Image Gaussian filtering method with dynamically configured mode and realization device thereof

Abstract

The invention relates to an image Gaussian filtering method with a dynamically configured mode and a realization device thereof. A field programmable gate array dynamically configures filtering modes according to different requirements; under the corresponding filtering mode, a two-dimensional image Gaussian filtering process is decomposed into two one-dimensional image Gaussian filtering processes by use of the separability of a Gaussian convolution kernel; the Gaussian filtering in one-dimensional direction is firstly carried out; the Gaussian filtering data is cached in a memory in FPGA (Field Programmable Gate Array); the Gaussian filtering in the other dimensional direction is carried out; and the Gaussian filtering in the dimensional direction is subjected to shift processing to obtain a final Gaussian filtering image in the filtering mode. The image Gaussian filtering method disclosed by the invention has a flexible use mode, the development difficulty is reduced, the developing period is shortened, the Gaussian filtering with large size can be effectively realized in a project, and the practicality of the Gaussian filtering in the project is enhanced.

Description

technical field [0001] The invention relates to a technology capable of realizing image Gaussian filtering with dynamically configurable modes, in particular to an image Gaussian filtering method with dynamically configurable modes and a device for realizing it. Background technique [0002] Image filtering technology is always an important technology in the field of image processing research, and it has important application value in specific engineering applications. The image containing noise is no longer what we want, but randomly contains a lot of local high-frequency information in the image, which brings many problems to our subsequent processing and application. The most common type of noise in images is Gaussian noise, that is, noise with a mean of zero and a normal distribution. Linear smoothing filters work well for removing Gaussian noise, and in most cases also work well for removing other types of noise. Linear filtering uses the weighted sum of pixels in a c...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the deficiencies in the prior art of image Gaussian filtering methods such as inflexible configuration methods, the technical problem to be solved by the present invention is to provide a low-cost, high-flexibility image Gaussian filtering method and an implementation device with dynamically configurable modes

Images