Spatial Statistics
Analyze patterns, clusters, and relationships in your spatial data.DBSCAN Clustering
Identify clusters of spatial points using density-based clustering. DBSCAN finds core points with enough neighbors within a distance threshold and groups connected points into clusters.
Hotspot Detection
Find statistically significant hotspots and coldspots with the Getis-Ord Gi* statistic. Identify areas with unusually high or low concentrations of values.
Outlier Detection
Detect spatial outliers using Local Outlier Factor (LOF). Identify points that are anomalous compared to their spatial neighbors.
K-Nearest Neighbors
Find the K nearest neighbors for spatial joins. Efficiently match points to their closest neighbors in another dataset.
Raster Inference
Run machine learning models on satellite and aerial imagery at planetary scale.Raster Inference Overview
Learn how WherobotsAI Raster Inference enables you to run ML models on raster data at scale using distributed computing.
Inference Tasks
Classification
Classify land cover and features in raster imagery. Assign categories to pixels or tiles based on their spectral signatures.
Object Detection
Detect objects like buildings, vehicles, and infrastructure in imagery. Get bounding boxes and confidence scores for detected objects.
Segmentation
Segment imagery into meaningful regions. Generate pixel-level masks for different classes or objects.
Text-to-Segments (SAM2)
Use natural language prompts to segment imagery with SAM2 and OWLv2. Describe what you want to find and get segmentation masks.
Custom Models
Bring Your Own Model
Deploy your own trained ML models for raster inference. Support for PyTorch models exported in PT2 format with MLM (Machine Learning Model Extension Specification) metadata.
Routing & Networks
Map Matching
Match GPS traces to road networks for accurate routing analysis. Clean noisy GPS data by snapping points to the most likely road segments.