Library graphic and text information denoising filter based on fractional order calculating equation

Abstract

The invention provides a library graphic and text information denoising filter based on a fractional order calculating equation and belongs to the technical field of interdiscipline of applied mathematics, digital image processing and digital circuits. On the basis of a special fractional order thermal conduction equation, fractional order, nonlinear, multi-scale and fast denoising of images is achieved. The related fractional calculus order v1 is not a traditional integer order but a non-integer order which is used in form of fraction or rational fraction in engineering application. The filter comprises a first differentiator, a second differentiator, a third differentiator, a model calculator, a first divider, a second divider, a fourth differentiator, a fifth differentiator, a first adder, a first multiplier, lambda n generator, a second multiplier, a second adder, a third multiplier, a fourth multiplier and a third adder which are in cascade connection. The filter is especially applicable to fast denoising of images rich in complex texture detail features.

Description

technical field [0001] A library graphic information denoising filter based on a fractional calculation equation proposed by the present invention is based on a special fractional heat conduction equation denoising algorithm to achieve fractional, nonlinear, multi-scale, and fast denoising of images. noise. The order v of fractional calculus involved in the present invention 1 It is not the traditional integer order, but a non-integer order. In engineering applications, fractions or rational decimals are generally taken. See figure 1 , the filter is Differentiator 2, differentiator 3, Differentiator 4, modulus finder 6, divider one 8, divider two 9, differentiator 11, Differentiator 12, Adder 15, Multiplier 16, λ n Generator 5, multiplier two 7, adder two 10, multiplier three 13, multiplier four 14 and adder three 17 are formed in a cascaded manner. This filter is especially suitable for fast denoising of images rich in complex texture details. The invention bel...

AI Technical Summary

Problems solved by technology

It is difficult to deal with some nonlinear, non-causal, non-minimum phase systems, non-Gaussian, non-stationary, non-integer dimensional (fractal) signals and non-white additive noise, etc.

If we directly apply the image denoising method based on integer-order partial differential equations to texture image denoising, there will be the following limitations: First, integer-order partial differential equations only process images according to local information, so they cannot maintain periodicity Texture features cannot restore the global features of the image; second, the integer-order partial differential equation model based on the integer-order variational principle realizes image processing by optimizing the energy functional

Since the integer-order differential value of the constant or DC component is zero, and the integer-order differential value of the slightly fluctuating AC component passes through a specific threshold current limit, its value is also zero, so the integer-order differential operation will cause extreme damage to the complex texture details of the image. big loss

When the traditional image denoising method based on integer-order partial differential equations is directly applied to texture image denoising, under low-resolution conditions, since the texture details in the original image are not rich and clear enough, the processing results can still Barely accepted; however, when the resolution gradually increases, the complex texture details in the processed image will be greatly lost, making it difficult for us to accurately interpret the texture details of the processed image

Images