Computer and machine vision are technologies that both capture and analyse images to achieve a useful result. Machine vision typically refers to the use of this technology to make simple decisions in automated industrial processes, whereas computer vision is used for more complex use cases such as drones and robotic vehicles.
Sensors used to acquire images for computer vision may include RGB and thermal cameras, video cameras, and LiDAR scanners.
The data from these may be augmented by other sources such as GPS receivers and environmental sensors.
Raptor Photonics’ Owl 640 M VGA camera for vision enhancement and machine vision
Computer vision image analysis may be performed on board the drone or unmanned vehicle, or the data may be streamed to a cloud server or other computing facility for offsite analysis, thus saving onboard processing power. A variety of analysis methods may be used on the images, including artificial intelligence (AI)-powered object detection and classification. The velocity of objects picked up by the imaging sensors can also be estimated, allowing for tracking.
Sentient Vision Systems’ Kestrel AI-enabled object detection & tracking software
Computer vision can be used by all forms of unmanned vehicle, including drones, UAVs (unmanned aerial vehicles), UGVs (unmanned ground vehicles), USVs (unmanned surface vessels), and UUVs (unmanned underwater vehicles).
A common application of the technology is in detect-and-avoid systems, allowing objects in flight, obstacles on the ground, or hazards on or in the water to be spotted and manoeuvred around. Computer vision is also used as part of SLAM (simultaneous localization and mapping), in which an autonomous vehicle can create a map of a previously unknown area and navigate around it.
Connect Tech’s RudiNX fully integrated solution channels the performance power of the Xavier NX platform for machine vision applications
Applications for computer vision systems in drones and unmanned systems include; detection of forest fires, target detection in counter-drone systems, defect spotting in automated inspection of pipelines, power lines and wind turbines, and identification of missing persons or casualties in search and rescue and other first responder missions.
Join Thousands of Unmanned Systems Professionals
Create a UST Pro account to get full website access, the weekly eBrief and exclusive event discounts.
Searching Companies & Products
Subscribe to the Weekly eBrief
The latest engineering and technical developments straight to your inbox - join thousands of engineers who receive it.
