Meta's DINOv3 Now Available
Latest integration
We've just added Meta's groundbreaking DINOv3 model thanks to the awesome integration with Hugging Face's 🤗 Transformers.js.
GeoAI for the modern
JavaScript developer
Open-source GeoAI. No backend required. Run models right in your JavaScript apps or edge devices!
Works with your favorite mapping providers
Seamlessly integrate with existing mapping infrastructure
pnpm add geoaiimport { geoai } from "geoai";
// Now with Meta's DINOv3 for image feature extraction
const mapProviderConfig = {
provider: "geobase", projectRef, apikey, cogImagery
};
const pipeline = await geoai.pipeline([
{ task: "image-feature-extraction" }
], mapProviderConfig);
const result = await pipeline.inference(polygon);
Seamless integration of state-of-the-art AI models from TransformersJS with Geobase's geospatial platform
Explore cutting-edge AI models for geospatial analysis
From object detection to feature extraction, discover how AI can transform your geospatial workflows. Try our interactive demos below.
⚡Image Feature Extraction with DINOv3
Powered by Meta's latest DINOv3 model - extract and analyze visual features from satellite imagery using state-of-the-art AI embeddings.
Oil Storage Tank Detection
Detects oil storage tanks in the imagery.
Object Detection
Detects and highlights objects in the imagery using AI models.
Building Detection
Identifies and outlines buildings present in the imagery.
Car Detection
Detects and marks cars and vehicles in the image.
Wetland Detection
Identifies wetland areas such as marshes and swamps in the imagery.
Solar Panel Detection
Detects solar panels and solar farms in the image.
Ship Detection
Detects ships and large vessels in water bodies.
Oriented Object Detection
Detects objects and provides their orientation in the imagery.
Building Footprint Segmentation
Generates segmentation masks for building footprints in satellite imagery.
Land Cover Classification
Classifies terrain and land cover types such as water, forest, or urban areas.
Zero Shot Object Detection
Detects objects without prior training on specific classes using advanced AI.
Zero Shot Segmentation
Segment objects without prior training on specific classes using advanced AI.
Interactive Mask Generation
Generates segmentation masks for features of interest in the image.
Embedding Similarity Search
Finds similar patches in the imagery based on embeddings.