【摘要】 针对互信息算法在配准图像过程中易陷入局部极值的问题,提出了一种新的算法,即将互信息和相关比按一定的标准组合在一起,生成一种新的测度算法,并将该算法应用到了多波束图像和侧扫声呐图像的配准中。该组合算法可以改善互信息测度易陷入局部极值的弊端,稳定性更强。通过实验验证,新的组合算法是准确的,可以较好地实现存在较大畸变的多波束和侧扫声呐图像的配准,配准速度和精度优于互信息算法和相关比算法,新的组合算法是对原有算法的改进和优化。更多还原

【Abstract】 For the problem that mutual information algorithm can easily fall into local extremum in the image registration,a new algorithm which combines the mutual information(MI)with the correlation ratio(CR)according to a certain standard is proposed.This new combination algorithm has been applied for the matching of multi-beam images and side scan sonar images.This algorithm can oercome the existing disadvantages of the mutual information measure and has a stronger stability.It has been verified through the experiments that the new combination algorithm is accurate and can realize the registrations both of the multi-beam and of the side scan sonar images which exist a big distortion.Both the speed and the precision of the registration are superior to the mutual information and the correlation ratio algorithms.The new combination algorithm is the improvement and optimization of the original one.更多还原