In this project, a new method is developed to optimize the performance of an Unmanned Aerial Vehicle (UAV) for autonomous detection and on-the-job view-planning of infrastructure elements with the purpose of their accurate three-dimensional (3D) modeling. The existing view-planning approaches in the literature have mostly modeled non-complex or small-scale objects and have rarely been adapted to flying robots. In addition, the target object is often identified by human operators. This research addresses these problems by training a drone to find the desired object of interest in an unknown environment during an inspection task without human interventions. To this end, first, a technique for object detection will be developed to recognize and locate the target object while the drone is exploring the environment. Second, based on the available information about the desired object, the drone will start next-best-view and motion planning to acquire an adequate photogrammetric network of images in order to reconstruct the inspection target in 3D both accurately and completely. This research will have important impacts on the evolution of infrastructure monitoring and assessment approaches using UAV systems.