Fractional calculus filter of digital images of high-precision computation

Abstract

The invention provides a fractional calculus filter of digital images of high-precision computation, which is a circuit device for fractionally enhancing or smoothing complex texture detail characteristics of digital images at high precision. The filter is formed by connecting a RGB-to-HSI converter, a line memory group, a phase locking / shift circuit group, a fractional calculus mask convolution circuit, a maximum comparator and an HSI-to-RGB converter in cascade. Special fractional calculus mask convolution algorithms are used by a first algorithm unit circuit to an eighth algorithm unit circuit in the fractional calculus mask convolution circuit to realize high-precision computation of fractional calculus. The fractional calculus filter of digital images of high-precision computation, which is provided by the invention, is especially suitable for application occasions for fractionally enhancing or smoothing complex texture detail characteristics of high-definition digital televisions, biomedical images, bank bills, satellite remote sensing images, biological characteristic patterns and the like at high precision.

Description

[0001] Field [0002] The fractional-order calculus filter of the digital image with high-precision calculation proposed by the present invention is a circuit device for high-precision fractional-order enhancement or smoothing of complex texture detail features of the digital image. The order of the fractional calculus involved in the present invention is not a traditional integer order, but a non-integer order, and fractions or rational decimals are generally used in engineering applications. See figure 1 , the fractional-order calculus filter of the digital image calculated with high precision adopts RGB to HSI converter 9, line memory group 10, phase-lock / shift circuit group 11, fractional-order calculus mask convolution circuit 12, maximum The value comparator 13 and the HSI-to-RGB converter 14 are configured in cascade. The first arithmetic unit circuit 1 to the eighth arithmetic unit circuit 8 in the fractional calculus mask convolution circuit 12 adopt a unique fraction...

AI Technical Summary

Problems solved by technology

Further research by the applicant of the present invention shows that although the fractional order differential filter of the above-mentioned digital image can preserve the low-frequency contour features in the smooth area of ​​the image as much as possible, it can also enhance the gray value in the image by fractional order, nonlinearity, and multi-scale High-frequency edge features with relatively large jump range, and can also enhance the significant advantages of high-frequency texture detail features with relatively small gray value jump range and frequency changes in fractional, nonlinear, and multi-scale enhancement images. There are following three major shortcomings to be further improved and improved: 1. the fractional order differential filter of this digital image can only complete the fractional order differential operation, and cannot utilize the same circuit structure to complete the fractional order integral operation

Therefore, this filter can only perform fractional enhancement processing on complex texture detail features of digital images, but cannot perform fractional smoothing processing; II. At the expense of the calculation accuracy, its relative error is relatively large, so it is not suitable for applications requiring high-precision calculation of fractional differential; III. Because when the order of fractional differential is large, it is easy to destroy the digital color image R, G, B The correlation between the three color components, so when the fractional differential order is large, the filter is prone to color distortion in the processing results of digital color images

Images