A workflow approach for mosaicking UAV images

Abstract

In recent years, the Unmanned Aerial Vehicles (UAVs) technologies are widely employed in many fields such as disaster monitoring, map revision, and aerial imagery. A UAV aerial remote sensing system has several advantages such as low cost, high spatial resolution, and flexibility. However, one single UAV image can only cover a small area due to the limited altitude of the vehicle and the restricted view of the UAV camera. In order to get the overview image of a target area, mosaic of dozens of images is necessary. In the mosaic process, the Speeded Up Robust Features (SURF) algorithm performs well and robustly in removing distortion. It works as the first step and can extract an enormous amount of feature points from images effectively and quickly, which later can be used in image registration. Based on the correct registration, image blending is applied to get the ultimate stitching image that covers the whole workspace. This paper presents a workflow approach on automatic mosaic of UAV images. Workflow technologies are widely used in the information domain for automation of business processes. By using the workflow approach, users could customize the whole process and plug in new subprocess when needed. The whole process could be archived and invoke on-demand for automatic UAV image mosaic.