摘要
针对基于灰度的图像配准速度较慢的弊端,给出了“图像迹”的定义,通过计算图像的迹差,求累积误差的最小值进行图像配准.通过实验验证,这种方法在保持高的配准概率的前提下,大大地提高了配准速率,降低了算法复杂度,不仅抗噪能力强,而且仅仅只有和差运算,也易于硬件实现.
To overcome the disadvantages of image registration based on image gray scale, the paper first gives the definition of image trace and then realizes a quicker image registration by computing the trace difference of images and getting the least value of accumulative errors. The experiments show that the method not only can keep high registration probability and greatly improve the registration speed, but also can restrain noise powerfully. Since the method only needs plus and minus operations, it can be easily implemented by hardware.
出处
《测试技术学报》
2006年第1期71-74,共4页
Journal of Test and Measurement Technology
基金
山西高校科技研究开发项目(200325)
关键词
图像配准
配准概率
累积误差
image registration
registration probability
accumulation error
