Underwater scene depth evaluation method and system based on monocular vision

Abstract

The invention discloses an underwater scene depth evaluation method and system based on monocular vision. The method comprises the following steps: acquiring scene image data shot and collected by a monocular camera at different visual angles in an underwater dynamic scene; performing feature extraction on video frames in the scene image data, and extracting feature points in the scene image data of different visual angles; performing depth estimation on the feature extraction result of the shot object based on a depth estimation model according to the change of the visual angle position of the camera to obtain a visual depth value of the shot object; performing depth processing on the scene image data based on different view angle pose parameters of the camera to optimize object surface vision depth values of feature points in the scene image data to obtain a scene depth estimation image; the system comprises an acquisition unit, a feature extraction unit, a depth judgment unit, a depth evaluation unit and a registration unit. According to the invention, underwater scene depth evaluation can be completed without considering influence factors of a camera baseline on system precision.

Description

technical field [0001] The invention relates to computer vision technology, in particular to a method and system for evaluating depth of underwater scene based on monocular vision. Background technique [0002] With the continuous exploration of the ocean and other waters, underwater operations, marine scientific investigations and other activities are also increasing. In order to improve the efficiency of underwater operations and achieve the purpose of marine scientific investigations, underwater operation equipment is usually used to assist operations. Due to the existence of various complex underwater scene environments, for the needs of task execution, to improve people's perception of underwater scenes, it is necessary to use technologies such as computer vision and image processing to evaluate the position of underwater operation equipment in unknown environments. The three-dimensional structure of the scene is reconstructed, and the observation of the underwater scen...

AI Technical Summary

Problems solved by technology

At present, in order to realize the reconstruction of underwater 3D scenes, SLAM methods are usually based on depth cameras or binocular cameras to achieve dense reconstruction of 3D scenes. However, the shortcomings of this type of method are obvious, and they rely too much on hardware devices, resulting in limited practical application scenarios.

In particular, the binocular system using a binocular camera needs to use a stereo camera to collect two images at the same time at each moment, which leads to high cost, and also needs to consider the impact of the camera baseline on the system accuracy, which is not conducive to the depth of the underwater scene. Evaluate

[0004] The SLAM method based on a depth camera or a binocular camera can achieve dense reconstruction of a 3D scene, but the shortcomings of this type of method are obvious, and the excessive reliance on hardware devices leads to limited practical application scenarios.

In particular, depth cameras have a limited working range and are sensitive to light, so they are only suitable for close-range indoor environments

Images