Comparative Study on Iterative-Optimization-Based Image Registration Algorithms

A detail comparative study is carried out on three representative iterative-optimization-based image registration methods, namely digital image correlation using Newton-Raphson iteration, iterative gray-gradient algorithm and least square image matching, which are originated and developed in the fields of optical experimental mechanics and stereo vision respectively. The mathematical consistency of the three algorithms is first demonstrated. Analysis and comparison of some further conclusions drawn from the three methods are then conducted. As the analysis and comparison show, these further conclusions can explain each other and can be mutually employed among the three methods. However, in view of algorithm implementation and extension, image correlation using Newton-Raphson iteration is more adaptable and flexible.