Research On Image Processing And Stitch Of CE-3

The detector will realize a soft landing and inspection in Lunar probe2. As oneof the most important load, The panoramic camera will achieve the goal ofresearching the structure and the morphology of the moon surface. Image stitchingtechnique for the panoramic camera was done in this article. Main contents of thisarticle were image distortion correction, image restoration, image enhancement andimage stitching.Image distortion was a important factor for the Panorama.But strictrequirements was executed in the design, manufacture and detect Process of theCE-3panoramic camera, Actually it’s distortion even can reach thousandthsaccuracy,so we didn’t do the distortion correction work for CE-3camera in thisarticle,but a high precision method of distortion correction was also given for thesituation of the image have large distortion.In the part of the image restoration,we first discussed the degradation model ofthe image,and finally a adaptive total variation image deblurring algorithm wasintroduced in this paper.we applied this new algorithm for the image of the CE-3.Itturned out that the clear quality, entropy and average gradient were substantiallyincreased.In the part of the image enhancement, we adopted the Retinex algorithm,whichwas widely used in the procession of the aerospace image. A enhance algorithm wasproposed for the images of CE-3camera based on the Retinex theory.A indexenhance algorithm was also pulled in for the purpose of meeting the vision feature ofthe human eye,in this algorithm, Noise was considered,and it was controled thoughtfully. The test of the image of the CE-3camera showed that it couldeffectively enhance the clear quality and average gradient.In chapter5, the image stitching algorithm for the two circles images of CE-3camera was discussed detail. A better algorithm was adopted which based on cornerdetect, a matching principle for SIFT coner based on system similarity wasproposed,which showed better behavior, finally it was used to the matching part ofthe stitching algorithm; Because strict demand was executed when capture images,Considering the time and accuracy, affine transformation was adopted; thenPROSAC and LM algorithm was used for the solution of transformation model;Based on the algorithm discussed above, two circle images of the CE-3camera werestitched, and the result was given in the last of the paper.