摘要
运用传统方法对超分辨图像的局部特征进行识别时,存在识别准确率较低和识别时间较长的问题,为此给出一种基于角点检测的图像局部特征识别方法。运用细节特征定位方法对图像的局部信息进行采样,采用灰度直方图重构方法对图像进行三维重建,将重建结果与边缘轮廓特点融合方法相结合,对图像的局部边缘信息特征进行提取。根据Harris角点检测方法提取图像的局部信息特征量,并依据卷积神经网络对提取结果进行分类,最终实现对超分辨图像局部特征的识别。分析实验结果可知,与传统方法相比,此方法的识别准确率较高,并且识别时间远低于传统方法。
Traditional methods of recognizing local features of super-resolution images have such problems as low recognition accuracy and more recognition time. Therefore,a method of image local feature recognition based on corner detection is proposed. The local information of the image is sampled by the method of locating minutiae features,and the image is reconstructed by the method of gray histogram reconstruction. The local edge information features of the image are extracted by combining the reconstructed results with the fusion method of edge contour features. According to Harris corner detection method,the local information feature of the image is extracted,and the extracted results are classified according to convolution neural network. Finally,the local feature of super-resolution image is recognized. The experimental results show that,compared with the traditional method,the recognition accuracy of this method is higher and the recognition time is much less than that of the traditional method.
作者
阮文惠
黄珍
RUAN Wenhui;HUANG Zhen(School of Digital Media,Lanzhou University of Arts and Science,LanZhou 730000,China)
出处
《西安工程大学学报》
CAS
2019年第1期106-110,115,共6页
Journal of Xi’an Polytechnic University
基金
甘肃省自然科学基金(17JR5RA007)
兰州文理学院种子基金(自然)(17XJZZ005)
关键词
角点检测
图像处理
超分辨
特征提取
特征识别
corner detection
image process
super resolution
feature extraction
feature recognition
