Matterport, Inc.

Costar Realty Information, Inc.

Matthieu Francois Perrinel

Kevin Balkoski

Eleanor Crane

Edward Melcher

Example systems, methods, and non-transitory computer readable media are directed to obtaining a point cloud that represents an environment based at least in part on a plurality of points in three-dimensional space; determining corresponding classifications of points in the point cloud as ground or not-ground based at least in part on a plurality of ground classification algorithms; determining respective point cloud features associated with the points in the point cloud; determining respective cell features associated with a plurality of cells that segment the point cloud; generating feature data for a machine learning model based at least in part on one or more of: the classifications of the points based on the plurality of ground classification algorithms, the point cloud features, or the cell features; and classifying the points in the point cloud based at least in part on an output from the machine learning model.

Machine learning techniques for ground classification

An example method comprises applying learning (e.g., weights) developed for training a first model for a first data set of image data to training a second model for a second data set of sensor data. The first and the second data sets may be from the same environment. The second data set has a greater number of channels than the first data set. Weights of layers determined in the first model training may be initially applied to training the second model for the second set of data. Channels of the second data set equal to the number of channels of the first data set may be utilized for each of the layers, using the same weights from the first model. All or some of the channels may be applied in training the second model and using the layers, but determining new weights for the generation of the second trained model.

Growing community of inventors

Berkeley, CA, United States of America

Matthieu Francois Perrinel