extract frequency characteristic while keeping space information of the image, so it is used to extract the characteristic of images. Then accord to the computing complexity, the advanced two-dimension correlation algorithm and the two-dimension dynamic programming elastic algorithm are used to match and recognize the images. The system is implemented by using hybrid programming with VC＋＋ and Matlab, and testing experiment is performed. The results of experiment show that the system has good query performance.