Research And Implementation Of Image Scaling Algorithm Based On Polyphase Filtering

With the development of digital television and people's visual need to gradually improve the situation,the video image quality to enhance the resolution becomes more rapid.Now,in the 4K TV is not a large area of occupation of the market,8K TV has achieved commercial in each big brand,and gradually appeared in the mass family.As the core function of image zoom image resolution,in video processing plays a crucial role.Therefore,this research has certain economic and scientific research value.This paper mainly through the two aspects to carry out research work on image zoom.On the one hand,the first is the image zoom principle is analyzed theoretically.Secondly,compared to some classical algorithms of image zoom,found that the classical algorithm is realized by using linear interpolation method.Classical interpolation algorithms: nearest neighbor interpolation,bilinear interpolation,bicubic interpolation etc..These algorithms have their own advantages,but there are some obvious limitations: Although the algorithm is simple,easy to implement,but it will cause the blurred edges and jagged edges or even other phenomena;some algorithm of zoom effect is good,but the computational complexity is too high,is not conducive to hardware implementation.In view of the shortcomings of the traditional method,comprehensive consideration,in recent years the emergence of a new nonlinear filtering algorithm: zoom.Image zoom algorithm proposed in this paper is a zoom algorithm based on polyphase filtering is a nonlinear filtering algorithm commonly used in the industry.This paper discusses the algorithm theoretical basis,ringing the ideal low-pass characteristics and the polyphase filter may cause are analyzed,and finally put forward a optimization algorithm to be improved.Next,the simulation of this algorithm by using the MATLAB software,and through the example of zoom effect chart compares it with the classical algorithm,the intuitive conclusion: this algorithm compared with classical algorithm has obvious advantages,zoom effect is good,the hardware implementation is not complicated.On the other hand,the optimized algorithm based on this,the paper presents an implementation scheme of hardware architecture.The first step to clarify the overall architecture of image zoom module.Then,according to the difference of internal function,the module is divided into several sub modules.Then,this paper introduces the data flow of the whole process of zoom,and a detailed description of the configuration module,color conversion module,the hardware structure of several important modules such as zoom module and Realization method.Especially in the realization of zoom module,using the human eye to brightness change sensitive characteristics,YUV pixel components with different zoom filter,reduces the computational complexity,saving hardware resources.Finally,using the Modelsim function simulation of image zoom module,that is before the simulation,to verify the correctness of code and module function.And a detailed description of the whole validation protocol and validation process,waveform simulation results are given.Finally,through simulation,prove the module function is correct and the circuit design is correct,the image zoom module available.