Experience
Oculus Vision has supported machine vision delivery across a wide range of industrial applications, working alongside OEMs, system integrators, and machine builders on projects where performance, reliability, and technical risk matter.
The examples below illustrate the types of challenges we are typically engaged to support, rather than full project case studies.
3D vision robot guidance
Design, development, supply, and commissioning of a 3D vision system to locate cast components and guide precise, repeatable robotic pick operations. Using a Balluff rc_visard stereo camera combined with MVTec Halcon's CAD surface matching, the system determines exact component position and orientation — including yaw, pitch and roll — allowing the robot to pick from a pre-defined point on the component with consistent accuracy across every cycle. Delivered as part of a wider automation project led by a system integrator, with Oculus Vision taking full responsibility for the vision scope.
The system was commissioned successfully and operates reliably at production throughput.
Assembly Verification
Design, development, and commissioning of a vision system for verification of correct assembly of metal components at high production throughput. The inspection challenge required the optical path to be redirected through 180° using a mirror assembly to overcome access constraints within the existing machine — ensuring reliable detection of three critical quality features without introducing bottlenecks or operational complexity.
The vision system was integrated into a wider automation solution by the project delivery team.
Deployed successfully, achieving 100% inspection coverage at full production speed.
High-speed OCR in production environment
Design, development, and commissioning of a custom vision system to read direct part marked codes on the underside of food cans at over 2,000 per minute.
T
he application involved significant technical challenges — variable print quality, unknown code orientation, and extreme throughput — addressed through a custom imaging approach combined with MVTec Halcon DeepOCR and GPU-accelerated batch processing.
Commissioned successfully and operating in production, automatically ejecting any can with an unreadable or incorrect code.
Tablet Inspection
Design, development, and commissioning of a multi-camera vision system retrofitted onto an existing blister packing machine, verifying presence, product identity, integrity, and correct placement of tablets across 64 cavities per pack. The application required a four-camera array with flat field correction and high-precision image stitching to achieve uniform illumination across the full field of view, combined with collimated backlighting to resolve the edges of fully transparent tablets. The system handles a wide range of SKUs — from fully opaque to fully transparent tablets — with an intuitive operator interface for product changeover.
Commissioned successfully and operating in production at one pack every two seconds, reliably detecting missing, incorrect, broken, and misplaced tablets across the full product range.
Handwriting recognition using Deep OCR
Evaluation and development of a deep learning OCR model capable of interpreting human handwriting using MVTec Halcon's Deep Learning Tool. The work explored the capabilities of the toolset with a real handwriting dataset, demonstrating the viability of the approach for applications where character variability and inconsistent presentation make traditional OCR unreliable.
Measurement of fibres
Design, development, and commissioning of a vision system to automatically measure the length and diameter of industrial reinforcing fibres presented randomly on a conveyor. Fibres are segmented and measured individually even when touching or crossing — with length determined by skeletonising each fibre's actual path to accurately capture curves and bends, and diameter sampled at equal intervals along the length. The system processes entire batch samples continuously, updating a real-time distribution display against defined specification thresholds.
Commissioned successfully and operating in production, significantly reducing manual preparation effort while enabling accurate, large-scale fibre measurement with minimal operator involvement.
These examples reflect the breadth of environments and constraints we have worked within — from high-speed production lines to complex, non-standard inspection challenges.
Across these projects, Oculus Vision has taken responsibility for the design, development, and commissioning of machine vision systems, while working alongside wider project teams to reduce delivery risk, support critical technical decisions, and ensure reliable performance under real operating conditions.