基于SURF的图像配准技术可行性研究

Feasibility Research of image registration technology based on SURF

当前基于特征的图像配准算法较多,而SURF算法由于其性能优越和运算效率较快,被众多学者广泛应用于图像匹配算法的研究中。在图像匹配算法研究的基础上,通过比较SURF算法和Harris算法在通用型、计算效率和配准精度三个方面的差异,验证SURF算法在图像配准应用中的可行性。实验结果及数据表明基于SURF的图像配准技术不仅能够适用于不同条件下的图像变化,并且可以实现快速、精确的图像配准。

Currently,there are many image registration algorithms based on characteristic. However,the SURF algorithm is widely used in the research of image matching by many scholars because of its superior performance and fast computational ef？ficiency. On the basis of the research of image matching algorithm,the feasibility of the SURF applied to the image registration is confirmed in this paper by means of comparing the differences of universality,computational efficiency and registration preci？sion between SURF and Harris. The experimental results and data indicate the image registration technology based on SURF not only can apply to image changes under different conditions,but also achieve the fast and accurate image registration.