DWE Marine Stabilization System
University of California, San Diego
Mechanical and Aerospace Engineering
MAE 156B: Senior Design Project (Spring 2024)
Background
DeepWater Exploration (DWE) designs camera hardware and complementary machine learning algorithms to improve the quality of their hardware in the marine environment. These cameras are used by customers to obtain footage from submersible remotely operated vehicles (ROVs) and autonomous surface vehicles (ASVs). However, the footage from these cameras can be easily disturbed due to their high zoom level. DWE’s cameras use a global shutter as opposed to a rolling shutter, which helps to somewhat counteract the problem, but the cameras are still vulnerable to slight disturbances. Thus, DWE needs a camera stabilization mechanism to eliminate these perturbations from the ocean environment.
Objective
The objective of this project is to develop a mechanism capable of stabilizing the footage from DWE cameras with the capability for DWE to implement their own control algorithms within this mechanism. This mechanism should be able to support a 12cm x 12cm x 12cm, 1.5kg payload (which encompasses several different camera configurations) and all relevant components should be water and corrosion resistant. This mechanism should also provide DWE with the capability to implement their own object tracking software for control purposes. This mechanism should also improve the image quality by 0.2 points as calculated by the image quality evaluation software when subject to wave conditions ranging from 0-0.61m with a frequencies from 0.07-5Hz.
Final Design
To achieve camera footage stabilization in the marine environment a 3-axis gimbal with an additional z-axis is utilized. This system aims to isolate the camera/s from high frequency small amplitude vertical ocean wave displacement as well as compensate for any roll, pitch, and yaw angular displacements that the vessel experiences. The final design uses a UM7 IMU to obtain state data which feeds into an Arduino Due with custom developed control software that counteracts ocean perturbations.
Performance Results Summary
The camera mounted on the gimbal and z-axis mechanism performed better at longer distances as compared to the camera mounted on the stationary base plate. The image score was computed based on the average gradient of grayscale value.
Narrated Video
Final Presentation
Poster
Final Presentation 6_5.pdfDWE Poster.pdf