Laser point cloud outdoor scene automatic segmentation method based on machine learning

Abstract

The invention relates to a laser point cloud outdoor scene automatic segmentation method based on machine learning. The laser point cloud outdoor scene automatic segmentation method comprises the following steps: extracting each type of voxel data in a database according to an approximately same proportion, storing the voxel data in a training set and a verification set, and compressing all data;importing the compressed training data set and the verification set into a three-dimensional convolutional neural network for operation, so that feature vectors of various ground objects are extracted; feeding back and adjusting the weight value of the deep neural network through a cost function between the feature vector and the actual value of each type of ground object, carrying out repeated iteration until the cost function is smaller than a set threshold value, and then obtaining a model storing the optimal weight value; optimizing the spatial information and the depth information by using the neighborhood algorithm, so that the removed voxel points can be classified, and the previous coarse classification result can be optimized, and fine classification can be realized, and the accuracy and efficiency of scene segmentation can be greatly improved.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence recognition, in particular to a method for automatically segmenting outdoor scenes of laser point clouds based on machine learning. Background technique [0002] Lidar is a non-contact and active technology to quickly acquire three-dimensional dense point clouds on the surface of objects. It can directly acquire massive and irregularly distributed three-dimensional point clouds with information such as three-dimensional coordinates and echo times, and is less affected by weather. At present, it plays an important role in the fields of global change, smart city, resource survey, environmental monitoring, basic surveying and mapping. However, in actual production at present, due to the complexity of terrain changes, the diversity of ground objects, and the uneven density of points, the reconstruction of point cloud object scenes is done manually or semi-automatically, requiring a lot...

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

Images