航空遥感影像匀光匀色方案研究与应用

发布时间：2018-01-31 08:03

本文关键词： 影像质量评价重采样单影像质量多影像一致性镶嵌　出处：《东华理工大学》2013年硕士论文　论文类型：学位论文

【摘要】：由于设备、天气和地形条件差异导致拍摄的航空遥感影像本身存在雾霭、偏色、阴影以及多影像间色调不一等问题，这些影像的质量问题是影响正射纠正成图的整体视觉效果和判读效果的直接因素，并且足以扰乱许多图像处理任务的执行，已经严重超出了传统意义上图像噪声对图像品质造成的影响。只有通过匀光匀色过程，在逐个去除这些影响因素的前提下，才能顺利执行后续的数字图像的处理过程，才能保证最后“一张图”成图的效果。虽然现在处理影像的质量问题是研究的热点，并且存在一些处理单个问题的算法，如去阴影算法、去雾算法等，但还没有一个完善的方案能比较全面的解决航空遥感影像在匀光匀色过程中的各种影响因素。本文研究并提出了一种航空遥感影像的匀光匀色方案，该方案从DOM处理流程出发，，在逐个解决问题的前提下，较为全面地处理影像的主要质量问题。基本过程为：首先根据原始影像的地面分辨率按照国家成图比例尺对应地面分辨率通过适当的采样方式进行重采样；接着在重采样的基础上进行单影像质量问题的处理，其中包括雾霭处理、阴影处理和偏色处理；然后在此基础上进行多影像色调和亮度的一致性调整；最后在镶嵌接边后的成图上通过调整镶嵌线走势和羽化值来消除接边线痕，进而完成“一张图”影像。此外，文中引用了一套（15张）一个测区范围内的航空遥感影像实际数据作为实验示例，将该匀光匀色方案与传统的、现有的匀光匀色软件的处理结果进行对比，并通过实验成果影像及其影像质量评价信息都证明，本文提出的航空遥感影像新匀光匀色方案大大优于现有的和传统的方案，同时弥补了单个匀色算法的不足，亦通过结合，能够形成面向海量影像成果匀光匀色制作的解决方案，能够满足了国家正射影像图成果的要求。
[Abstract]:Because of the difference of equipment, weather and terrain conditions, the aerial remote sensing images themselves have the problems of haze, color deviation, shadow and different colors among the multiple images. The quality of these images is a direct factor affecting the overall visual and interpretation effects of orthophoto correction, and can disrupt the execution of many image processing tasks. It has gone beyond the traditional image noise impact on image quality. Only through the uniform color process, in the premise of removing these factors one by one. In order to smoothly carry out the subsequent digital image processing process, can ensure the final "one map" mapping effect, although the quality of the image processing is now the focus of research. And there are some algorithms to deal with a single problem, such as shadow removal algorithm, de-fog algorithm and so on. However, there is no perfect scheme to solve the influence factors of aerial remote sensing image in the process of uniform light color. In this paper, a uniform color scheme for aerial remote sensing images is proposed. The scheme is based on the DOM process and solves the problems one by one. The basic process is as follows: firstly, according to the ground resolution of the original image, the ground resolution of the original image is resampled by appropriate sampling method according to the national mapping scale corresponding to the ground resolution; Secondly, on the basis of resampling, the single image quality problem is processed, including haze processing, shadow processing and color deviation processing. Then, the consistency of color and brightness of multi-image is adjusted on this basis. Finally, by adjusting the trend of mosaic line and the value of Eclosion, we can eliminate the line mark and then complete the "one picture" image. In this paper, a set of 15 pieces of aerial remote sensing image data in a measuring area is cited as an experimental example, and the results of the scheme are compared with that of the traditional uniform color software. And through the experimental results of images and image quality evaluation information, it is proved that the proposed new uniform color scheme of aerial remote sensing images is much better than the existing and traditional schemes, and makes up for the shortcomings of a single uniform color algorithm. Through combination, the solution of uniform color production for mass image results can be formed, which can meet the requirements of national orthophoto image results.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P237

【参考文献】