Up-sampling filtering method for ultrasonic signal processing

Abstract

The invention discloses an up-sampling filtering method for ultrasonic signal processing. Resources used by a filter are reduced by a method of regulating filtering coefficients to be exactly divisible for 2^n, wherein n is as great as possible; the number of multipliers as well as system resource occupation is reduced by utilizing the characteristic of symmetry of the filtering coefficients; themultiplication of any filtering coefficients is realized in ways of shift addition and addition coefficient matching; and the resources are reduced by the method of multiplexing an adder by the same parts of different filtering coefficients. By the method, the to-be-consumed resources of the up-sampling filter can be reduced to be fewer than those originally to be consumed; control can be simply realized by the three methods of shifting, addition and delay; and the bit width of the adder is reduced, so the master working efficiency of a system is remarkably increased, and the stability and processing capability are both improved effectively.

Description

technical field [0001] The invention relates to the field of ultrasonic signal processing, in particular to an upsampling filtering method. Background technique [0002] In data processing, in order to increase the sampling rate and improve the performance of the system, upsampling is usually performed. Therefore, it is necessary to improve the performance of upsampling and reduce the resources for its implementation, especially in multi-channel systems. If each All channels must be upsampled, and if the design and implementation of upsampling are unreasonable, it will have a relatively large adverse impact on the system. [0003] Let's take the upsampling of a multi-channel system as an example to illustrate: [0004] The schematic diagram of the multi-channel processing part is shown in figure 1 As shown, the data of a certain channel needs to be processed after AD conversion, serial-to-parallel conversion, and before summation. The previous processing of delay summatio...

AI Technical Summary

Problems solved by technology

[0012] It can be seen from this that the 4-point filter requires at least 4 multipliers and 3 adders. If there are 64 channels in the device, 256 multipliers and 192 adders are required, and the ultrasonic echo performs AD conversion The latter is at least 12bit, and each channel of the echo data is continuously input at a rate of 12bit above 10M, which is a large consumption for the carry chain of the adder and the bit width of the multiplier. This resource consumption It is a big challenge for current hardware devices

[0013] Another way to achieve upsampling is: in the figure 1 The place where the cache output is shown, that is, the dotted line in the figure, is upsampled. This method directly increases the data rate of the cache output by the upsampling rate, and the control is more complicated, resulting in a large number of control resources.

[0014] The coefficients of the above two common processing methods are completely fixed. This processing method is less flexible and has a limited scope of application.

The existing technical control of upsampling interpolation is relatively complicated. In the case of many channels, the resources required will be large, which will cause a relatively large adverse impact on the system. Due to hardware limitations, the existing technology results in poor upsampling filtering effect

[0015] However, the existing algorithms with variable coefficients, large orders, and better upsampling and filtering effects cannot be realized in ordinary ultrasound systems due to the consumption of huge hardware multipliers and other resources. Even if they are realized, the cost performance is very low

Images