J.N. Sarvaiya and S.S. Patnaik (India)
Image registration, Control Point, Bilinear Mapping, Radon transform.
This paper presents an automated image registration algorithm suitable for remote sensing, computer vision and image mosaicking applications. In general, such images have different gray-level characteristics and simple techniques such as those based on only correlation cannot be applied directly. The proposed algorithm will be suitable for applications which need to deal with shape information, independent of the orientation, scaling and translation. It removes possible rotational and scale variances before implementing control point matching and translation estimation. It is a two step algorithm. First step deals with intensity based segmentation analysis to obtain possible rotation and scaling information. The mathematical background of this step is based on the radon transform. Images with same scale and without any mutual rotation can be registered efficiently by many available feature based mapping algorithm, which is the second step of the proposed algorithm. As the algorithm is a combination of intensity and feature analysis we call it hybrid algorithm. Simulation results obtained from the proposed hybrid method are comparable with other automatic algorithms with less computational complexity.
Important Links:
Go Back
