Research On Image Detail Enhancement Based On Fractional Differential

In image detail enhancement,in order to improve the visual effect of the image,it is desirable to enhance the high-frequency components of the image,thereby highlighting the texture information of the image,and further smoothing the low-frequency components of the image.Traditional image detail enhancement algorithms cannot solve the enhanced image well.Contradictions of image noise are removed while texture details are removed.Based on the theory of fractional differential calculus,this paper uses the advantage of weak derivative properties of fractional differential operators to apply it to the field of image detail enhancement.Fractional differentiation makes the enhanced image non-linear while highlighting the texture details.Preserve the smooth areas in the image,sharpen the image details,and achieve better visual effects.Fractional order differentiation is used significantly in image detail enhancement.First,the application of the fractional calculus theory in image processing is explained,and the development status of the fractional calculus in the field of image processing is introduced in detail.The basic theory of fractional calculus is analyzed.Several commonly used time domain definitions of fractional calculus are given.The classic fractional differential operator Tiansi for image detail enhancement is introduced.The Tiansi operator is relative to the integer order differential.Operators have better differential characteristics.Secondly,in view of the limitation that the Tiansi operator will treat the texture region as a smooth region at individual pixels and thus cannot achieve the enhancement effect,two improved operators are proposed.One is the improved Tiansi operator.On the one hand,the improved Tiansi operator still retains good differentiation characteristics of fractional order,on the other hand,in the smooth area of the image,the processed pixel value changes little,and in the texture area The processed pixels change more.The other is an improved fractional differential operator,which derives the Taylor expansion of the binary function in a certaindirection,and then combines the definition of the fractional calculus GL to derive the Taylor expansion of the fractional differential in a certain direction.An improved fractional differential operator.This improved fractional differential operator makes full use of the surrounding information of the pixels to be processed,overcomes the problem of under-enhancement of the Tiansi operator,and obtains a better image detail enhancement effect.Then,for the current status of the fractional differential operator requiring multiple experiments to determine the best differential order,combined with the local feature information of the image,a fractional order model with adaptive ability is constructed,saving the manual search for the best fractional order Order time,get more detailed information than the original image.Finally,on the basis of the above theory,several objective image quality evaluation indexes are introduced,including objective evaluation indexes such as information entropy,average gradient,image contrast.Image detail enhancement experiments were performed on the images of the haze weather and the night scene respectively.The analysis of the experimental results finally confirmed that the algorithm proposed in this paper has achieved good results in detail enhancement,and proved the effectiveness and applicability of the algorithm.