Low-Cost Multispectral Aerial Imaging using Autonomous Runway-Free Small Flying Wing Vehicles

摘要

Aerial imaging has become very important to areas like remote sensing and surveying. However, it has remained expensive and difficult to obtain with high temporal and spatial resolutions. This paper presents a method to retrieve georeferenced aerial images by using a small UAV (unmanned aerial vehicle). Obtaining aerial images this way is inexpensive, easy-to-use and allows for high temporal and spatial resolutions. New and difficult problems are introduced by the small image footprint and the inherent errors from an inexpensive compact inertial measurement unit (IMU). The small image footprint prevents us from using the features from the images to help negate the errors from the IMU, which is done in conventional methods. Our method includes: using the data from the IMU to georeference the images, projecting the images on the earth and using a man-in-the-loop approach to minimize the error from the IMU. Sample results from our working system are presented for illustration.