A frame center matching technique for precise registration of multitemporal airborne frame imagery

Accurate spatial registration between multitemporal imagery is necessary if pixel-level changes are to be detected. Registration of multitemporal airborne frame imagery is complicated by image distortions resulting from wide view angles and variable terrain. A novel technique for acquiring and precisely registering multitemporal airborne frame imagery is presented. This technique, referred to as frame center (FC) matching, may enable precise registration between a long series of multitemporal airborne imagery. Polynomial warping and orthorectification algorithms were tested for registering multitemporal imagery acquired with matched FCs. Registration results were compared to those derived using imagery not acquired with matched FC locations. The FC matching approach to image acquisition yielded substantially lower misregistration errors between all multitemporal image sets for each method evaluated. Special tools and protocols required for operationally replicating FCs in time sequential image acquisitions were evaluated. The effectiveness of these tools and protocols for frame-center matched acquisition and registration of multidate imagery is demonstrated.