Study Of UAV Remote Sensing Images Processing Based On Evidence Theory

In recent years, geographic resource exploration and natural disasters wrecker task,unmanned aerial vehicles and gradually evolved into a new type of remote sensingplatforms, unmanned aerial vehicles, compared with the satellite remote sensing andaerial remote sensing, UAV remote sensing with real-time high, the cost is relativelylow mobility, excellent anti-interference performance, weather conditions, requiring alower level, as well as remote sensing data of high quality advantages of remote sensingdata collected by unmanned aerial vehicles, you can achieve the quality of theinformation needed to establish a remote sensing system. For this reason, this papermakes the choice of their real reasons for the UAV remote sensing image processing asthe direction.In this paper, the purpose is the original remote sensing image information to acertain degree of simplification and screening level at the pixel level to achieve moreeffective results, so that later became the regional geographic information to get a full,prompt and accurate understanding. Pixel-level image information is often the basis forthe latter characteristic level, at a strategic level analysis, For this purpose, we propose afusion algorithm based on the evidence of improved theoretical data on the basis ofimage morphology, the pixel gray values contained redundancy, the uncertainty of thedata, analysis, deduction, and ultimately to the intuitive description of the regionaltargets.The contents which in the Paper as follows:Combined with a more mature grayscale morphological operations, a new imagepre-processing technology, while regional image denoising and simplify the imagedetails, the way in high-, low-resolution grayscale different emphasis on the value of thedegree of the corresponding information data, and then analyze the uncertainty andredundancy in data characteristics, the pixel gray value description contains a newtheory of evidence for the morphological preprocessing for high, the low degree of pixelgray fuzziness and uncertainty of information caused by the characteristics of solutions,can be more effective to eliminate irrelevant details in the original image, and getvaluable information in the original image. The above results by the open sourceOpenCV library of visual programming experiment confirmed this method to meet therequirements of the research purposes.