Underwater Object Recognition Using Point-Features, Bayesian Estimation and Semantic Information
Khadidja Himri1, Pere Ridao1, Nuno Gracias1
1Underwater Robotics Research Center (CIRS), Computer Vision and Robotics Institute (VICOROB), University of Girona, Parc Científic i Tecnològic UdG C/Pic de Peguera 13, 17003 Girona, Spain.
Sensors (Basel, Switzerland)
|April 3, 2021
Summary
This study introduces a 3D object recognition method for sub-sea structures using point features. Incorporating Bayesian estimation and semantic information significantly boosts recognition accuracy for industrial inspection, maintenance, and repair (IMR).
Area of Science:
- Robotics and Computer Vision
- 3D Point Cloud Processing
- Industrial Inspection Technologies
Background:
- Automated inspection of sub-sea industrial structures requires robust 3D object recognition.
- Existing methods struggle with non-coloured point cloud data common in underwater environments.
- Inspection, Maintenance, and Repair (IMR) operations demand precise identification of components like pipes, valves, and connectors.
Purpose of the Study:
- To develop and evaluate a 3D object recognition method for non-coloured point clouds.
- To enhance the accuracy and reliability of object recognition for sub-sea industrial applications.
- To investigate the impact of feature descriptors, Bayesian estimation, and semantic information on recognition performance.
Main Methods:
- A 5-stage recognition pipeline including plane segmentation, pipe detection, semantic object segmentation, feature-based recognition, and Bayesian estimation.
- Development of an Interdistance Joint Compatibility Branch and Bound (IJCBB) algorithm for object tracking within Bayesian estimation.
- Utilizing the Clustered Viewpoint Feature Histogram (CVFH) descriptor for feature extraction.
Main Results:
- The Clustered Viewpoint Feature Histogram (CVFH) descriptor demonstrated superior performance.
- Bayesian estimation improved recognition rates by 18%.
- Inclusion of semantic information further increased recognition rates by an additional 21%, reaching up to 90%.
Conclusions:
- The proposed 3D object recognition method effectively handles non-coloured point clouds for sub-sea industrial structures.
- Bayesian estimation and semantic information are crucial for achieving high recognition accuracy in challenging IMR scenarios.
- The CVFH descriptor combined with the proposed pipeline offers a promising solution for automated sub-sea inspection.
Related Concept Videos
01:11Buoyancy and Stability for Submerged and Floating Bodies
2.3K
In fluid mechanics, buoyancy and stability are key concepts for understanding the behavior of submerged and floating bodies. When a stationary body is fully or partially submerged in a fluid, the fluid exerts a force on the body known as the buoyant force. This force acts vertically upward through a point called the center of buoyancy, which is the center of the displaced fluid volume. According to Archimedes' principle, the magnitude of the buoyant force is equal to the weight of the fluid...
2.3K
00:59Buoyancy
11.5K
When an object is placed in a fluid, it either floats or sinks. All objects in a fluid experience a buoyant force. For example, a metal ball sinks, while a rubber ball floats. Similarly, a submarine can sink and float by adjusting its buoyancy. The concept of buoyancy raises several interesting questions. For instance, where does this buoyant force come from? How much buoyant force is required to make an object sink or float? Do objects that sink get any support at all from the...
11.5K