Nicheloom

Market intelligence for builders — see what's gaining traction before it's crowded.

Satellite imagery object detection using text prompts

Details

External ID
47305979
Source
HN
Company
—
Product
Satellite imagery object detection using text prompts
Website domain
useful-ai-tools.com
Launched
March 9, 2026
Cohort
—
Upvotes
53
Upvotes percentile
0.8400984009840098
Tags
—
Fetched at
Sept. 7, 2026, 9:26 p.m.
Updated at
Sept. 7, 2026, 9:26 p.m.

Description

I built a browser-based tool for detecting objects in satellite imagery using vision-language models (VLMs). You draw a polygon on the map and enter a text prompt such as "swimming pools", "oil tanks", or "buses". The system scans the selected area tile-by-tile and returns detections projected back onto the map as GeoJSON.Pipeline: select area and zoom level, split the region into mercantile tiles, run each tile with the prompt through a VLM, convert predicted bounding boxes to geographic coordinates (WGS84), and render the results back on the map.It works reasonably well for distinct structures in a zero-shot setting. occluded objects are still better handled by specialized detectors like YOLO models.There is a public demo and no login required. I am mainly interested in feedback on detection quality, performance tradeoffs between VLMs and specialized detectors, and potential real-world use cases.

Enrichment

Theme
ai image prompts and developer tools
Vertical
Horizontal
Function
Analytics & BI
Audience
B2B
AI stance
AI-native
Project type
Commercial product
Normalized one-liner
object detection on satellite imagery with text prompts
Manually corrected
False

Could you build this?

Partial The map interface and tile slicing can be built with standard web mapping tools, but tiling high-resolution satellite imagery and running low-latency VLM geospatial object detection requires specialized spatial inference pipelines.

What it would actually take: The architecture requires a client map UI (e.g., MapLibre/Leaflet) connected to a geospatial backend using GDAL/Rasterio to slice high-resolution imagery into standard Slippy map tiles. The backend orchestrates batch inference across tiles using vision models (e.g., Grounding DINO or open-vocabulary VLMs) running on GPU clusters, followed by Non-Maximum Suppression (NMS) and projection back into EPSG:4326 GeoJSON coordinates. High throughput and precision require deep familiarity with geospatial pipelines and computer vision serving.

Discussion

20 comments analyzed.

Competitors mentioned: Google Street View object detection, Planet Labs (ship detection), Maxar satellite imagery, AIS/ADSB tracking systems

Concerns raised: Model struggles with non-generic/obscure objects from above, Search not responding for certain locations (Las Vegas, San Antonio), Mobile selection polygon size too small, Low satellite image resolution for matching objects like ships, Accuracy vs. LLM interface usability tradeoff

Feature requests: Image-based object search (find similar areas from screenshot/photo), Change detection using Sentinel-2 data over time, Notification/sound alert when scan completes, Support for uploading custom aerial/satellite imagery, Better UI feedback for scan progress/results loading

Competitors

Other products that read as similar to this one — 40 launches clear the similarity bar, closest 8 shown.

Attention rank: #7 of 41 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).

Launched 129 days after the earliest competitor.

Other launches for this product

Same idea, different domain

Nobody's really built a analytics & bi tool for Legal yet.