Jessica Rowbury reports from the Image Sensors 2015 conference, which took place in London earlier in the month, on the need for 4K imaging in security
The demand for higher resolution in the security sector is creating a challenge for camera manufacturers where the size and cost of cameras have to be taken into consideration. The comments were made by Dr Anders Johannesson, senior expert engineer at surveillance video firm Axis Communications, at the Image Sensors 2015 conference in London, which took place 18-20 March.
A major challenge Johannesson noted with imaging in the security sector is providing enough resolution in both the temporal and spatial domain.
In the last ten years, resolutions have increased from the standard 4CIF video up to HDTV, which almost all security cameras now employ. Now though, 4K video cameras have started to be introduced to the market. These imagers deliver 8 megapixel resolutions at speeds of at least 30 frames per second.
But as the size of sensors has decreased over recent years in order to make cameras smaller and cheaper, it makes it more difficult to provide higher resolutions. ‘The size of the sensors that we’re using is basically centred on about a 1/3-inch,’ Johannesson commented. ‘What that means is we can no longer use nice 7 x 7µm pixels; we have to go down in pixel size, and it’s basically driven by size limitations of the camera, and cost.’
Cameras with 4K resolution are beneficial for recognising several different faces at once, for example at football matches. ‘To recognise faces, we always count pixels form one ear to the other. The police say that you need to count 100 pixels from ear to ear to actually do something with the image,’ Johannesson explained. ‘So, if you really want to see everyone at a football game, for example, you’d have to sell a lot of [HDTV] cameras, which is probably not very inviting to customers. But in a 720 feed at 4K resolution, you can see quite a lot.’
However, it creates a challenge for camera manufacturers to keep up with demand for such high resolutions, especially when people want smaller and smaller devices. ‘It’s not easy to go up to 4K resolution, because we are using much smaller pixels, and there are lots of different defects in the optics and so on.’
Another limiting factor that Johannesson has noticed has been getting enough signal-to-noise ratio (SNR) and signal-to-architect ratio (SAR) in low light conditions – SAR is other defects, such as filter defects, optical defects, and shading issues. ‘We really want to get an image free of those limiting factors [SAR] in low light,’ he said.
Johanesson continued: ‘With security, there could be lots of things in the scene that complicate the image. You have to have good SNR and good SAR to take good images in complex scenes.’
Another technology being used by security cameras operating in different levels of illumination is capturing multiple exposures to get a better image, Johanesson noted.
Related articles:
Further information:
