Research On Stereoscopic Image Synthesis Method Based On MapReduce Model

In the mobile Internet era,the number of information is rapidly expanding,the information on the Internet is increasing at the speed of geometric ratio every day.The information includes text,image and audio,etc.Among them,the image information is one of the most important multimedia information,it is larger than the scale of the text information and more complex.Three-dimensional image processing is an important issue in the field of computer studies,related industry is also very popular.The stereo image processing mainly includes the stereoscopic image synthesis and stereo image display.Lots of researchers have put effort into the algorithms of fulfilling above jobs.Although many effective and efficient algorithms have been put forward,they cannot be applied to data with huge volume due to the limits of single processing machine.As a result,the large amount of data has put forward new requirements for the realization of the stereo image synthesis algorithm.Through the analysis of the characteristics of Hadoop platform,the image file is not suitable to be directly processed by Hadoop platform.In order to solve this problem,it is necessary to pre process the image file.According to the different methods of image file preprocessing,this paper presents three kinds of distributed implementation method of stereo image synthesis algorithms.The first implementation method uses the HIPI framework to merge and cut the image file for making the image file suitable for Hadoop platform processing.The second implementation method stores the file name in the text block.Text block is a format suitable for the processing of Hadoop platform.Hadoop platform reads the corresponding image file by the file name.The third implementation method uses the way of video storage.Video file is a collection of a large number of image files to avoid the problem that image files is not suitable for Hadoop platform.In order to improve the synthesis effect of the stereo image synthesis algorithm,this paper introduces a method of adaptive disparity adjustment to the stereo synthesis algorithm.Firstly,the average disparity of the left and right view images is calculated by the image matching algorithm.After comparing with the optimal disparity value,the disparity adjustment value is obtained to make the corresponding adjustment.The experimental results show that compared with the single processing,the three kinds of distributed implementation of stereo image synthesis algorithm’s running time is greatly shortened.The algorithm efficiency is improved in different degree.And the effect of the subjective evaluation and the crosstalk of the stereo images are improved by the adaptive adjustment of the disparity.