Image Registration Using Conformal Log Polar Mapping

Image Registration is the process of aligning, or overlaying two images of the same scene that were taken at different times and/or from different viewing angles and/or by sensors with different modalities or resolutions. The variations in the imaging environment induce the difference between the images of the same scene. In our situation, we have two images of the same scene taken with two sensors, one in the visible and the other in the infrared (IR) domain. The cameras are placed adjacent to each other on a stable platform, and the images are captured almost simultaneously. This means that the only variation we are dealing with is the sensor modality, and not variations in spatial and temporal coordinates. Using image registration we project the visible image into the IR image to rectify both images to the same co-ordinate system. This is done to match the sensor outputs and then produce an information product from the two sensors, which can be used to further analyze and assess the scene. In this thesis we use the conformal log polar mapping (CLPM) for image registration. The CLPM is invariant to changes in image size, rotation and scale: images of different sizes are mapped to the same CLPM space, rotational changes map to a shift along one of the axes and scale changes a shift long the other. Thus the CLPM for any two images that differ only in rotational and scale changes should be the same. The amount of shift can be used to determine the angle of rotation and the change in scale with respect to the original image. The computed parameters can be used for rectifying the reference image to the original image or vice-versa. In this thesis, a robust approach for image registration is presented, that uses CLPM shifts to compute the parameters representing rotation and scale change, and registers a pair of images. An application of this approach is in multi-sensor fusion that provides illumination-independent surveillance capabilities for day/ night operations.