Spaces:
Running
title: Connectivia Labs
emoji: 🛰️
colorFrom: indigo
colorTo: blue
sdk: static
pinned: false
Connectivia Labs
Applied computer vision for defense, ISR, and agriculture.
Connectivia Labs builds production-grade vision models and platforms that turn raw sensor feeds into operational intelligence. Our work spans defense — drone surveillance, maritime domain awareness, and threat detection — and precision agriculture, where the same core CV stack drives crop health, pest and weed detection, and biomass estimation.
Platforms
| Platform | Domain | Description |
|---|---|---|
| OmniSense Defense | Defense ISR | Multi-sensor ISR prototype combining live CV detection, VLM tactical assessment, fire and satellite intel, and operator workflows |
| MediaSense | Media analysis | Content and media analysis platform built on the OmniSense core |
| YieldPulse | Agriculture | Precision agriculture analytics — biomass, yield, and plant health |
Models
Defense / ISR
- YOLO26-Detection — general object detection (YOLO26 Large)
- DroneDetection — drone-as-object detector
- Maritime_Custom — vessel and ship detection (YOLO26, multi-class)
- ThreatDetection-YOLOv8n — 4-class threat (gun / knife / explosive / grenade)
- MilitaryConvoy-YOLO26L — military vehicle and convoy detection
- yolo26l-pose — pose and keypoint estimation
- ThreatDetection-RFDETR — RF-DETR based threat detection (alternative architecture)
Aerial / VisDrone
Agriculture & Aquaculture
- YieldPulse Biomass — biomass estimation for crop yield forecasting
- WeedDetection-YOLOv8s — weed detection for precision spraying
- PestDetection-YOLO11s — pest detection for IPM workflows
- FishDetection — underwater fish detection for aquaculture
Access
Most models in this organization are private. For licensing, evaluation, or collaboration inquiries, please reach out through our main site.
Models here are trained on a mix of public datasets (VisDrone, ABoships, CCTV-Gun, MilitaryVehicleRecognition, and others) and proprietary data. Individual model cards document their training data, intended use, and limitations.