哈希快速图像匹配算法研究

Fast Image Matching Algorithm Base on Hash

如何快速有效地在大量数据中将图片筛选匹配出来,是图像匹配技术研究的重点课题之一。通过分析感知哈希算法及Surf算法各自的优点,提出用感知哈希算法进行初步图片搜索,利用Surf算法提取相似图片局部特征,从而更精准地确定最相似图片,增加图片匹配的鲁棒性。实验结果表明,在对图片进行处理后,哈希快速图像匹配算法仍能快速地从本地图片库中将最相似图片搜索出来。

Nowadays, data and image update each minute. How to match the image quickly and effectively is the key issue in image matching technology and research. By analyzing the merits of the perceptual hash algorithm and Surf algorithm, this paper proposes a preliminary image search using the perceptual hash algorithm, and then uses Surf algorithm to extract the local characteristics of similar images to determine the most similar images more accu-rately and increase the robustness for image matching. Experimental results show, when the image is processed, the Hash Fast Image Matching algorithm can still quickly search the most similar images from the local gallery.