Research On Blind Convolution Fuzzy Image Processing Based On Cortex-A9 Processor

Fuzzy image restoration is an important branch of digital image processing,which is widely used in military industry,medical diagnosis,traffic monitoring and other industries as well as in daily activities and learning.In recent years,fuzzy image restoration has been paid more and more attention by scholars.However,most of the algorithms used have many defects,such as complex calculation process,low efficiency of the algorithm and poor effect of processing large size point diffusion function(PSF),and some algorithms will have serious ringing phenomenon when restoring images.Therefore,there is still a big gap between the image restoration technology and the expected effect.As the application of fuzzy image restoration becomes more and more extensive,the significance of studying fuzzy image restoration becomes more and more important.Generally,many factors may lead to fuzzy images,such as the tiny camera jitter or the relative motion between the object and the camera.This paper takes the fuzzy image which is taken by the embedded mobile device as the research object and also discusses the effective method of fuzzy image processing.This paper mainly does the following work:1.Based on the real-time,stability,reliability and other factors of image processing,this paper adopts the built-in ARM Mali-400 dual-core GPU Cortex-A9 microprocessor to build an experimental platform for image processing.U-boot,Linux kernel and root file system are ported in the setup platform.CMOS camera driver,software design processing and bootloader processing are analyzed in Linuxoperating system.CMOS camera driver,MJPG_streamer server and Linux interface function are used to realize the function of single frame image acquisition and processing.2.The basic knowledge system of digital image processing is introduced,and the previous researchers' research on fuzzy image restoration is deeply analyzed.In the process of iterative estimation of PSF,conjugate gradient method was used to optimize the energy equation and the first and second derivatives of the energy equation were used to improve the convergence rate of the algorithm.In addition,hysteresis threshold can also be used to suppress PSF noise.The algorithm can recover the clear edge and texture of the image,and has good suppression effect on ringing effect and image noise.3.This paper devotes to improve the rate of blur image restoration.After analyzing and studying the blind deconvolution algorithm based on Gaussian distribution of image gradient and sparse distribution of image gradient,this paper introduces the theory of two-dimensional wavelet transform and proposes the wavelet pyramid model to reconstruct the fuzzy image,and adopts the fast restoration method from low resolution to high resolution.Based on this model,a fast fuzzy image restoration algorithm suitable for Cortex-A9 processor is designed.Finally,the experimental results show that this algorithm is better and more efficient than the traditional fuzzy image restoration algorithm.Moreover,it can quickly and stably restore the clear edge and texture of the image and inhibit the occurrence of ringing phenomenon.