Development of Brightness Correction Method for Mosaicking UAV Images

초록

Remote Sensing using unmanned aerial vehicles (UAV) can acquire images with higher time resolution and spatial resolution than aerial and satellite remote sensing. However, UAV images are photographed at low altitude and the area covered by one image is relatively narrow. Therefore multiple images must be processed to monitor large area. Since UAV images are photographed under different exposure conditions, there is difference in brightness values between adjacent images. When images are mosaicked, unnatural seamlines are generated because of the brightness difference. Therefore, in order to generate seamless mosaic image, a radiometric processing for correcting difference in brightness value between images is essential. This paper proposes a relative radiometric calibration and image blending technique. In order to analyze performance of the proposed method, mosaic images of UAV images in agricultural and mountainous areas were generated. As a result, mosaic images with mean brightness difference of 5 and root mean square difference of 7 were avchieved.

키워드