Real Time Features Will Expand The Market for Machine Vision

Real-Time Machine vision software is bound to have a ground-breaking impact on the industrial and manufacturing sectors in 2020 and beyond. Most vision systems in machine control applications today do not rely on a real-time feature to include vision data as feedback to guide control of the machine. Machine vision usually augments an operation such as supporting an inspection process to verify that a part was made correctly. However, leading edge machine builders are starting to use vision-directed motion to guide the machine to perform better and to produce higher quality parts at faster throughput rates.

These leading edge machine builders are enhancing machine vision with determinism (or hard real time) so that it can applied in machine automation setting rather than just limited to the brick and mortar facilities, inspection or other settings that do not required an immediate response of less than a few milliseconds.

The expanding industry 4.0 systems heavily depend on real time machine vision enabled devices to feed precise directions so that soft motion machine controllers can increase performance, quality and throughput. This will only be another source of data that can be forwarded to the cloud in order for artificial intelligence to provide more accurate actionable insights to improve overall machine health and performance. Here are some of the machine vision trends to watch out for in 2020. Further, new lens and camera technologies will support the demands of a real-time, deterministic machine vision system.

Vision guided robotics and metrology

The demand for guidance systems, faster, more intelligent 3D measurement is required to grow as vision guided robotics advance and gain more mobility. This is because, 3D imaging algorithms are required to process all images effectively and in real-time. This entails measuring the distance between points to ensure that the moving robots know where they are in the 3D space. 3D image systems are designed for use in applications such as bin picking of random identical and oriented objects. Therefore, they must have improved accuracy when applied in picking heterogeneous and unknown objects like boxes and parcels. This has triggered the need for additional features like faster 3D reconstruction of object features in a bid to enhance automation production and speed.

Embedded vision

Embedded vision combines image processing and capturing capabilities into one device. This is a real time machine vision technology that has been embraced by numerous industries that utilize the sorting and inspection systems. Meaning that, embedded vision is gaining the recognition it deserves in the manufacturing industries. As a result it is expected that it will be applied in more automation systems as pre-programmable vision processing applications facilitate industries to implement deep learning and artificial intelligence solutions. It is also expected that more embedded vision products will be designed for application in specific commercial tasks that require simplified plug and play real-time solutions.

Importantly, real-time machine vision does not stand alone. Using off-the-shelf cameras with off the shelf cables that are all standardized on the Ethernet enabled GigE Standard means that a $2000 setup on an open Machine Automation software platform can replace a camera that costs $50,000 or more. This doesn’t even include the integration costs because the all-software solution can run on an IPC and is easily integrated.

Liquid lenses

Component lens suppliers are required to provide better optics as the camera pixel sizes decrease and resolutions increase. The lens performance is expected to improve in quality and speed. Liquid lenses are preferred because they can adjust focus instantly by changing the voltage or current as opposed to making mechanical changes as is the case with conventional lenses. In a bid to offer better quality images, the liquid lenses have advanced from the smart cameras and sensors to complex industrial applications. As such, they can offer better vision where quick focus is necessary and in applications with varied image distances. As such, it is expected that more component cameras with embedded processing capabilities that control the liquid lenses automatically will be designed with the aim of boosting flexibility.


Because of heavy use in non-deterministic settings, cameras for machine vision are dropping and this creates a new opportunity to create a breakthrough in machine vision applications for machine controllers. This coupled with open standards allows machine builders to create a real-time vision system for a fraction of the cost of a proprietary alternative Leading edge machine builders are starting to use vision-directed motion to guide the machine to perform better and to produce higher quality parts at faster throughput rates.

The machine builders that adopt real-time machine vision will have a tremendous first mover and price/performance advantage of competitors who don’t.