局部响应函数耦合特征匹配的图像拼接

摘要

In view of shortage that the current image mosaic mechanism is a lack of universal applicability for parameters,and the stitching is an object with the global angular point,leading to low speed of splicing and be difficult to eliminate the false corners,in this paper,se-lect the largest local Harris response function value as the effective angular point and remove response function value adjustment factor, designing a new Harris corner detection operator. Introducing the normalized cross-correlation function,considering geometric angular point distribution features,coupling space transformation model,accurate feature point matching algorithm is designed. Finally propose the image splicing mechanism of precision coupling feature matching of local Harris response function,and test it with the help of simulation experiment. Experimental results show that compared with the current image splicing mechanism,the mechanism proposed has higher pre-cision of mosaic,stitching is complete,effectively eliminating the false corners,and splicing speed is higher.%针对当前图像拼接机制对参数缺乏普遍适用性，且其以全局角点为对象展开拼接，导致拼接速度慢，以及难以消除伪角点等不足，文中选取最大局部范围Harris响应函数值为有效角点，并去除Harris响应函数值调节因子，设计了新的Harris角点检测算子。引入归一化互相关函数，考虑角点分布几何特性，耦合空间变换模型，设计了特征点精准匹配算法。最后提出了局部范围Harris响应函数耦合精准特征匹配的图像拼接机制，并借助仿真实验测试该机制。结果表明：与当前图像拼接机制相比，文中机制具有更高的拼接精度，缝合处完整，有效消除了伪角点，且拼接速度更高。