Academy of Interdisciplinary Studies Division of Integrative Systems and Design 217 ChatGPT for CAD and Graphics Design Supervisor: YEUNG Sai Kit / ISD Student: ZHOU Hangan / CPEG Course: UROP 1000, Summer Our world is stepping into the AI era. These assistants could deal with natural language and execute orders. To utilize it, our CADgpt program aims to develop a blender add-on that uses GPT-4 to enable 3D modeling and optimize the design workflow. Users could generate and run blender Python codes by typing commands in the textbox. This addon also integrates with the shape-e library, which enables it to use the more mature models made by previous researchers. This report focuses on the testing of its abilities in performing operations, generating models, and exploring the limits. In addition, examine the datasets in shape-e to compare their similarity with real objects. Finally, to get a systematic evaluation. Image Processing Method for Underwater SLAM Supervisor: YEUNG Sai Kit / ISD Student: FANG Ruoyun / ISD Course: UROP 1000, Summer This report explores the intricacies of managing data in underwater settings and implementing image processing methodologies for Underwater SLAM utilizing data gathered underwater. Underwater SLAM plays a pivotal role in facilitating autonomous navigation and mapping in demanding aquatic surroundings. However, in contrast to conventional SLAM, Underwater SLAM encounters challenges stemming from factors like diminished visibility and intricate lighting scenarios. Consequently, advanced image processing methods are essential to augment SLAM efficacy and harness the outcomes for real-world applications. The present emphasis lies in scrutinizing image data collection, diverse image processing techniques, and delving into prospective research avenues in this domain. Underwater Robot for Biodiversity Study Supervisor: YEUNG Sai Kit / ISD Student: YIP Chi Ho / COMP Course: UROP 1000, Summer The protection of marine biodiversity and the conservation of ecosystem is essential to ensuring the existence of ocean species. Constant monitoring of species by means of an underwater robot such as an ROV (Remotely Operated Vehicle) can reduce the financial and labor costs of underwater inspection. Underwater simulation plays a significant role in testing the ROV functions before deployment in water. Orca4, maintained by Clyde McQueen, is used as the basis for the research. Enhancements to the simulation’s authenticity were made based on Orca4. Future work will focus on enhancing the visual aspect to prepare ROVs for real-world underwater tasks.
RkJQdWJsaXNoZXIy NDk5Njg=