Desenvolvimento de um método que aplica realidade aumentada em imagens aéreas

Abstract

This dissertation presents the method RAIA - Augmented Reality in Aerial Images - which was developed to insert available information on an aerial image. The goal is to facilitate the activities involving the analysis of the images captured by cameras fixed on aircraft. Also, there are currently selling products that offer the functionality to complement aerial images with information, although there are few academic records on this type of application and no records were found that detail methodologies that complement with Augmented Reality images acquired from cameras fixed on aircraft. In the images processed by the method is employed the concept of Augmented Reality, because it is displayed a junction of virtual objects, that represent information about the site being captured, and the image itself. This is done by calculating the image georeferencing, which means that the geographic coordinates are determined for each pixel. The method was evaluated in tests on a prototype tool. These tests were divided into three case studies that were used to measure the accuracy that the method overrides information for images captured at distances close to the ground, captured from satellite cameras and in a real aerial surveillance situation. The positioning error results presented by RAIA, in a real aerial surveillance situation, were about approximately 5 meters for images captured with little inclination. Images captured with inclination greater than 40° showed positioning errors of approximately 40 meters.