›› 2010, Vol. 12 ›› Issue (2): 254-260.

• ARTICLES • Previous Articles     Next Articles

Fast Image Processing Method of UAV without Control Data

GONG Adu1,3, HE Xiaoying2, LEI Tianjie2, LI Jing1,3   

  1. 1. Academy of Disaster Reduction and Emergency Management,Beijing Normal University,Ministry of Civil Affairs/Ministry of Education of China,Beijing 100875,China;
    2. College of Resources Science and Technology,Beijing Normal University,Beijing 100875,China;
    3. Key Laboratory of Environmental Change and Natural Disaster,Beijing Normal University,Ministry of Education,Beijing 100875,China
  • Received:2009-10-30 Revised:2010-03-15 Online:2010-04-25 Published:2010-04-25

Abstract: Unmanned Aerial Vehicles(UAV) images quickly processing method without other GCP(Ground Control Point) data is discussed in this paper,and the UAV images of Disaster Areas of Wenchuan Earthquake in Sichuan Province are used as the typical test data source.In the handling progress,only the images and auxiliary data recorded by the UAV system itself are used to stitch and rectify the image mosaics.The main work contains images,which were recorded by digital camera on the UAV and auxiliary data,which were recorded by GPS(Global Positioning Satellite) system on the UAV analyzing,flying area blocking,image auto-stitching after blocking,image rectifying and image mosaic.The image auto-stitching is the key point of the whole research.Firstly,a detailed analysis on UAV images and auxiliary data is done.With the analysis result,many questions are put out,such as the number of images is so large and UAV image distortion is worse than that of traditional photogrammetry.These bring a lot of difficulties to the work,that the normal methods can not be used.Base on this situation,a new strategy is proposed in this paper.That is,in the small area,which is determined by the experiment,the auto-stitching method base on image matching is raised,then the regional images after auto-stitched are corrected according to auxiliary data of UAV.In the image matching progress,SIFT(Scale-Invariant Features Transform) algorithm is applied in order to achieve high efficiency and high precision.Then pseudo center points collected from auxiliary data are used to rectify the regional images.From the result image after stitching and rectifying,a conclusion can be drawn that the relative accuracy is high and the mosaic image is visually dislocation-free.

Key words: Unmanned Aerial Vehicles, image, GCP, quickly stitch, rectification