Autonomous Drone Vision System for KrattWorks

AI Solutions • Defense & Public Safety • Computer Vision

Autonomous Drone Vision System for KrattWorks

Agmis partnered with KrattWorks to develop onboard machine vision algorithms that enable autonomous object detection, transforming aerial surveillance for defense operations and emergency response.

Real-Time
Object Detection
Autonomous
Edge Processing
Instant
Data Transmission

KrattWorks - Pioneering Autonomous Aerial Intelligence

KrattWorks is an ambitious Estonian startup that emerged from the Agmis Accelerator Program, now recognized as a leader in autonomous drone technology. The company develops drones equipped with onboard machine vision and GSM connectivity, enabling real-time object detection and instant data transfer to control centers via mobile networks. KrattWorks recently secured a €15 million contract with the Estonian Centre for Defence Investment to supply aerial target drones to the Estonian Defence Forces over a 7-year period - a testament to their technological capabilities and market trust.

Traditional Surveillance Drones Fall Short in Mission-Critical Scenarios

Bandwidth-Heavy Video Streams

Conventional drone surveillance requires continuous video transmission, consuming massive bandwidth and creating delays that compromise real-time decision-making in emergency situations.

Manual Monitoring Dependency

Traditional systems demand human operators to constantly watch video feeds, introducing fatigue-related errors and limiting the scale of surveillance operations.

Delayed Situational Awareness

First responders often rely on outdated technology where every wasted minute can cost human lives - whether in wildfire detection, search and rescue, or defense operations.

Limited Autonomous Capability

Most drone platforms lack the intelligence to identify and classify objects independently, requiring constant human intervention and limiting their effectiveness in BVLOS (Beyond Visual Line of Sight) missions.

Onboard Computer Vision for Autonomous Object Detection

Real-Time Edge Processing: Our computer vision algorithms run directly on the drone's onboard module, enabling autonomous detection of humans, vehicles, fire outbreaks, and other critical elements without relying on ground-based processing or bandwidth-heavy video streams.

Instant Vectorized Data Sharing: Instead of streaming raw video, the system transmits only essential detection data to control centers via GSM networks - dramatically reducing bandwidth requirements while accelerating response times.

Multi-Target Recognition: The machine vision system simultaneously identifies and classifies multiple object types, providing comprehensive situational awareness across defense, firefighting, and public safety applications.

Seamless Integration: Our algorithms integrate with KrattWorks' existing hardware architecture, including their Ghost Dragon multirotor UAV and clip-on gimbal systems, extending computer vision capabilities to various drone platforms.

Autonomous Decision-Making: The system enables true edge autonomy - no manual surveillance, no delay, no large data streams - just intelligent detection and instant alerts when it matters most.

Fail-Safe Operation: Combined with KrattWorks' fail-safe features, the computer vision system maintains reliable performance even when connectivity is intermittent, ensuring mission continuity in challenging environments.

Mission Impact

Edge AI

Autonomous Detection at Point of Observation

Measurable Improvements Across Defense and Public Safety Operations

Eliminated Manual Surveillance

Autonomous object detection removes the need for constant human monitoring, freeing operators to focus on strategic decision-making rather than watching endless video feeds.

Reduced Data Transmission Load

Vectorized data sharing slashes bandwidth requirements compared to traditional video streaming, enabling reliable operation even in areas with limited connectivity.

Accelerated Response Times

Real-time detection and instant alerts compress the window between threat identification and response - critical in firefighting where early detection can prevent disasters.

Expanded Operational Range

Onboard processing combined with GSM connectivity enables unlimited operational range within mobile network coverage, supporting true BVLOS missions without range restrictions.

Vision for the Future

Our joint work with Agmis to develop real-time onboard object detection is a game-changer for our platform. We envision a future where there's no manual surveillance, no delay, no large data streams - just autonomous decision-making at the edge. It takes elite engineering to make that happen, and Agmis delivers exactly that.

Mattias Luha Founder of KrattWorks

Autonomous Fire Detection

The computer vision system identifies fire outbreaks in their earliest stages, transmitting precise location data to first responders before flames gain uncontrollable momentum - potentially saving up to 10x daily firefighting operational costs.

Defense-Grade Performance

Technology proven through KrattWorks' €15 million contract with Estonian Defence Forces, demonstrating reliability and performance that meets military specifications.

Scalable Multi-User Access

Detection data distributes simultaneously to multiple recipients - rescue teams, mission control centers, and authorized third parties - creating seamless situational awareness across all stakeholders.

Platform Flexibility

Computer vision capabilities extend beyond KrattWorks' proprietary drones through clip-on gimbal solutions, bringing autonomous detection to existing drone fleets.

Transforming Aerial Surveillance for Critical Missions

Each year, wildfires destroy an area equal to half of US territory, taking 300,000 lives and causing 15% of global CO2 emissions. In defense scenarios, situational awareness directly impacts mission success and personnel safety. This computer vision-driven approach delivers transformative value through:

Edge Intelligence

Processing happens onboard the drone itself, eliminating latency and enabling autonomous decision-making at the point of observation rather than in distant control rooms.

Mission-Critical Reliability

Technology validated through defense contracts and emergency response testing with Estonian and Portuguese rescue services, proving performance when lives depend on it.

Resource Optimization

Autonomous detection multiplies the effectiveness of limited personnel, enabling broader coverage without proportional increases in human operators.

Future-Ready Architecture

The platform supports KrattWorks' vision of fully autonomous aerial systems for defense, disaster management, and critical infrastructure monitoring - positioning both companies at the forefront of drone innovation.