Large-scale point cloud semantic segmentation method based on lightweight neural network
A semantic segmentation and neural network technology, applied in biological neural network models, neural learning methods, neural architectures, etc., can solve the problem of high computational complexity of semantic segmentation methods, achieve simplified point cloud semantic segmentation networks, reduce time complexity, The effect of reducing training time
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Publication Date
- 2021-11-09
Smart Images

Figure 1 
Figure 2 
Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the field of three-dimensional point cloud and pattern recognition, and the field of deep learning technology, and more specifically relates to a large-scale point cloud semantic segmentation method based on a lightweight neural network. Background technique
[0002] Semantic segmentation is one of the most important research techniques in the field of computer vision, which aims to divide each pixel or point in the scene into several regions with specific semantic categories. Semantic segmentation is the basis of 3D scene understanding. It has achieved very good results in the fields of map geographic information, navigation and positioning, computer vision, pattern recognition, etc. It has important research significance and broad application prospects. The 3D point cloud semantic segmentation methods based on deep learning are mainly divided into indirect segmentation methods and direct segmentation methods. The indirect me...
Examples
Embodiment Construction
[0027] The specific implementation manners of the present invention are described below, so that those skilled in the art can implement with reference to the description.
[0028] Such as figure 1 , 2 , the present invention provides a 3D point cloud semantic segmentation method based on a lightweight neural network, comprising:
[0029] S1. Use the 3D point cloud training set to train the semantic segmentation neural network model, and the training samples are 3D lidar point clouds {P i |i=1,2,...,n}, where each point P i is a vector (x, y, z, int, ret, num) representing the original (x, y, z) coordinates, intensity, number of returns, and number of returns of the lidar data, respectively, and corresponds to a label y i , representing the real semantic category.
[0030] Perform rasterization, sampling, and normalization preprocessing on the point cloud dataset to establish a training set.
[0031] The data set used in the present invention is a remote sensing three-dime...