Sequence Of Aerial Images Based On D-H Theory Of Key Element Consistency

Unmanned aerial technology with the accelerated development of science andtechnology more and more widely attention and practical applications, unmanned aerialsystem has a flexible, simple operation and maintenance and low cost etc. With thematuring of unmanned aerial vehicle technology, the wide application of unmannedaerial vehicle advantages in the field of aerial photography, especially the UAV remotesensing platform as a kind of flexible obtained widespread application in recent years,especially in the rapid access to small regions or aerial environment difficult areas ofimage has very obvious advantage, can undertake high risk or high-tech flight aerialmission.Small volume, lighter weight is the most obvious feature of the unmanned aerialvehicle, and because of the unmanned aerial vehicle load capacity and the limitation ofthe engine room condition such as size, make the whole flight aerial system is generallynot suitable for using heavy large aerial equipment, so the whole weight of the unmannedaerial vehicle flight platform is relatively light, in the process of its flight is vulnerable tothe influence of the external environment factors such as wind and air flow, cannotguarantee carrying on unmanned aerial vehicle digital camera shots always perpendicularto the ground, because the body's flight not smooth leads to obtain the projection of aerialimage geometric distortion problem, it will directly affect the final quality of the imagematching.In view of the problems arising from the above situation, this paper puts forward akind of unmanned aerial vehicles based on D-H theory of aerial image geometriccorrection method, it is also a key elements air in image sequences to research the mainproblem of consistency. In this paper, first of all, the cause of the UAV aerial imagegeometric distortion is analyzed, and then expounds the principle of micro unmannedaerial vehicle flight attitude angle of the transformation, as well as the used D-Hprinciple are discussed in detail in this paper, the theory of knowledge, finally unifies theD-H theory related to the image geometric correction model is established, and completethe corresponding image geometric correction experiment, finally the experimentalresults show the feasibility and validity of the method, the basis for the next mosaicmatching program. Secondly, this paper some common basic splicing algorithm forimage Mosaic made were discussed in detail, including registration method based ongray level information, based on the transform domain registration and registrationmethod based on feature. Stitching algorithm is used in this paper, based on thecharacteristics of SURF algorithm, with the characteristics and steps of the algorithm hasmade the detailed elaboration, and then through the experiment of after geometriccorrection of ten aerial images sequence for joining together, using MATLABprogramming eventually joining together to complete a basic conform to the actualdemand of panoramic stitching images, the final image mosaic of effect shows that theerror can be better control within a pixel, to satisfy the actual application needs.