About

The frequent inspection of railway tracks is essential to ensure safe and efficient operation of trains. Major derailments and fatal accidents may happen due to human failure or infrequent and improper monitoring systems. Traditionally, track inspection is carried out by human inspectors or via human-operated ground vehicles and involves investigation of various types of rail components and their properties to detect visually noticeable defects, such as missing fasteners, clips, and encroaching vegetation. However, this manual human-operated process is time-inefficient, costly, and complex, while it also incurs a very suboptimal and sparse pattern of observability of tracks. StreamOwl provides a novel full-fledged solution for aerial monitoring of railways using Unmanned Aerial Vehicles (UAVs), or drones, for railway inspection and identification of defects and anomalies in the structure of the rails. The drone operates in an autonomous unsupervised manner, even in GPS-denied areas, and is able to navigate over a railway track for a long distance, inspect, detect and identify anomalies along the railway (rail defects, fallen leaves, sun kinks, gauge deviation, sideline vegetation etc.), thanks to industry-leading Computer Vision and Deep Learning capabilities.

Solution & Product

StreamOwl provides a novel full-fledged solution for aerial monitoring of railways using Unmanned Aerial Vehicles (UAVs), or drones, for railway inspection and identification of defects and anomalies in the structure of the rails. The drone operates in an autonomous unsupervised manner, even in GPS-denied areas, and is able to navigate over a railway track for a long distance, inspect, detect and identify anomalies along the railway (rail defects, fallen leaves, sun kinks, gauge deviation, sideline vegetation etc.), thanks to industry-leading Computer Vision and Deep Learning capabilities. Integrated 3D cameras and peripheral sensors (e.g., lidars, ultrasound sensors) enable adaptive 3D path planning, and safe navigation based on advanced flying capabilities (including collision-avoidance, object detection and target proximity estimation). The provided solution enables faster inspections, large coverage even in adverse conditions, ease of deployment, and safety during operation.

Innovation & Scalability

The use of drones instead of ground vehicles (e.g., rovers) is motivated by the need for quick overview, inspection, and large coverage of the railways in adverse terrains, where regular inspection operations cannot be applied, are costly, and time-inefficient. A drone can cover large distances and can overcome possible large obstacles (e.g. fallen trees) on the railway (while a rover cannot go through a large tree that has fallen on the rails). Drones are also easier to deploy since they can be carried in a typical passenger car and they do not interfere with the normal operation of the trains as they fly above the train level (~3-20m above the ground).t StreamOwl has developed a UAV with a programmable flight controller that enables route planning in an autonomous non-supervised manner and an on-board companion computer and 3D cameras to record, scan, and analyze vegetation along the railway in real-time. A set of sensors (lidar, ultrasonic sensor) allow the adaptation of the route to slopes, the avoidance of obstacles and collisions, and the automatic adaptation of the route in every terrain.t Moreover, the UAV is powered by industry-leading AI skills based on advanced Computer Vision and Deep-Learning capabilities which enable the detection and identification of anomalies and problems along the railway lines (objects, fallen trees/leaves, sun kinks, gauge deviation, etc.) and 3D cartography of encroaching vegetation.t

Traction & achievements

The company was incorporated in 2014, in Athens.t - 2016: StreamOwl selected in the top-10 startups by MITEF Greece and attended a bootcamp at MIT.t - 2016: StreamOwl selected in top-10 startups by NBG Seeds programme.t - 2019-2020: StreamOwl receives 200k grant by EC projects ESMERA and ROSIN on autonomous UAVs.t 2020: our solution has reached TRL-9. In the context of our above EC projects, we have successfully demonstrated and validated the UAV in operational environment and the system has been proven and made ready for full commercial deployment.tWe havet conducted successful demos with NetworkRail, UK’s largest rail network operator, and published a research paper on the developed solution in the Int. Conf. on Unmanned Aircraft systems (https://bit.ly/3rCsh1d).t 2020: We have joined the RAIL-Activation project to enhance our strategic partnerships with European rail operators.

Team & Why Us

Dr. Savvas Argyropoulos is the CEO and founder of StreamOwl and leads the R&D of computer vision and machine learning algorithms. He holds a Ph.D. (2008) in electrical and computer engineering from the Aristotle University of Thessaloniki. He was a Senior Research Scientist at Deutsche Telekom Innovation Laboratories (2009-2014), being responsible for technology transfers to the Business Units of Deutsche Telekom. He has 30+ publications (h-index:14), and filed 10 patents, which have been transferred to Deutsche Telekom. He has 15+ years of technical and entrepreneurial experience.t George Papadopoulos holds a diploma degree from the Dept. of Mechanical Engineering of the Hellenic Air Force Academy and M. Sc. in “Polymer Science and Technology” from Univ. of Patras. He leads the design and development of the UAV (hardware+software) and he is responsible for field testing. In 2016, he founded “Drone2Ship” and joined the accelerator program of EUROBANK “EGG”.

Funding Needs

We are actively looking for Seed funding to scale our activities and commercialise our solutions.