Issue link: https://resources.mouser.com/i/1437726
21 INTEL 2021 f aster and better-performing machine vision is a mainstay of moves toward next-generation industrial automation, driverless cars and the managing of smart cities. Better image quality, more rapid image capture, plus lower equipment cost and complexity are key objectives for designers of automated equipment, inspection systems and robotics, with the aims of strengthening quality assurance and increasing productivity. Similarly, advanced machine vision will be pivotal to autonomous vehicles, to enable them to recognize signage, road markings and potential hazards ahead. Here, the emphasis is going to be on reducing system response times and improving image-recognition accuracy. As far as smart city applications are concerned, improved image clarity from city centre CCTV can help law enforcement agencies protect citizens by anticipating disturbances and improving identification capabilities. In addition, new opportunities for high-performance machine vision are starting to emerge. These include onboard drones to assist with vehicle guidance and also for data-gathering purposes (like surveying agricultural land or architectural sites). To extract more information from captured images within shorter timeframes requires better-quality images and enhanced signal-processing performance. To this end, some important innovations are coming forward that apply to cameras and image sensors, and also – thanks to commercialization of machine- learning technologies – the image processing techniques employed. Next-Generation Lenses Enhance Focus and Vision At the front end of the entire system, camera lenses are the subject of some powerful new engineering advances that can improve flexibility, reduce cycle times and streamline equipment design by performing multiple tasks with a single camera or single lens. Among these, liquid lenses (as shown in Figure 1) are an emerging class of optics that extend the depth of field of a conventional lens without the cost and complexity of a traditional motorized focusing system. Motorized focusing is also relatively slow, and so avoiding this will help to increase cycle time in applications – such as industrial inspection involving objects positioned at various distances. Introducing a liquid lens enables a standard optical system to adjust focus on the fly within a couple of ms – from infinity to less than 100mm, depending on spacing. The liquid lens contains an optical liquid sealed inside a flexible membrane. Changing the lens radius by only a few µm, by moving the membrane or adjusting the volume of optical liquid, has an effect comparable to moving the lens a few cm using a conventional motorized focusing system. In addition to faster focusing and simpler construction (with fewer moving parts and thus less chance of operational failure), liquid lens systems also benefit from lower inertia and reduced power consumption. Figure 1: A liquid lens allows the focal point to be adjusted with just a few µm change in shape.