opengeos/segment-geospatial
Segment Geospatial
Segment Geospatial is a Python toolkit for isolating geographic features in remote sensing imagery using the Segment Anything Model. It functions as a remote sensing image processor that converts map tiles into georeferenced formats to generate segmentation masks from satellite data.
The system enables the extraction of geographic objects through automatic mask generation or manual prompts, such as text descriptions, bounding boxes, and interactive markers. It supports timeseries imagery segmentation to track or identify objects across sequences of images over different dates and provides a geospatial mask visualizer for rendering results on interactive maps.
The project covers a broad range of spatial operations, including map tile acquisition, raster-to-vector conversion, and feature edge reconstruction to refine object boundaries. It also includes a REST API that exposes these segmentation and data processing functions to remote applications.
Export capabilities support georeferenced raster images and standard vector formats including GeoJSON, Shapefile, and GeoPackage.
Features
- Geospatial Imagery Segmentation - Isolates specific geographic features in satellite and aerial imagery using the Segment Anything Model to create georeferenced masks.
- SAM-Based Implementations - Uses the Segment Anything Model (SAM) to generate high-quality binary masks from text, bounding box, or point prompts.
- Automatic Mask Generators - Produces object masks across an entire image automatically without requiring manual user input or prompts.
- Raster Export - Saves segmentation masks as spatial imagery files that maintain a coordinate reference system.
- Geospatial Adaptations - Adapts the Segment Anything Model specifically for isolating geographic features within remote sensing imagery.
- Prompt-Based Semantic Segmentations - Uses a pre-trained Segment Anything Model to generate semantic masks from bounding boxes or text descriptions.
- Remote Sensing Machine Learning - Analyzes sequences of satellite imagery over multiple dates to track geographic objects and identify changes over time.
- Prompt-Based Segmentations - Identifies map objects using text descriptions, bounding boxes, or interactive foreground and background markers.
- Timeseries Analysis - Analyzes sequences of remote sensing images over multiple dates to track geographic objects across time.
- Object Tracking - Identifies and tracks specific geographic objects across sequences of remote sensing images over different dates.
- Map Tile Downloaders - Fetches map imagery from remote tile services and converts the data into GeoTIFF files for local processing.
- Georeferenced Format Converters - Converts map tiles from remote services into georeferenced formats to enable automated object detection.
- Georeferenced Raster Transformations - Converts raw map tiles into GeoTIFF formats to maintain coordinate reference systems during the segmentation process.
- Raster-to-Vector Conversions - Transforms pixel-based segmentation masks into georeferenced vector formats like GeoJSON and Shapefiles.
- Batch Mask Generation - Generates multiple segmentation masks in a single execution cycle using collections of point coordinates from vector layers.
- Segmentation APIs - Provides a REST API to expose remote sensing segmentation and geospatial data processing capabilities to remote clients.
- Geospatial Boundary Refinements - Applies a feature edge reconstruction algorithm to refine the boundaries of segmented geospatial objects.
- Geospatial Visualizations - Provides an interactive mapping interface to render and verify segmentation masks using geographic coordinates.
- REST API Interfaces - Provides a server-side REST API to expose geospatial segmentation and data processing functions to remote clients.
- Geospatial Mask Visualizers - Ships an interactive mapping interface for rendering and verifying the spatial accuracy of extracted segmentation masks.
- RESTful Services - Exposes geospatial segmentation and data processing functions through a stateless RESTful service layer.
- Computer Vision and Image Processing - Geospatial data segmentation using foundation models.
- Geospatial Machine Learning - Adapts foundation models for geospatial segmentation.
