Research on an improved MAP Super-Resolution in the Intelligent Transportation system
Haitang Chen, Guoyong Huang, Wuqiang Yang · 2015
An improved MAP super-resolution was proposed aiming to the problem that the image dealt by traditional super-resolution was quite smooth and the marginal information was vanished.This algorithm constructs a 2×high resolution image from a low resolution image by Curvelet interpolation.The constructed image is divided into four directions and the corresponding directivity factor was calculated.Then, interpolate image by combining the directivity factor and its weight coefficient.And then, the interpolated image was interpolated by Bicubic interpolation.And the interpolated image was restored by MAP super-resolution.The experiment shows that this algorithm can remain the marginality of image better and restrain the appearance of blur and serration within the area where the margin and texture are rich compared to the traditional super-resolution algorithm.