XONIX
VISION ENGINEERING
Xonix Vision System
The Xonix Vision System (xVS) is a state of the art embedded computer solution offering real-time image processing using Artificial Intelligence and advanced image processing algorithms
​
The xVS utilises the latest in GPU technology, providing power efficient AI computing. The deep learning solutions are designed to be embedded and produce fast image processing, either in real-time or offline, to provide automation in the areas of damage identification, routine inspection, object classification and object recognition
TECHNOLOGY
With a completely scalable model, our technology allows us to adapt to individual requirements and provide a highly specialised system. Whether using 3D imaging, thermal imaging, inspection imagery, LIDAR or a combination of different imaging techniques, our system is designed to be modular and extensible.
Having a system that is both modular and autonomous limits the amount of unnecessary working time, creating an extremely cost effective solution. The xVS operational workflow ultimately means far less personnel will be required, to complete the designated tasks, when compared with conventional workflow practices.
 ModularWith a completely scalable model, our technology allows us to adapt to individual requirements and provide a highly specialised system. Whether using 3D imaging, thermal imaging, inspection imagery, LIDAR or a combination of different imaging techniques, our system is designed to be modular and extensible. The PVS optical solution can be completely modified creating a truly bespoke system. |  CostHaving a system that is both modular and autonomous limits the amount of unnecessary working time, creating an extremely cost effective solution. The PVS operational workflow ultimately means far less personnel will be required, to complete the designated tasks, when compared with conventional workflow practices. |
|---|---|
 AutonomyFrom the moment the embedded devices are trained, we aim to provide a solution that is completely autonomous. ​ For instance, if mounting the embedded device on a UAV, our systems can be used to recognise objects, gather the images and then process the images in real-time. ​ With UAV technology and our collision avoidance, personnel are not exposed to undue risk. |
The xVS optical solution can be completely modified creating a truly bespoke system.
Autonomous, real-time analaysis of imagery reduces the volume of data, post process, which reduces costs.
From the moment the embedded devices are trained, we aim to provide a solution that is completely autonomous.
​
For instance, if mounting the embedded device on a UAV, our systems can be used to recognise objects, gather the images and then process the images in real-time.
​
With UAV technology and our collision avoidance, personnel are not exposed to undue risk.
UAV INTEGRATION
Conventional UAV inspection workflow would generally have a dedicated pilot to fly to the inspection point, traverse the object, take images at pre-defined locations, return to base and then analyse the images.
Using the xVS provides a different method of workflow:
-- Initially the artificial intelligence system is trained with images representing known issues.
-- The mission is uploaded to the UAV.
-- The UAV will fly to the initial object -- using advanced object recognition techniques, the xVS starts to autonomously traverse the object, whilst controlling the flight controller to ensure movement is smooth and directional.
-- Images are taken at a pre-defined frame rate and analysed in real-time.
-- The result is then classified and stored in the xVS relational database system.
-- The UAV moves onto the next waypoint when the object inspection is complete.
The analysis software will allow the inspector to view all of the classifications, where anomalies were found, simplifying the previously laborious task of inspecting
each individual image.
During the inspection the xVS will have stored everyimage that was taken of the object building a data-base of future, historic data, which can be used to further train the xVS artificial intelligence system.
 AnalysisThe analysis software will allow the inspector to view all of the classifications, where anomalies were found, simplifying the previously laborious task of inspecting each individual image. During the inspection the PVS will have stored everyimage that was taken of the object building a data-base of future, historic data, which can be used to further train the PVS artificial intelligence system. |  FutureUsing the PVS provides a different method of workflow: -- Initially the artificial intelligence system is trained with images representing known issues. -- The mission is uploaded to the UAV. -- The UAV will fly to the initial object -- using advanced object recognition techniques, the PVS starts to autonomously traverse the object, whilst controlling the flight controller to ensure movement is smooth and directional. -- Images are taken at a pre-defined frame rate and analysed in real-time. |
|---|---|
 WorkflowConventional UAV inspection workflow would generally have a dedicated pilot to fly to the inspection point, traverse the object, take images at pre-defined locations, return to base and then analyse the images. |
Post UAV mission, the UAV will return
to base where the inspection data
can be downloaded and imported into the Xonix xVS front end software.
SENSE AND AVOID
Xonix has implemented its own form of Detect & Avoid technology providing full 360 degree monitoring up to a range of 100m. Support is also provided for downwards monitoring up to a range of 120m
The collision avoidance system is integrated into the flight controller system and will take over the control of UAV when objects are detected within user controlled ranges.
Using the collision avoidance system allows the UAV to fly at safe distances providing avoidance data at a rate of 50 times per second
The xVS also incorporates a Proportional Distance Algorithm (PDA), within the flight controller interface, allowing the UAV to decrease speed when objects are detected.