Study On The Frame Rate Up Conversion Algorithm And Implementation By Simulink

Large size LCD is prone to produce moiton blur in displaying dynamic images,while frame rate up conversion is an effective solution to this problem by increasing the LCD's refreshing rate.The simple algorithm directly uses each kind of frame combination to interpolate the middle frame without considering the object movement,while complex algorithms use motion compensation to insert the frame.Based on the research on the frame rate up conversion algorithm,this paper presented a novel frame rate up conversion algorithm with higher complexity but good quality of interpolaed frames.This algorithm was based on the segmentation of moving objects and background.It deals with the frame difference image by opening and closing morphological operators to obtain a segmentation mask,and this mask was regarded as the basis for further motion estimation and motion compensation.Then we set up the simulation module of floating-point simulation and fixed-point simulation for the algorithm.The simulation results were compared to uni-directional motion estimation and bi-directional motion estimation algorithms and it verified the superiority of the proposed algorithm by both the visual quality and PSNR.Finally,from the point of view of hardware implementation,the algorithm was improved to reduce the amount of calculation while maintaining the quality of interpolated frames.This paper presented a new fast matching block search algorithm,namely double trace three step searches,to replace the full search method in estimating the initial motion vector to improve the original algorithm.It could not only reduce the search points effectively but also ensure the accuracy of the search results.In addition,the segmentation based on frame difference could accurately segment the moving objects and background when the background was static,but it could not achieve the desired segmentation result if there are movements in the background.We removed the segmentation to improve algorithm and compared the simulation results of the unimproved and the improved algorithm.The comparison demonstrated that the improved algorithm got a banlance between quality of interpolation frame and complexity of the algorithm.