1. Space Security Research Center, Space Engineering University, Beijing 101416, China 2. 91039 Troops, Beijing 102400, China 3. School of Geospatial Information, Information Engineering University, Zhengzhou 450001, China
Aimed at tackling the problems that there are many mismatched points in the bundle adjustment of unmanned aerial vehicle (UAV) images, the variance loss function may cause the distortion of the solution parameters, and there exists large deviation from the true value which even causes impossibility of converge, the authors applied a robust bundle adjustment method considering the reliability of the observation value. This method uses loss function as a strategy to suppress gross errors, and is a variant designed on the basis of Cauchy loss function. The main idea of this method is adjusting the total loss function adaptively according to the mean value and variance of overlap degree and the residual of feature points, so as to overcome the influence of mismatched points on the computation of image parameters. Correspondingly, a practical accuracy evaluation method independent of ground control point (GCP) was designed. Experiments show that the method can still get robust adjustment results with high mismatch rate, and hence it is practical.
Colomina I, Molina P. Unmanned aerial systems for photogrammetry and remote sensing:A review[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2014,92:79-97.
Lei Y Z, Li Z W, Zhong K, et al. Mismatching marked points correction method based on random sample consensus algorithm[J]. Acta Optica Sinca, 2013,33(3):1-8.
Shu H, Shi W Z. Extension of reliability theory of surveying adjustment into spatial data analytics[J]. Geomatics and Information Science of Wuhan University, 2018,43(12):1989-1993.
[7]
张保明, 龚志辉, 郭海涛 . 摄影测量学[M]. 北京: 测绘出版社, 2008.
Zhang B M, Gong Z H, Guo H T. Photogrammetry[M]. Beijing:Surveying and Mapping Compilation, 2008.
[8]
李德仁, 袁修孝. 误差处理与可靠性理论[M]. 武汉: 武汉大学出版社, 2012.
Li D R, Yuan X X. Error processing and reliability theory[M]. Wuhan: Wuhan University Press, 2012.
[9]
Aravkin M, Styer Z, Moratto A, et al. Student’s trobust bundle adjustment algorithm [C]// International Conference on Image Processing, 2012: 1757-1760.
[10]
吴福朝 . 计算机视觉中的数学方法[M]. 北京: 科学出版社, 2008.
Wu F C. Mathematic method in computer vision[M]. Beijing:Science Preess, 2008.
[11]
Madsen K, Nielsen H B, Tingleff O. Methods for Non-Linear Least Squares Problems[M]. Copenhagen:[s.n.], 2004.
[12]
Moulon P, Monasse P, Marlet R. Adaptive structure from motion with a contrario model estimation [C]//Asian Conference on Computer Vision.Springer, 2012: 257-270.
[13]
薛武. 无人机影像定位优化技术研究[D]. 郑州:信息工程大学, 2017.
Xue W. Research on optimization technology of UAV images positioning[D]. Zhengzhou:Information Engineering University, 2017.
[14]
Talker L, Moses Y, Shimshoni I. Using spatial order to boost the elimination of incorrect feature matches [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.IEEE, 2016: 1809-1817.
[15]
Masiero A, Guarnieri A, Vettore A, et al. On the use of INS to improve feature matching[J]. The International Archives of Photogrammetry,Remote Sensing and Spatial Information Sciences, 2014,40(1):227.
[16]
Tsai C H, Lin Y C. An accelerated image matching technique for UAV orthoimage registration[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2017,128:130-145.
Yuan X X, Chen S Y, Zhong C. Oblique aerial image relative orientation based on fundamental matrix[J]. Geomatics and Information Science of Wuhan University, 2016,41(8):995-1000.
[20]
耿则勋, 张保明, 范大昭 . 数字摄影测量学[M]. 北京: 测绘出版社, 2010.
Geng Z X, Zhang B M, Fan D Z. Digital photogrammetry[M]. Beijing:Surveying and Mapping Compilation, 2010.