Research On Mosaic Technology Of Space Camera Multi-Channel Output Image

Space camera which designed as sub-regional imaging using multi TDICCD and multi-channel output has to mosaic separate channel images into a complete whole view picture, a new mosaic method is required in this particular case. The thesis analyses the image registration and fusion algorithms in the mosaic process. A mosaic method which suits to multi-TDICCD stitching space camera multi-channel output images is proposed, it depends on the structure and imaging model of camera for coarse matching, then makes the matching result more accurate based on the image content processing while the search space has been reduced, finally fuses the image to finish the mosaic work.In this thesis, an imaging model is built based on the multi-TDICCD stitching structure and principle of optical imaging. Using the model to analyses image motion changes under different attitudes, and with the time delay integration characteristic of TDICCD, getting the corresponding relationship of row between channel images under different side swing angle and TDICCD control, calculating the range of overlap region between channel images under the drift angle control accuracy. By comparison among three typical gray's similarity measures, mutual information registration is chosen. In order to overcome the shortcomings of mutual information that only uses the information of pixel gray and easy to appear partial extreme value, an improved algorithm is proposed. Edge information is added into the new method, used to template self-adaptive selecting. Mutual information registration integrated with edge membership grade has strong robustness, and gets pixel matching accuracy. After the image fusing processing, it gets a complete whole view picture.Bases on the integrated simulation testing platform of space camera and the high speed real-time image date recording system, a multi-channel images mosaic software has been designed and finished. It successful mosaics the images from multi-channel in the field real imaging test of space camera. The proposed method is stable, can meet the space camera multi-channel output images mosaic applications, and gets a satisfactory mosaic result visually.