International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Download
Citation
Volume XLII-3/W10
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-3/W10, 595–599, 2020
https://doi.org/10.5194/isprs-archives-XLII-3-W10-595-2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-3/W10, 595–599, 2020
https://doi.org/10.5194/isprs-archives-XLII-3-W10-595-2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

  07 Feb 2020

07 Feb 2020

RESEARCH ON ACCURACY EVALUATION METHOD BASED ON THE RESULT OF IMAGE CORRECTION AT 10000-SCENES CLASS

M. Li, H. P. Chen, B. Qiu, W. J. Xie, and Y. H. Chen M. Li et al.
  • National Quality Inspection and Testing Center for Surveying and Mapping Products, China

Keywords: Orthorectification, Quality control, Accuracy, 10000-Scenes class

Abstract. In the project of National Fundamental Geographic Information Database Updating, the high-qualified orthorectification to a bulk of remote sensing images are the foundations that guarantee the reliable geographic information for national economic construction and social developments. Therefore, quality control for orthorectification, as a crucial step during remote sensing, is of great significance. Basically, the image orthorectification have been achieved by automatic matching of millions of frame-referenced images. In this paper, a make-to-measure method is devised with improvement in the way of sampling and plane precision evaluation. The improved method is verified through nine 1:1million scenes across the nation. Stratified sampling is deployed according to the topographical features and locations geo-information is collected referring to the digital orthorectification model (DOM) results of geographical conditions Monitoring in 2017. The results show that the root mean square errors (RMSEs) calculated by improved methods described in this paper are highly consistent with the RMSEs that are calculated by automatic image matching, which means the improved method can adequately evaluate the nationwide accuracy of image correction and enable to provide reference and instruction for the large-scale, 10000-scenes class quality assurance of the image correction.