Underwater scene reconstruction method based on motion recovery, and storage medium

Abstract

The invention discloses an underwater scene reconstruction method based on motion recovery, and a storage medium. The reconstruction method comprises: introducing an improved motion recovery algorithm, extracting a motion matrix, establishing a mutual connection between video images, and after completing redundant image culling, performing feature point matching and point cloud generation in two steps: firstly performing feature point matching on the binocular image, generating patches based on the matched feature points in order to obtain more dense point cloud data; then spreading the patches to all angles of view, to complete the reconstruction of the scene model, and finally performing the color correction of on point cloud model according to the imaging characteristics of the underwater scene. The invention may still achieve better reconstruction results when only a few input images are obtained, has relatively good efficiency and precision, and at the same time certainly improvesthe accuracy and robustness of the reconstruction scene.

Description

Technical field [0001] The invention relates to a three-dimensional reconstruction method, in particular to an underwater scene reconstruction method based on motion recovery and a storage medium, which can take into account both efficiency and accuracy. Background technique [0002] The real world is three-dimensional. In order to facilitate the observation and analysis of the real world and even expand it, it is necessary to reconstruct a three-dimensional model in a computer environment. In recent years, with the rapid advancement of computer hardware technology and the rapid development of software, there are more and more methods for constructing 3D models. Related software is widely used in medical image processing, 3D printing, computer games, virtual reality, mapping, simulation military training, Film and television entertainment and other fields. According to the different ways of obtaining reconstruction data, 3D model construction technology is mainly divided into th...

AI Technical Summary

Problems solved by technology

However, the existing algorithms also have many disadvantages: the binocular stereo vision algorithm based on feature point matching has high accuracy and low time complexity, but can only obtain the sparse point cloud of the scene, and the reconstruction effect is not ideal; Although the matching binocular stereo vision algorithm can obtain dense point cloud data, the accuracy is reduced, and the time complexity is greatly improved compared with feature point matching. When performing 3D reconstruction of large scenes, it takes too long

At the same time, because the traditional binocular stereo vision algorithm only considers the relationship between the corresponding frames, the reconstruction point cloud lacks the interconnection of different perspectives, and the reconstruction point cloud connection is unnatural, which affects the reconstruction effect.

[0009] During the reconstruction process of binocular stereo vision, similar or repeated parts of the image still have shortcomings such as mismatching caused by similar or repeated parts. Therefore, in response to these problems, multi-eye stereo vision reconstruction is proposed, and a new camera is further added on the basis of binocular to provide a more accurate image. More constraint information to improve the final reconstruction accuracy

Although multi-eye stereo vision can reduce mismatching and edge blurring to a certain extent during the reconstruction process, with the addition of additional cameras, the number of images that need to be processed in each viewing angle increases, and the device structure and physical relationship are further complicated. , the operation difficulty and cost are greatly increased, and the effect is not ideal

Images