Dense Incremental Metric-Semantic Mapping via Sparse Gaussian Process Regression

We develop an online probabilistic metric-semantic mapping approach for autonomous robots relying on streaming RGB-D observations. We cast this problem as a Bayesian inference task, requiring encoding both the geometric surfaces and semantic labels (e. g. , chair, table, wall) of the unknown environment. We propose an online Gaussian Process (GP) training and inference approach, which avoids the complexity of GP classification by regressing a truncated signed distance function representation of the regions occupied by different semantic classes. Online regression is enabled through sparse GP approximation, compressing the training data to a finite set of inducing points, and through spatial domain partitioning into an Octree data structure with overlapping leaves. Our experiments demonstrate the effectiveness of this technique for large-scale probabilistic metric-semantic mapping of 3D environments. A distinguishing feature of our approach is that the generated maps contain full continuous distributional information about the geometric surfaces and semantic labels, making them appropriate for uncertainty-aware planning.

Authors

• Ehsan Zobeidi
• Alec Koppel
• Nikolay Atanasov 0001

Keywords

• Three-dimensional displays
• Semantics
• Gaussian processes
• Probabilistic logic
• Planning
• Task analysis
• Surface treatment
• Gaussian Process
• Kriging
• Sparse Gaussian Process Regression
• Training Data
• Probability Function
• Finite Set
• Probabilistic Approach
• Semantic Labels
• Geometric Surface
• Signed Distance Function
• Training Dataset
• Posterior Probability
• Number Of Observations
• Precision And Recall
• Class I
• Object Classification
• Online Training
• Mean Function
• Test Locations
• Covariance Function
• Finite Set Of Points
• Online Prediction
• Sensor Observations
• Arbitrary Position
• Semantic Segmentation Algorithms
• Short Values
• Matérn Covariance Function
• Precision Matrix
• RGB-D Sensor
• Closed Set