April 2021

MACHINE VISION n Making machine vision count during Covid Various machine vision technologies have been redeployed during the Covid pandemic to carry out new duties such as tracking people and ensuring social distancing. Ian Alderton, technical sales director of Alrad Instruments, looks at some of the technologies involved. Since the outbreak of the Covid pandemic, various thermal, visible light and lidar technologies have been used to count people in buildings and to ensure social distancing. Cameras can detect people’s faces and thus count them as they enter and exit a building. These systems not only control entrances, but also register how many people are leaving a building. For example, without needing extra staff, they can help to ensure that the number of people allowed in a shop is not exceeded. If too many people are detected, red“traffic light”signals can be generated and door controls implemented automatically. The data can also be used to inform customers via screens about waiting times. If thermal cameras are used, you can also measure body temperatures and trigger I/Os, thus preventing people with a fever from entering buildings. Another option is to use 3D lidar (light detection and ranging) motion sensors. A ceiling- mounted sensor sends out rays of laser light and tracks the time it takes for the light to bounce back, painting a rough picture of objects and people in the area. By calculating the“time of flight”, you can determine the distance and the shape of any humans, which the software can recognise. Paired with machine learning analysis, such systems can count how many people step over a virtual line, or pass a specific point, with 99% accuracy. Lidar systems of this sort were originally developed to understand how people move through buildings by tracking where they linger and creating“heatmaps”. But the technology is also proving a useful tool for ensuring social distancing. Lidar cannot identify an individual, so there’s built-in privacy protection. Unlike with camera tracking, no images are captured – just a rough outline of a person, essentially as a set of black- and-white dots. The data collected is not a privacy risk and you do not need to put the layers of security that you might need for GDPR using traditional cameras. JJSMANUFACTURING .COM OUTSOURCED MANUFACTURING MADE EASY ELECTRO MECHANICAL PCB ASSEMBLY BOX BUILD ASSEMBLY CABINET BUILD OVER 150,000 SQ. FEET OF CONNECTED FACILITIES IN THE UK & CZECH REPUBLIC