Real-time image processing on handheld devices and UAV

Detta är en Kandidat-uppsats från Linköpings universitet/Institutionen för datavetenskap

Sammanfattning: The forest industry needs an up-to-date overview of certain areas of a forest, to either estimate damages after a storm or assess its overall health. Today, the use of unmanned aerial vehicles (UAV) have exploded. Almost anyone can own one and they are very easy to operate. They are often equipped with accurate sensors and cameras that can be used for several productive applications. This paper investigates if a UAV equipped with positional sensors and a high resolution camera can be used for taking aerial photographs and together with a mobile devices create a coherent orthophoto in real- or near real-time. Three different seam stitching algorithms are tested and evaluated based on speed, accuracy and visual appearance. The results of using a DJI Matrice 100 UAV together with an iPad Air is presented both qualitative and quantitative.

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)