Modular AI Software Suite for Highly Reliable Perception & Autonomy in UAVs, Drones, & Self-Driving Vehicles

Spleenlab is a developer of safe artificial intelligence (AI) and machine learning (ML) technologies that enable highly reliable perception and autonomy for UAVs, drones, and self-driving vehicles.

We have developed unique solutions that enable perceptual AI to be approved for highly safety-critical systems, using a unique approach that blends large-scale data training and sophisticated sensor fusion. Our rigorous machine learning concepts and algorithms have been submitted to the highest-ranked and most competitive AI conferences in the world, and we have collaborated closely with leading academic institutions, laboratories and industry figures to ensure the success and utmost safety of our software.

Our state-of-the-art VISIONAIRY® modular software suite provides a wide range of capabilities for airborne and ground-based vehicles.  It can be integrated with all leading drone flight controllers and robotic operating systems including PX4, Auterion and ROS.

VISIONAIRY® for UAVs & Drones

GPS-Denied

VISIONAIRY® provides cutting-edge SLAM (simultaneous localization and mapping) capabilities that enable precision autonomous navigation even in the absence of GPS and GNSS signals. This allows for vision-based landing and the traversal of challenging environments such as underground and in urban canyons.

Collision Avoidance

VISIONAIRY®’s collision avoidance software provides reliable object detection at ranges of up to 100 metres, and is compatible with any RGB and infrared cameras and gimbals on the market. This critical capability is essential for BVLOS and fully autonomous operations, and enables a wide range of use cases including drone delivery, wide-area mapping and surveying, and navigation in GNSS-denied areas.

Safe Precision Landing

VISIONAIRY® enables safe fully-automated precision with centimetre-level accuracy both during the day and at night, detecting all obstacles and moving objects within the landing zone including vehicles, people and animals. The software supports all UAS cameras and LiDAR scanners, and utilizes redundant design to ensure safe operation.

Industrial Inspection

Our powerful software enables fully autonomous inspection of power lines, wind turbines, roads and railways, pipelines, and a wide array of other industrial assets. With real-time navigation capabilities and the highest performance-to-weight ratio on the market, VISIONAIRY® allows you to minimize downtime and save significantly on costs and manpower.

Surveillance

With state-of-the-art real-time detection, identification and tracking of multiple objects and diverse object classes, VISIONAIRY®’s Surveillance module ensures excellent performance even under highly challenging conditions.

Detect & Avoid

Robust detect-and-avoid (DAA) capabilities are crucial for certified BVLOS flights and fully autonomous drone operations. VISIONAIRY® provides real-time geolocalization and tracking of air traffic at long ranges up to several kilometers and through 360 degrees.

VISIONAIRY® for Self-Driving Cars & Autonomous Ground Vehicles

Online Calibration & Sensor Fusion

VISIONAIRY® provides multi-sensor calibration of automotive cameras, radar and LiDAR scanners for sensor fusion systems, ensuring that accuracy is kept constant as factory calibrations drift. The advanced software can simultaneously detect errors and calculate the correct calibration factors in real time without the need for targets or hand-crafted features.

MonoDepth & Object Detection

The world’s first AI system for functional automotive safety, VISIONAIRY® MonoDepth & Object Detection supports a wide range of use cases including ADAS, parking, emergency braking, and highly automated driving. Providing true depth estimation from just a single sensor, it is compatible with all types of cameras including fisheye, pinhole, 360-degree, and infrared.

SLAM for Driving

The VISIONAIRY® SLAM for Driving module is specifically designed for automotive applications such as Level 2+ ADAS (advanced driver assistance systems), trained parking, and high levels of automated driving (L4+). The ASIL-ready software supports all automotive cameras and LiDAR scanners on the market.