摘要
针对单一单板机的图像分类效率低的缺陷,提出一种基于云计算的图像分类算法。首先通过特征提取算法提取图像的多种特征,然后采用Map/Reduce模型对图像进行匹配和分类,根据匹配结果得到图像的最优分类结果,最后采用Matlab软件实现图像分类的仿真实验。结果表明,相比于其他图像分类算法,该算法减少了图像的分类时间,提高了图像的分类速度,尤其对大规模图像分类优势更加明显。
Aiming at the defect of low image classification efficiency of the single board computer, an image classificationalgorithm based on cloud computing is proposed. The image multiple features are extracted with the feature extraction algorithm,then the Map/Reduce model is used to match and classify the image, and obtain the image optimal classification result accordingto the matching result. The simulation experiment of the image classification was realized with Matlab software. The results showthat, in comparison with other image classification algorithms, the algorithm can reduce the image classification time, improvethe image classification speed, and has obvious superiority especially for the large-scale images classification.
作者
孙沫丽
SUN Moli(Changchun Guanghua University, Changchun 130033, China)
出处
《现代电子技术》
北大核心
2017年第1期57-60,共4页
Modern Electronics Technique
关键词
云计算
图像分类
特征提取
特征匹配
cloud computing
image classification
feature extraction
feature matching
