A Machine Learning-Based Automatic Segmentation Method of Laser Point Cloud Outdoor Scene

Abstract

The present invention relates to a method for automatically segmenting outdoor scenes of laser point clouds based on machine learning, comprising: extracting each type of voxel data in a database in approximately the same proportion, storing them in a training set and a verification set, and compressing all data; Then import the compressed training data set and verification set into the three-dimensional convolutional neural network for calculation, so as to extract the feature vectors of various types of features; feedback and adjust through the cost function between the feature vectors and actual values ​​of each type of features The weight value of the deep neural network is iterated repeatedly until the cost function is less than the set threshold, and then a model storing the optimal weight value can be obtained; then the spatial information and depth information are used to optimize the neighborhood algorithm, so that both The removed voxel points are also classified, and the previous rough classification results can be optimized to achieve fine classification and greatly improve the accuracy and efficiency of scene segmentation.

Description

An automatic segmentation method of laser point cloud outdoor scene based on machine learning technical field The present invention relates to artificial intelligence identification technology field, relate in particular to a kind of outdoor laser point cloud based on machine learning Scene automatic segmentation method. Background technique Lidar is a non-contact active technology that can quickly acquire three-dimensional dense point clouds on the surface of objects. Obtain massive, irregular spatially distributed 3D point clouds with information such as 3D coordinates and echo times. It is currently playing an important role in the fields of global change, smart city, resource survey, environmental monitoring, basic surveying and mapping, etc. use. However, in actual production, due to the complexity of terrain changes, the diversity of ground objects and the uneven density of points The scene reconstruction of point cloud objects is done manually or semi-autom...

AI Technical Summary

Problems solved by technology

[0003] The embodiment of the present invention solves the technical problems of low recognition efficiency, low recognition accuracy and incapable of full automation in the field of outdoor point cloud scene reconstruction in the prior art by providing a machine learning-based automatic laser point cloud outdoor scene segmentation method

Images