Taking a picture from a smartphone is pretty much hit and miss however the technology has been improving all the time, as witnessed by the excellent camera built into the Nokia Lumia 920.
Things are looking better as researchers at MIT Microsystems Technology Laboratory have created a new processor chip. This new chip is said to enhance images within milliseconds and reportedly uses less power than the image processing software installed on many devices.
According to the group, the chip works by splitting the photographs into a matrix of small blocks, known as a ‘bilateral grid’. A histogram is then created for each block in which the block’s X and Y axis represent the location within the photo as a whole. This is then combined with another histogram for that same block, featuring brightness levels.
The chip can create a High Dynamic Range (HDR) image. HDR can expose dark and bright sections in a photograph perfectly without one area being too ‘dark’ or ‘bright’. The chip records three ordinary photographs of each shot – one normally exposed, one overexposed and one underexposed. The three images are then merged to create the HDR photo. The chip is able to process this within a few hundred milliseconds for a 10 megapixel photo meaning there is no noticeable delay.
Researchers say it can also handle a similar process for dark environments, taking a photograph with flash, and then one without and merging both photographs
The researchers are saying that the new chip uses very little power for the processing. The chip can also clean up noise in photographs, blurring ‘undesired’ pixels into the pixels closest. The blurring won’t take place when pixels closeby have significantly different brightness level. This means the algorithm should not affect main image areas.
No information yet on when this chip might be incorporated into mainstream cameras.
More information is available over here.
Kitguru says: Will this chip revolutionise smartphone camera technology in the future?