Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


39310.pdf (33.33 MB)

ETD Abstract Container

Abstract Header

Quality Analysis of UAV based 3D Reconstruction and its Applications in Path Planning

Rathore, Aishvarya
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1627658323958222

Abstract Details

2021, MS, University of Cincinnati, Engineering and Applied Science: Computer Science.
Unmanned Ariel Vehicles (UAV) have become increasingly popular and have emerged as a powerful tool for many applications, including commercial, military, civilian, and emergency relief. The ease of availability, the ability to carry high-resolution cameras, and the small size have made UAVs a popular choice for ariel photography required for 3D reconstruction. 3D reconstruction is helpful in surveillance, mapping street view, inspection, disaster area evaluation, and emergency relief. Compared to the traditional laser-based approach, the use of UAVs for generating the 3D model is cost-effective and provides access to areas beyond human reach. However, due to the complex process of flying the UAV in an optimal path, the energy budget of the UAV, and difficulty in efficiently ensuring image acquisition, the quality of UAV-based 3D reconstruction and efficient path planning remain challenging problems. In addition, evaluating the accuracy of a 3D model can be time-consuming as it requires the 3D model to be fully reconstructed before accuracy analysis can be performed. This thesis comprises two interconnected parts of work aiming to provide a complete solution for quality analysis of UAV-based 3D reconstruction and its applications in path planning. In the first part of the work, a regression-based accuracy model is proposed to predict the accuracy of a UAV-based 3D model using the image data based on the overlap ratio between the images, the number of images, the sensing direction, distance from the object, and the spatial index of the scene. For this model, a method to compute the overlap ratio, considering the spherical distances (great circle distances) between every two consecutive camera locations in a dataset, is presented. Evaluation results show that the overlap ratio calculation follows the expected trend. The accuracy is predicted correctly, with an RMS (Root Mean Square) of 0.88 and RSME (Root Mean Square Error) of 0.0045. For the second part, using the accuracy model proposed, an accuracy-based energy-aware path planning algorithm is introduced. In this algorithm, using the user-provided accuracy requirement, energy budget, and spatial resolution, a set of waypoints for an optimal circular path are computed while minimizing the energy requirement and ensuring the user requirement of spatial resolution, accuracy, and coverage are met. The algorithm also aims to reduce the total number of pictures to reduce the computational time complexity of the 3D reconstruction. Evaluation results show that the energy is minimized. Furthermore, the computational time is reduced significantly, up to 70%, while ensuring the user's accuracy, spatial resolution, and coverage requirements are met.
Rui Dai, Ph.D. (Committee Chair)
Seokki Lee, Ph.D. (Committee Member)
Nan Niu, Ph.D. (Committee Member)
71 p.
Computer Science
UAV; 3D Reconstruction; Drone; Quality Analysis; Path Planning; Algorithm

Recommended Citations

Citations

  • Rathore, A. (2021). Quality Analysis of UAV based 3D Reconstruction and its Applications in Path Planning [Master's thesis, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1627658323958222

    APA Style (7th edition)

  • Rathore, Aishvarya. Quality Analysis of UAV based 3D Reconstruction and its Applications in Path Planning. 2021. University of Cincinnati, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1627658323958222.

    MLA Style (8th edition)

  • Rathore, Aishvarya. "Quality Analysis of UAV based 3D Reconstruction and its Applications in Path Planning." Master's thesis, University of Cincinnati, 2021. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1627658323958222

    Chicago Manual of Style (17th edition)

ucin1627658323958222
41
© 2021, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.