基于立体匹配片的国产卫星影像无缝立体测图方法研究

发布时间：2018-06-13 01:30

本文选题：国产卫星影像 + 立体匹配片　；参考：《中国地质大学(北京)》2015年硕士论文

【摘要】：随着现代科技的快速发展,遥感卫星在测绘领域中的应用越来越广。近年来,我国着重大力发展自己的测绘卫星系统,先后发射了资源系列和天绘系列等国产卫星,实现了我国国产测绘卫星的自主化。国产卫星系统的建立,需要有相应的立体测图方法对采集的卫星影像数据进行处理。针对国产三线阵卫星影像数据,本文提出了基于立体匹配片的无缝立体测图模型方法。首先利用已知的原始影像和RPC参数,采用基于投影基准面的近似核线影像生成方法进行核线重采样,然后通过SGM密集匹配方法,得到同名点的对应关系,内插生成数字高程模型,并对原始影像进行数字纠正生成正射影像,最后通过对数字高程模型人为的引入视差,对正射影像作重采样,生成该正射影像的立体匹配片。选取资源三号卫星同一条航带的三景影像数据,利用该数据处理流程生成立体正射影像像对,左影像为正射影像,右影像为立体匹配片。采用蚁群拼接优化算法对立体正射像对的左片和右片分别进行影像拼接,生成了正射影像和立体匹配片。对本文方法构建的无缝立体测图模型进行了可行性的验证,分别对两幅拼接后的大范围正射影像和立体匹配片的灰度进行线性拉伸,并用红、绿波段叠加显示,利用构建的测图模型和已有商业软件测图模型选取若干坐标点进行比较验证,得到本文提出的模型实现方法的测图精度符合精度标准,完全可以满足立体测图的要求。
[Abstract]:With the rapid development of modern science and technology, remote sensing satellites are more and more widely used in the field of surveying and mapping. In recent years, China has made great efforts to develop its own surveying and mapping satellite system, and has successively launched a series of domestic satellites, such as resource series and Tianyi series, so as to realize the autonomy of China's domestic surveying and mapping satellites. The establishment of domestic satellite system requires the corresponding stereo mapping method to process the collected satellite image data. In this paper, a method of seamless stereo mapping model based on stereo matching is proposed for the domestic three-line array satellite image data. Firstly, using the known original image and RPC parameters, the method of generating approximate kernel line image based on projection datum is used to resample the kernel line, and then the corresponding relation of the same name point is obtained by SGM dense matching method. The digital elevation model is generated by interpolation, and the original image is digitally corrected to generate orthophoto image. Finally, by introducing the parallax artificially to the digital elevation model, the orthophoto image is re-sampled to generate the stereo matching piece of the orthophoto image. In this paper, the three scene image data of the same belt of Resource-3 satellite are selected, and the stereo orthophoto image pair is generated by using the data processing process. The left image is a orthophoto image and the right image is a stereo matched image. The left and right images of stereo orthotropic image pairs were spliced by ant colony mosaic optimization algorithm, and orthophoto images and stereo matching pieces were generated. The feasibility of the seamless stereo mapping model constructed by this method is verified. The gray scales of two large range orthophoto images and stereo matching pieces are stretched linearly respectively and displayed in red and green bands. By comparing and verifying some coordinate points by using the constructed mapping model and commercial software mapping model, it is obtained that the mapping accuracy of the method proposed in this paper accords with the precision standard, and it can completely meet the requirements of stereoscopic mapping.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P237

【相似文献】