The influence of active projection speckle patterns on underwater binocular stereo vision 3D imaging

Due to the phenomena of refraction, absorption and scattering, underwater stereo vision faces more difficulties than in air. These phenomena will bring various problems such as blurred details, reduced contrast and brightness to underwater images, which will affect the accuracy and reliability of stereo matching and increase the difficulty of stereo matching. The method based on active speckle pattern is a more simple and reliable method to enhance the surface texture of underwater objects. A good-quality speckle pattern can provide more robust and accurate results for matching the corresponding points of the left and right images. This is a more important research problem in water than in air. In this paper, the row entropy of speckle pattern is adopted to quantify the quality of the projection speckle pattern suitable for underwater stereo vision, and the relevant experiment and analysis are carried out to prove the correctness and effectiveness of new parameters. Further, the influence of different projection speckle patterns on underwater 3D imaging is researched. The actual underwater experiment indicates that the speckle pattern exerts a great influence on the accuracy of underwater 3D imaging based on the binocular stereo vision, and the row entropy of higher-quality projection speckle pattern should be as small as possible. This work can give a reference for the selection of projection speckle patterns in underwater binocular stereo vision. An appropriate speckle pattern can effectively improve the accuracy of underwater 3D measurement.

Automated summary

Due to refraction, absorption, and scattering, underwater stereo vision encounters more challenges than in air. These phenomena result in blurred details, reduced contrast and brightness in underwater images, which affect the accuracy and reliability of stereo matching and increase the difficulty. The method based on active speckle pattern enhances the surface texture of underwater objects, providing robust and accurate results for matching corresponding points. This research problem is more important in water. The experiment demonstrates that the speckle pattern significantly affects the accuracy of underwater 3D imaging and higher-quality patterns with lower row entropy should be preferred. This work can guide the selection of projection speckle patterns in underwater stereo vision and improve the accuracy of underwater 3D measurement.