Plane extraction method and system based on sparse points and electronic equipment

Abstract

The invention discloses a plane extraction method and system based on sparse points and electronic equipment. The plane extraction method based on sparse points comprises the following steps: performing spatial point extraction processing on acquired image information to obtain a plurality of spatial points in a world coordinate system; constructing one or more candidate planes based on four spatial points randomly selected from all the spatial points; plane covariance calculation is carried out on all the candidate planes, so that the candidate planes meeting a preset plane screening condition are used as planes to be confirmed; and performing singular value calculation of a plane matrix on all the to-be-confirmed planes, and outputting the to-be-confirmed plane meeting a preset output condition as a finally extracted plane.

Description

technical field [0001] The invention relates to the technical field of SLAM, in particular to a plane extraction method based on sparse points, a system thereof and electronic equipment. Background technique [0002] At present, augmented reality (English Augmented Reality, AR for short) applications based on mobile devices have become a reality, and are favored by major mobile device manufacturers and users. Since the anchor point of the virtual object needs to be determined based on the plane when integrating the virtual object into the real environment and interacting with multiple people, so as to render the virtual object at the determined anchor point, the plane detection technology is extremely important. [0003] There are many existing plane detection methods: the first plane detection method is a method based on deep learning. First, the neural network information and plane segmentation information of the image are obtained, and then the above information is input ...

AI Technical Summary

Problems solved by technology

[0004] However, the problem with the above-mentioned first method is that it is usually impossible to correctly segment different plane regions through the network model, especially in a building scene with multiple planes, it is even more impossible to correctly segment different plane regions; the above-mentioned No. The second method is because the dense point cloud is obtained by extracting feature points from multiple images, and the dense point cloud contains more feature points, which leads to high computational complexity of this method, high hardware requirements, and high cost. restricts its scope of use

However, the computing power of mobile devices is limited, and the algorithm with high complexity is used for plane detection, which is slow and the user experience is poor.

Images