The commercial drone industry is rapidly maturing – across industries, geographies, use-cases, and business models – thanks to advances in UAV hardware and regulation. An incredibly rich variety of drone hardware has been built, tested, piloted and deployed for enterprise applications – ranging from ‘nano’ category drones for covert surveillance to large, custom-built drones designed to carry multi-kg payloads for last-mile e-commerce delivery.
Nevertheless, a large percentage of commercial use-cases involving drones revolve around the ‘eye in the sky’ capability – making live, high-quality video feeds from drones a key capability of enterprise-grade solutions. With the recent advances in AI/ML technology, the data captured by the drones’ camera can then be automatically processed – in the specific context of that use-case – to derive insights and make informed business decisions.
Live Video Streaming from Drones
Live views of an environment, asset, person, etc. – enabled by drones – are central to use-cases such as situational awareness during natural disasters, inspections of wind turbines and solar farms, intruder detection at secure sites, etc. Drones that stream live video are expected to augment, if not replace, humans as well as fixed monitoring equipment such as CCTVs – driven by their ability to use the 3D space, be remotely operated, carry payloads, be deployed in fleets, and made fully autonomous.
Video from Remote Drone Operations
In use-cases such as public safety, drones can provide live video streams to the emergency operators centers and help accelerate the incident response. In the security & surveillance context, drones can be operated remotely to capture live video, thus preventing humans from being put at risk.
Firefighters can deploy drone fleets all around the incident site to get a 360 degree, real-time video feed and thus employ the most suitable equipment and tactics to control the fire while minimizing safety risks.
From Drone Video Streams To Remote Drone Control
The latency of live video feeds from drones has been brought down to the 500-ms level, and the latency of drone telemetry is in fact much lower at 50ms. Enterprises can thus not only rely on near-real-time video feeds (with quality as high as 4K) from drone fleets – but they can even incorporate remote control of the drone and camera gimbal, into their workflows. Subject-matter experts, for example, need no longer go to the site for asset inspections; instead, they can control the drone from a corporate office, using the live video feed as immediate feedback, saving time, effort and expenses – and improving worker safety at the same time!
A variety of remote stakeholders can thus easily access drone videos simultaneously – with only the drone pilot/operations manager having to be physically on-site. In fact, varying levels of access can be designed to telemetry, images, videos and sensor data – thus ensuring data privacy and security, while empowering the right stakeholders to fulfill their roles in the context of enterprise drone operations.
Enhancing Live Video Streams from Drones
With the availability of cost-effective off-the-shelf drones and airframes, as well as a wide variety of payloads, drone solution providers are crafting the optimal solutions for commercial use-cases. The software of course plays a vital role in enabling drone stakeholders to plan, execute, log, monitor and repeat drone missions; the software will also be crucial in maturing most drone operations from manual to fully autonomous.
However, drone payloads such as thermal cameras, IR cameras, sirens, lights, camera gimbals, charging pads, etc. also are crucial for delivering value. Live video streams from dual cameras in drones can power night-time missions, while the ability to record video streams (on local servers or on the cloud) can support the investigation of security incidents and audits of security services.
Insights from Video Streams from Drones
The rich, real-time image and video data captured by drone cameras serve as a rapidly increasing repository of ‘training data’ for AI/ML algorithms to use as part of intelligent automation of enterprise automation. For example, live video feeds from drones that monitor industrial premises can be complemented with trained AI/ML models that can automatically detect humans, animals, and objects – and trigger security alarms. Similarly, video streams from public safety situations can be automatically analyzed to aid emergency response teams to identify suspects and/or victims in that context. Drone solution providers are in fact rapidly building capabilities to automatically detect weeds, pests, crops, cattle, etc. – with high-quality video data from drone fleets as the key enabler.