Date of Award
12-2023
Degree Type
Thesis-Restricted
Degree Name
M.S.
Degree Program
Computer Science
Department
Computer Science
Major Professor
Atriya Sen
Abstract
We are more interested in recovering and salvaging containers which may be loaded with either valuable or harmful substances. Therefore, we designed various container models to account for different loading conditions and then exported them to a 3D printer. Six small container models were dropped into the trailer pool at the University of New Orleans (UNO) as planned at angles of 0°, 45°, and 90°.
After collecting all the videos using a high-definition camera set outside the tank, we perform pre-processing tasks on the videos in preparation for model training. The two-phase angle classification method uses a pretrained ResNet50 model trained on ImageNet as a feature extractor to generate latent features to train a gated recurrent unit (GRU) model to classify the entry angle of dropped objects.
Furthermore, the output of the recognition phase, including the shape of the dropped object and the initial drop angle, is used as input to another in-house tool, the Dropped Object Simulator (DROBS), which was developed in MATLAB using simulate the trajectory and determine the final landing position.
Recommended Citation
Yu, Xiaochuan, "Dropped Objects Recognition in Offshore Operations Based on Computer Vision and Artificial Intelligence" (2023). University of New Orleans Theses and Dissertations. 3116.
https://scholarworks.uno.edu/td/3116
Rights
The University of New Orleans and its agents retain the non-exclusive license to archive and make accessible this dissertation or thesis in whole or in part in all forms of media, now or hereafter known. The author retains all other ownership rights to the copyright of the thesis or dissertation.