Three-dimensional reconstruction method, point cloud fusion method, device, equipment and storage medium

Abstract

The present invention relates to the field of data fusion technology, and specifically discloses a three-dimensional reconstruction method, point cloud fusion method, device, equipment and storage medium, wherein the fusion method includes the following steps: obtaining the first point cloud information and point cloud sequence information; The self-attention mechanism converts the point cloud sequence information into the second point cloud information set by the fusion parameters; according to the confrontational self-supervised learning algorithm, the comparison loss function about the first point cloud information and the second point cloud information is designed; based on the reverse The propagation algorithm uses gradient descent to update the fusion parameters until the contrast loss function converges to determine the best fusion parameters; according to the best fusion parameters, the first point cloud information and point cloud sequence information are fused to obtain the fusion point cloud information; the fusion The fusion process of the method does not need to pre-label the point cloud information, realizes the fusion of data collected by drones and unmanned vehicles, simplifies the process of point cloud fusion, and improves the efficiency and accuracy of point cloud fusion.

Description

technical field [0001] The present application relates to the technical field of data fusion, in particular, to a three-dimensional reconstruction method, a point cloud fusion method, a device, a device, and a storage medium. Background technique [0002] 3D reconstruction is a research hotspot in the field of computer vision and computer graphics, which aims to recover the given reality-based data into a geometric model. [0003] One of the key points of 3D reconstruction technology is how to obtain the depth information of the target scene or object. Under the condition that the depth information of the scene is known, the three-dimensional reconstruction of the scene can be realized through the registration and fusion of point cloud data. [0004] In the 21st century, with the emergence of various depth cameras for ordinary consumers, 3D scanning and reconstruction technology based on depth data has developed rapidly. [0005] 3D reconstruction can be carried out based ...

AI Technical Summary

Problems solved by technology

[0005] 3D reconstruction can be carried out based on the fusion of point cloud data collected in various ways. The existing point cloud fusion method needs to pre-process the point cloud information when merging the point cloud data collected by unmanned vehicles and the point cloud data collected by drones. Labeling, there are problems of low fusion efficiency and low accuracy

[0006] For the above problems, there is no effective technical solution

Images