基于投影特征的自适应图像匹配算法

针对传统图像匹配算法存在的计算量大、匹配速度缓慢的缺点，对基于相关函数法的投影法进行了改进，提出了一种基于投影特征的自适应图像匹配算法。得到投影序列后，选定初始阈值，在后续各点的计算中不断自适应更新，遵循超过当前阈值即停止匹配计算的原则，大大减少了对非匹配点的计算。实验结果表明，该算法抗干扰性强，匹配速度快；模板图像和待匹配图像大小越接近，该算法精度越高，速度优势越明显。

Algorithm for images matching based on projection with adaptive threshold

Aiming at the disadvantage of slow matching velocity existing in traditional image matching algorithms, projection based on correlation coefficient algorithm was improved to eliminate the complicated calculations. An algorithm for images matching was proposed based on projection with adaptive threshold. The initial threshold was set according to the error on the first loaction of macting region. If the error was lower than the current value , the threshold would be updated, otherwise calculation must be stopped, which can eliminate the calculation on non matching points. The experiment result shows that the improved algorithms can reduce the influence of noises and increase the matching velocity greatly. The algorithm becomes more efficient as the size of the template image gets closer to that of the source image.