Please use this identifier to cite or link to this item:
|Title:||Visual analytics for aircraft identification||Authors:||Yeo, Tiffany Yu Ling||Keywords:||DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision||Issue Date:||2019||Abstract:||Computer vision has been used to tackle various problems in object recognition and image classification. In the air traffic control scene, the identification of aircraft is a process that can be automated with the advancement in computer vision solutions. This report aims to study and evaluate the effectiveness of 2 different image classification methods, SIFT and CNN, for identifying aircraft based on images and to subsequently propose a feasible implementation. Performance is primarily evaluated using the metric of accuracy, but other factors like ease of computation and implementation will be considered as well. For this project, we conducted a test on the feasibility of using SIFT as a feature extractor. However, the tests show that it is not able to draw accurate keypoints apart from the airline’s paintjob and hence is not a viable solution. We then develop and test implementations of MobileNet and find that it achieves impressive accuracies of 70.3%, 64.8%, and 47.7% for identification of manufacturer, family and variant respectively.||URI:||http://hdl.handle.net/10356/77867||Rights:||Nanyang Technological University||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||MAE Student Reports (FYP/IA/PA/PI)|
Files in This Item:
|Visual Analytics for Aircraft Identification FYP.pdf|
|3.12 MB||Adobe PDF||View/Open|
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.