The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Download
Citation
Articles | Volume XXXVIII-1/C22
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XXXVIII-1/C22, 271–276, 2011
https://doi.org/10.5194/isprsarchives-XXXVIII-1-C22-271-2011
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XXXVIII-1/C22, 271–276, 2011
https://doi.org/10.5194/isprsarchives-XXXVIII-1-C22-271-2011

  06 Sep 2012

06 Sep 2012

A NEW APPROACH TO FAST MOSAIC UAV IMAGES

Q. Liu1, W. Liu2, L. Zou1, J. Wang2, and Y. Liu1 Q. Liu et al.
  • 1Department of Geomatic Engineering, Central South University, Changsha, Hunan, China
  • 2School of Surveying and Spatial Information Systems, University of New South Wales, Sydney, Australia

Keywords: Mosaic UAV Images, Coordinate Transformations, Displacement of Images, Image Mosaic Errors

Abstract. Unmanned Aerial Vehicles (UAVs) have been widely used to acquire high quality terrain images of the areas of interest, particularly when such a task could potentially risk human life or even impossible as the areas cannot be accessed easily by surveyors. Once the images have been obtained, traditional photogrammetric processing process can be used to establish a relative orientation model and then, absolute orientation model with the procedures of space resection and intersection. In many such applications, the geo- referenced images which are stitched together to represent the geospatial relationships for the feature objects are sufficient. A fast or near real-time processing approach for UAV images using GPS/INS data has being investigated for years. One beneficial application of such approach is the capability of quick production of geo-referenced images for various engineering or business activities, such as urban and road planning, the site selection of factories and bridges, etc. In this paper, we have proposed a new fast processing approach for the UAV images collected with an integrated GPS/INS/Vision system. The approach features that the corresponding points between images have been determined, and then coordinate transformation is carried out to implement image stitching. The accuracy of corresponding points normally affects the quality of stitched images, but the results of our experiments revealed that the image stitching errors were obvious even the accuracy of corresponding points was high. The stitching errors could be caused by the changes of surface elevation.