基于尺度制约规则耦合距离约束的图像匹配算法

Image Matching Algorithm Based on Scale Restriction Rule and Distance Constraint

为了提高图像的匹配精度及其鲁棒性,本文提出了基于尺度制约规则耦合距离约束的图像匹配算法.首先,采用箱式滤波器对高斯函数二阶偏导进行逼近,对特征点进行检测;同时,利用特征点对应的空间尺度来建立尺度制约规则,剔除伪特征点.然后,以特征点为中心,形成圆形区域,计算其Haar小波响应,获取特征点的主方向以及特征向量,形成特征描述子.随后,利用特征点的尺度相似性以及角度相似性来建立空间相似法则,完成特征点的匹配.最后,利用特征点欧氏度量的结果,建立距离约束模型,对匹配特征点之间的距离进行约束,剔除错误的匹配特征点.实验结果显示,与当前图像匹配算法相比,本文算法匹配的图像具有更好的匹配准确度及匹配精度.

In order to improve the accuracy and robustness of image matching,an image matching algorithm has been proposed on the basis of scale restriction rule and distance constraint.First of all,the box filter has been used to approximate the two order partial derivative of Gauss function,and the Hessian matrix determinant been obtained to detect the feature points,and the scale restriction rule has been set up by using the spatial scale function corresponding to the feature points,and the pseudo feature points been removed.Then,a feature descriptor has been formed by computing the Haar wavelet in the circular region formed by the feature point as the center.The spatial similarity rule has been established by using the scale similarity of feature points and angle similarity,which has been used to complete the matching of feature points.Finally,the distance constraint model has been established by using the Euclidean distance measurement of feature points.The distance between the matched feature points has been constrained,and the error matching feature points been eliminated.The experimental results show that compared with the current image matching algorithms,the proposed algorithm has better matching accuracy and matching accuracy.