3D point cloud semantic segmentation method based on position attention and auxiliary network

Abstract

The invention provides a 3D point cloud semantic segmentation method based on position attention and an auxiliary network, mainly solves the problem of low segmentation precision in the prior art, andadopts the implementation scheme which comprises the following steps of acquiring a training set T and a test set V; constructing a 3D point cloud semantic segmentation network, setting a loss function of the network wherein the network comprises a feature down-sampling network, a position attention module, a feature up-sampling network and an auxiliary network which are cascaded in sequence; performing P rounds of supervised training on the segmentation network by using the training set T: in each round of training process, adjusting network parameters according to a loss function, and afterthe P rounds of training are completed, taking the network model with the highest segmentation precision as a trained network model; and inputting the test set V into a trained network model for semantic segmentation to obtain a segmentation result of each point. The 3D point cloud semantic segmentation precision is improved, and the method can be used for automatic driving, robots, 3D scene reconstruction, quality detection, 3D drawing and smart city construction.

Description

technical field [0001] The invention belongs to the technical field of data processing, and in particular relates to a 3D point cloud semantic segmentation method, which can be used in automatic driving, robots, 3D scene reconstruction, quality inspection, 3D mapping and smart city construction. Background technique [0002] In recent years, with the wide application of 3D sensors such as lidar and RGBD cameras in the fields of robots and driverless vehicles, the application of deep learning in 3D point cloud data has become one of the research hotspots. 3D point cloud data refers to: a collection of a set of vectors in a three-dimensional coordinate system. These vectors are usually expressed in the form of x, y, and z three-dimensional coordinates, and are generally used to represent the outer surface shape of an object. In addition, in addition to the geometric information represented by (x, y, z), it may also contain information such as RGB color, intensity, gray value, ...

AI Technical Summary

Problems solved by technology

[0005] Compared with the traditional method, the above two methods directly process 3D point cloud data, the calculation is simple, effectively solve the characteristics of point cloud disorder and improve the segmentation accuracy, but PointNet++ does not consider the relationship between the characteristics of each center point. The relationship between them, that is, the context information, so the feature representation is relatively weak. In addition, PointNet++ complies with the general framework of encoding-decoding, and does not consider more information at the bottom layer. Therefore, the segmentation accuracy is not high, and there is still room for improvement.

Images