研究目的
To explore the use of structured light to achieve high-resolution 3D imaging of objects in an underwater environment, adapting a Gray Code Phase Stepping approach to be robust to contrast degradation caused by light scattering and attenuation in water.
研究成果
The adapted structured light approach using inverse Gray codes significantly improves segmentation robustness in turbid water, reducing mis-segmented pixels and enabling high-resolution 3D imaging with depth precision from 1.4mm to 6.4mm. This facilitates potential automation of underwater tasks such as inspection and manipulation, though performance is affected by water turbidity.
研究不足
The sensor's range is limited by the attenuation length of water, and depth precision decreases with increasing turbidity due to reduced signal-to-noise ratio. The current setup uses a USB 3.0 camera with a 10m cable limit, which could be extended with different hardware. The housing size could be reduced with a fixed baseline.
1：Experimental Design and Method Selection:
The study uses a structured light methodology combining Gray codes and phase stepping for depth estimation. It involves projecting binary patterns (Gray codes) and phase-shifted sinusoidal patterns to enhance resolution. The approach is tailored with inverse Gray codes to improve robustness in turbid water.
2：Sample Selection and Data Sources:
Experiments were conducted in a controlled aquarium and an 8m x 4m pool. An object (a flat plate with small structures) was placed at 1.1m distance for turbidity tests, and a 3D printed fish-tail handle was imaged in the pool.
3：1m distance for turbidity tests, and a 3D printed fish-tail handle was imaged in the pool.
List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Includes a DMD projector (DLP LightCrafter 4500), a machine vision camera (IDS UI-3160CP), an underwater housing with windows, a turbidity measurement device, and materials like brown clay for adjusting water turbidity.
4：Experimental Procedures and Operational Workflow:
The sensor projects patterns onto the scene, captures images with the camera, and processes them using algorithms for phase shift and Gray code segmentation. Turbidity was varied by adding clay to water, and repeated measurements were taken to evaluate segmentation and depth precision.
5：Data Analysis Methods:
Depth precision was estimated as the standard deviation from repeated measurements. Segmentation errors were counted by comparing to a reference in non-turbid water. Triangulation routines were used to compute depth from phase estimates.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
