Computer vision
Computer vision is about using cameras to understand the world in 3D. We can use computer vision to locate and find, count, measure (size and distance), model and recognise a multitude of things in controlled environments and in the field.
We utilise machine learning, 6D pose estimation, real-time embedded systems, and 3D modelling and measurement to provide solutions in mobile platform location and orientation, product tracking and logistics, industrial metrology and the measurement and counting of natural products.
We provide customised solutions to maximise information extraction and optimise system performance.
Our customers target applications in dimensional metrology, as-built surveying, quality control and CAD models from scans.
Moving platforms
A ‘platform’ is a moving instrument or vehicle that knows where it is in 3D. Handheld instruments, mobile robots or drones are all moving platforms.
Moving platforms are used to:
- observe from ideal distances and angles, enabling views of hard-to-see surfaces
- get around obstructions and avoid occlusions, enabling more complete models
- scan in small spaces where long stand-off is not possible
- observe details up close – it all gets stitched together in one 3D mode
- keep the operator in the loop through intuitive ergonomic scanning
Locate
We can locate and recognise objects from a scene by using:
- algorithms
- machine learning models
- 2D/3D imaging
Map
We can use computer vision to collect information such as:
- colour
- depth
- thermal
- hyperspectral/IR
Track
We can enable simultaneous location and mapping (SLAM) to enable deep learning.
Measure
We can measure everything from products in tin cans to logs, colourful stones, activities in industrial plants or workshops.
Our people
Meet our Computer Vision Team.
Case study
Thermal imaging for robotic carcass processing
Business opportunity
Automation of pelt cut over animal belly region.
What we did
Develop thermal camera imaging data analytics for robotic processing of animal carcases.
Result
Increased processing line productivity using machine vision to locate midline, identify insertion point and animal profile.