Image by Taras Shypka
Simera-Africa---White.png

Modular Machine Vision Solutions

 

Machine Vision technology relies on cameras with specialised sensors, filters, and optics to acquire images for processing by dedicated hardware and software to analyse and measure various decision-making characteristics. Applications include but are not limited to automatic inspection, biomedical inspection, process control and robot guidance.

 

Simera Africa developed a modular machine vision architecture, whereby commercial machine vision cameras and lenses from industry-leading suppliers are combined with embedded processors and custom packaging to satisfy your specific requirements. We rapidly integrate leading technologies in new ways and apply them to solve real-world problems. 

We use leading image processors such as NVidia Jetson, providing the capability to implement onboard machine learning algorithms. These may be trained (for instance) to identify and track the specific objects in the images in real-time. Images or video frames are timestamped and geo-located for later integration with Geographic Information Systems, where required. Based on this standardised architecture, Simera Africa can supply machine vision solutions to satisfy a broad range of requirements.

Examples of previously delivered systems include: 

  • Hardware to explore underwater vision enhancement in murky water, based on spectral and polarisation characteristics of the water. This system consisted of co-located monochrome and polarisation cameras, with changeable spectral filters, packaged in waterproof enclosures and mounted on a structure that can be manipulated by a diver. 

  • Hardware to aid identification and management of invasive plant species. Using an array of colour cameras, time-stamped and geo-located videos and high-resolution still images provide analysts with recordings of survey observations, for later detailed analysis and processing. 

  • Stereo-Vision Aircraft Approach Tracker: Two Vision Units (each based on our standardised architecture and displaced at a set distance from each other) are set up next to a runway. By simultaneous observation of an aircraft on approach, stereo-vision techniques can be employed to determine the distance and range to the aircraft. This can be used to reconstruct the aircraft’s actual approach path and compare it to the ideal. Intended to be used in conjunction with the TADPOLE system, this will be a valuable training aid for pilots of high-performance aircraft. This system is currently under development. ​This project is executed with funding obtained from the Department of Trade & Industry South Africa (DTIC) through the Aerospace Industry Support Initiative (AISI), hosted and managed by the Council for Scientific and Industrial Research (CSIR).