Super-resolution image restoration method based on hardware platform

Abstract

The invention discloses a super-resolution image restoration method realized on the basis of a hardware platform. The method is realized on the basis of the Zynq SOC hardware platform, and employs an idea of parallel processing. An original image is developed; the developed image is cached into a buffer memory buff, the buffer memory buff is realized via a BlockRAM in the SoC, a series of operations in an image restoration algorithm are carried out while caching, and operation results are cached into the corresponding buffer memory buff at the same time. The operation steps in the algorithm are executed in parallel instead of in order, intermediate values of the operations are cached while parallel operation, the executing efficiency of the algorithm is improved, only memory resources, as hardware resources, in a PL part of a Zynq chip are consumed, consumption of hardware memory resources is reduced greatly, and the method is helpful for research and development of small, low-power-consumption and low-cost equipment.

Description

technical field [0001] The invention relates to a super-resolution image restoration technology, in particular to a super-resolution image restoration technology based on a hardware platform. Background technique [0002] At present, various imaging devices are widely used in various fields, but due to various conditions, the imaging quality is often unsatisfactory, so people are eager to find a way to improve the image quality. Since the optical system is affected by the diffraction of the optical system, image information beyond the cutoff frequency of the transfer function cannot be acquired. How to observe image information beyond the cutoff frequency has become a research boom. [0003] At present, the super-resolution image restoration methods mainly include: spectral extrapolation method, POCS method, neural network restoration method and Bayes analysis method, etc. Among them, the algorithm with strong super-resolution restoration ability is the Poisson-ML image re...

AI Technical Summary

Problems solved by technology

However, due to the need to store a large number of intermediate calculation values ​​in the process of multiple iterations, it is a big test for hardware storage resources.

Images