Noise reduction device, noise reduction method, and noise reduction program

Abstract

To reduce noise while suppressing an increase in the amplitude.SOLUTION: A noise reduction device includes: a first Hilbert transformer that performs Hilbert transformation on an input signal; a second Hilbert transformer that is connected in cascade with the first Hilbert transformer, performs Hilbert transformation on a first Hilbert transformed signal output from the first Hilbert transformer and has characteristics opposite to those of the first Hilbert transformer; a delay unit which delays the input signal based on the order of the first Hilbert transformer and the order of the second Hilbert transformer; and an adder which adds a delayed signal output from the delay unit and the second Hilbert transformed signal output from the second Hilbert transformer.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The disclosed technology relates to a noise reduction device, a noise reduction method, and a noise reduction program. [Background technology]

[0002] Patent Document 1 describes a signal processing method for receiving a time-series signal obtained by A / D conversion processing of an output signal of a sensor as an input signal, removing high-frequency noise components higher than a predetermined frequency (Fa) contained in the input signal, and removing low-frequency noise components equal to or lower than the predetermined frequency from a first processed signal (Y(k)) from which the high-frequency noise components have been removed, to obtain a direct-current component of the input signal. The process for removing the low-frequency noise components includes a step (S5) of amplifying the first processed signal with a fixed gain set in advance to compensate for a gain reduction of the low-frequency noise components to be removed that is predicted to occur due to a phase shift process, and a step (S6) of performing two stages of Hilbert transform processing on the amplified signal to generate a first phase-shifted signal (Ni) having a phase difference of 180 degrees with respect to the amplified signal, and a second phase-shifted signal (Nq) having a phase difference of 90 degrees with respect to the first phase-shifted signal. The disclosed signal processing method includes a step (S6), a step (S7) of calculating, based on the first phase-shifted signal and the second phase-shifted signal, an instantaneous amplitude, an instantaneous phase and an instantaneous frequency of the low-frequency noise component to be removed, which is included in the amplified signal, a step (S8) of detecting an error between the calculated instantaneous amplitude and the instantaneous amplitude of the low-frequency noise component to be removed, which is included in the first processed signal, a step (S9) of generating, based on the calculated instantaneous amplitude, instantaneous phase, instantaneous frequency and the error in the instantaneous amplitude, a first noise signal having the same amplitude and frequency as, and an inverted phase with, the low-frequency noise component to be removed, which is included in the first processed signal, and a step (S10) of adding the first processed signal and the first noise signal to obtain a second processed signal (Z(k)) in which the low-frequency noise component to be removed from the first processed signal is removed.

[0003] Patent Document 2 discloses a signal processing method including a step (S1) of converting a sensor output signal into a digital original signal sequence by oversampling, a step (S2) of detecting a signal sequence of AC components contained in the original signal sequence, a step (S3) of matching the phase and amplitude of the original signal sequence and the extracted signal sequence of AC components, a step (S4) of adding or subtracting both signal sequences whose phases and amplitudes have been matched to detect a signal sequence of DC components of the original signal sequence, a step (S5) of performing a thinning process on the signal sequence of DC components, and a step (S6) of performing a high-frequency cut-off process on the thinned signal sequence, in which the high-frequency cut-off processed signal sequence is used as a predicted convergence value of the sensor output signal.

[0004] Patent Document 3 discloses a signal processing device comprising an input unit for inputting an audio signal, an acquisition unit for acquiring the audio signal input to the input unit, a calculation unit for performing a Hilbert transform on the audio signal acquired by the acquisition unit to convert it into an analytic signal and generating harmonics by multiplying the analytic signal, and an addition unit for adding the signal after the calculation by the calculation unit to the audio signal input to the input unit and outputting the result. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2009-276268 A [Patent Document 2] JP 2005-274320 A [Patent Document 3] JP 2011-119892 A Summary of the Invention [Problem to be solved by the invention]

[0006] In mass measurement in a transport system or on a ship, there is a problem of how to measure the mass of an object with high accuracy. To do so, it is necessary to efficiently remove the vibration noise of the transport system.

[0007] Generally, in mass measurement of a transport system, the mass can be regarded as a DC component, and other noise components as AC components.

[0008] FIG. 12 shows a first Hilbert transformer H1 e , the second Hilbert transformer H2 e , two first Hilbert transformers H1 e A Hilbert transformer (H1 e +H1 e ), the first Hilbert transformer H1 e and the second Hilbert transformer H2 e A Hilbert transformer (H1 e +H2 e The amplitude characteristics of the simulation when the mass was measured on a ship for each of the configurations of the 1000-kV gyroscope are shown. The horizontal axis of Fig. 12 is normalized angular frequency, and the vertical axis is amplitude.

[0009] In addition, the Hilbert transformer H1 e The order N1 is 26, the second Hilbert transformer H2 e The order N2 of is 30. Therefore, the Hilbert transformer (H1 e +H1 e ) has a degree of 52, and the Hilbert transformer (H1 e +H2 e ) has an order of 56. The bandwidth is 0.15 to 0.85.

[0010] As shown in FIG. 12, the first Hilbert transformer H1 e can perform a 90-degree phase shift on a signal having an angular frequency band from 0.05 to 0.95. For this reason, the same first Hilbert transformer H1 e A Hilbert transformer (H1 e +H1 e ) can shift the phase of a signal by 180 degrees.

[0011] However, generally, when Hilbert transformers with the same characteristics are connected in cascade, as shown by "Cascade" in Figure 13, the stopband attenuation becomes larger and the noise reduction effect is improved compared to the characteristics of a single Hilbert transformer shown by "Single", but the amplitude of the passband through which the signal passes becomes larger, amplifying the signal.

[0012] Since an ideal Hilbert transformer converts all AC components into a 90-degree phase shift, the first Hilbert transformer H1 e A Hilbert transformer (H1 e +H1 e ), as shown in FIG. 12, a single first Hilbert transformer H1 e As a result, the amplitude of the first Hilbert transformer H1 e By cascading these, a single Hilbert transformer H1 e It is possible to change the phase of the AC component in by 180 degrees, but there is a problem in that the amplitude becomes large.

[0013] The disclosed technology has been developed in consideration of the above-mentioned facts, and aims to provide a noise reduction device, a noise reduction method, and a noise reduction program that can reduce noise while preventing amplitude from increasing. [Means for solving the problem]

[0014] In order to achieve the above object, a noise reduction device according to a first aspect includes a first Hilbert transformer that performs a Hilbert transform on an input signal, a second Hilbert transformer that is cascaded with the first Hilbert transformer and performs a Hilbert transform on a first Hilbert transform signal output from the first Hilbert transformer, the second Hilbert transformer having inverse characteristics to the first Hilbert transformer, a delay device that delays the input signal in accordance with the orders of the first Hilbert transformer and the second Hilbert transformer, and an adder that adds the delayed signal output from the delay device and the second Hilbert transform signal output from the second Hilbert transformer.

[0015] In the noise reduction device according to the first aspect, the order of the first Hilbert transformer and the order of the second Hilbert transformer may be odd numbers.

[0016] A noise reduction method according to a second aspect includes a computer executing processes including: converting an input signal into a first Hilbert transform signal by a first Hilbert transform process, converting the first Hilbert transform signal into a second Hilbert transform signal by a second Hilbert transform process having inverse characteristics to the first Hilbert transform process, generating a delayed signal by delaying the input signal in accordance with an order of the first Hilbert transform process and an order of the second Hilbert transform process, and adding the delayed signal and the second Hilbert transform signal.

[0017] A noise reduction program according to a third aspect causes a computer to execute processes including: converting an input signal into a first Hilbert transform signal by a first Hilbert transform processing; converting the first Hilbert transform signal into a second Hilbert transform signal by a second Hilbert transform processing having inverse characteristics to the first Hilbert transform processing; generating a delayed signal by delaying the input signal in accordance with an order of the first Hilbert transform processing and an order of the second Hilbert transform processing; and adding the delayed signal and the second Hilbert transform signal. Effect of the Invention

[0018] The disclosed technology has an effect of reducing noise while suppressing an increase in amplitude. [Brief description of the drawings]

[0019] [Figure 1] FIG. 1 is a diagram illustrating the configuration of a noise reduction device. [Diagram 2] FIG. 13 is a waveform diagram of an input signal and a signal that has been Hilbert transformed by an even-order Hilbert transformer in a simulation of mass measurement on a ship. [Diagram 3]FIG. 13 is a diagram showing the amplitude characteristics of an odd-order Hilbert transformer in a simulation of mass measurement on a ship. [Figure 4] FIG. 13 is a waveform diagram of an input signal and a signal that has been Hilbert transformed by a Hilbert transformer of odd order in a simulation of mass measurement on a ship. [Diagram 5] FIG. 11 is a waveform diagram of a signal that has been Hilbert transformed by an even-order Hilbert transformer and an odd-order Hilbert transformer when mass is measured on a ship. [Figure 6] FIG. 13 is a diagram showing the amplitude characteristics of an even-order Hilbert transformer in a simulation in which mass is measured on a conveyor. [Figure 7] FIG. 13 is a waveform diagram of an input signal and a signal obtained by Hilbert transforming using an even-order Hilbert transformer in a simulation in which mass is measured on a conveyor. [Figure 8] FIG. 13 is a diagram showing the amplitude characteristics of an odd-order Hilbert transformer in a simulation in which mass is measured on a conveyor. [Figure 9] FIG. 13 is a waveform diagram of an input signal and a signal obtained by Hilbert transforming using an odd-order Hilbert transformer in a simulation in which mass is measured on a conveyor. [Figure 10] FIG. 11 is a waveform diagram of a signal that has been Hilbert transformed by an even-order Hilbert transformer and an odd-order Hilbert transformer when mass is measured on a conveyor. [Figure 11] 13 is a flowchart of a Hilbert transform process. [Figure 12] FIG. 13 is a diagram showing the amplitude characteristics of an even-order Hilbert transformer in a simulation in which mass is measured on a ship. [Figure 13] FIG. 1 is a diagram showing the amplitude characteristics of a single Hilbert transformer and a Hilbert transformer in which Hilbert transformers are cascaded. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0021] Fig. 1 is a diagram showing the functional configuration of a noise reduction device 10. As shown in Fig. 1, the noise reduction device 10 includes a first Hilbert transformer H1, a second Hilbert transformer H2, a delay unit DL, and an adder AD.

[0022] The first Hilbert transformer H1 performs a first Hilbert transform on the input signal x(n) to generate a first Hilbert transform signal h1(n). The first Hilbert transformer H1 outputs the generated first Hilbert transform signal h1(n) to the second Hilbert transformer H2.

[0023] The second Hilbert transformer H2 is connected in cascade with the first Hilbert transformer H1. The second Hilbert transformer H2 performs a second Hilbert transform on the first Hilbert transform signal h1(n) output from the first Hilbert transformer H1 to generate a second Hilbert transform signal h2(n). Here, the amplitude characteristics of the second Hilbert transformer H2 are inverse to those of the first Hilbert transformer H1. The second Hilbert transform signal h2(n) generated by the second Hilbert transformer H2 is output to an adder AD.

[0024] The delay unit DL generates a delayed signal d(n) by delaying the input signal x(n) according to the order of the first Hilbert transformer H1 and the order of the second Hilbert transformer H2. The delayed signal d(n) generated by the delay unit DL is output to the adder AD.

[0025] The adder AD generates and outputs an output signal y(n) by adding the delayed signal d(n) output from the delay unit DL and the second Hilbert transform signal h2(n) output from the second Hilbert transformer H2.

[0026] When distinguishing between even and odd-order Hilbert transformers, the even-order Hilbert transformer is designated by adding an "e" to the end of its symbol, and the odd-order Hilbert transformer is designated by adding an "o" to the end of its symbol.

[0027] The first Hilbert transformer H1 of order 26 e , the second Hilbert transformer H2 of order 30 e In this case, the first Hilbert transformer H1 e A Hilbert transformer (H1) is a cascade connection of a first Hilbert transformer H1 and a second Hilbert transformer H2. e +H2 e ) has an amplitude characteristic as shown in FIG. 12, which makes it possible to reduce the amplitude and effectively reduce noise.

[0028] FIG. 2 also shows an input signal x(n) and a Hilbert transformer (H1 e +H1 e ) is used to transform the input signal x(n) into a Hilbert transform signal, and the first Hilbert transformer H1 e and the second Hilbert transformer H2 e A Hilbert transformer (H1 e +H2 e The figure shows the waveform obtained by simulating the signal that was Hilbert transformed by the Hilbert transformer (H1 e +H2 e ) is transformed into a Hilbert transform signal by the Hilbert transformer (H1 e +H1 e ) the amplitude is smaller than the signal that has been Hilbert transformed, and it can be seen that the noise has been reduced.

[0029] 12, a Hilbert transformer with an even order cannot pass a signal with a normalized angular frequency close to 1. Therefore, the orders of the first Hilbert transformer H1 and the second Hilbert transformer H2 may both be odd numbers.

[0030] FIG. 3 shows the results of a simulation of mass measurement on a ship using the first Hilbert transformer H1 o The order N1 is 27, and the second Hilbert transformer H2 o The amplitude characteristics are shown when the order N2 is set to 31. As shown in Figure 3, when the order of the Hilbert transformer is an odd number, it can pass signals with normalized angular frequencies close to 1, and the bandwidth is from 0.05 to 1.

[0031] Also, as shown in FIG. 3, a first Hilbert transformer H1 o A Hilbert transformer (H1 o +H1 o ), the amplitude becomes large, as in the case of an even order. On the other hand, in the case of a Hilbert transformer (H1) in which a first Hilbert transformer H1 with an odd order and a second Hilbert transformer H2 with an odd order are cascaded, o +H2 o ), the Hilbert transformer (H1 o +H1 o ), the amplitude is smaller, and it can be seen that noise is effectively reduced.

[0032] FIG. 4 also shows an input signal x(n) and the same first Hilbert transformer H1 o A Hilbert transformer (H1 o +H1 o ) and a signal obtained by Hilbert transforming the input signal x(n) using a first Hilbert transformer H1 with an odd order and a second Hilbert transformer H2 with an odd order. o A Hilbert transformer (H1 o +H2 o ) and the waveform obtained in a simulation of mass measurement on a ship. o +H2 o ) is transformed into a Hilbert transform signal by the Hilbert transformer (H1 o +H1 o) the amplitude is smaller than the signal that has been Hilbert transformed, and it can be seen that the noise has been reduced.

[0033] FIG. 5 also shows a first Hilbert transformer H1 e and a second Hilbert transformer H2 of even order. e A Hilbert transformer (H1 e +H2 e ) and a signal even obtained by Hilbert transforming the input signal x(n) using a first Hilbert transformer H1 o and a second Hilbert transformer H2 of odd order. o A Hilbert transformer (H1 o +H2 o ) and the waveform obtained in a simulation in which the input signal x(n) is Hilbert transformed and the mass is measured on a ship. As shown in Figure 5, the signal odd has a smaller amplitude than the signal even. This effect becomes more pronounced as the input signal x(n) contains more high-frequency components.

[0034] Furthermore, as is clear from FIG. 1, when adding and outputting an input signal x(n) and a signal obtained by converting the input signal x(n) using the first Hilbert transformer H1 and the second Hilbert transformer H2, it is necessary to delay the input signal x(n) by an amount equal to 1 / 2 the total order N (=N1+N2) of the Hilbert transformers H1+H2, and add it to the signal converted by the first Hilbert transformer H1 and the second Hilbert transformer H2.

[0035] This requires a delay element to delay the signal by N / 2, but a single Hilbert transformer with order N being an odd number is not normally used because N / 2 is not an integer.

[0036] In contrast, the first Hilbert transformer H1 o and the second Hilbert transformer H2 oBy cascading these, half of the total order N becomes an integer, so that a Hilbert transformer with an odd order can be used.

[0037] 2 to 5, 12, and 13 show the results of a simulation in which mass was measured on a ship, while FIGS. 6 to 10 show the results of a simulation in which mass was measured on a transport system, specifically, on a conveyor.

[0038] As shown in FIG. 6, the simulation result of mass measurement on a conveyor is similar to the simulation result of mass measurement on a ship shown in FIG. 12, and the first Hilbert transformer H1 e and the second Hilbert transformer of even order H2 e A Hilbert transformer (H1 e +H2 e ), the Hilbert transformer (H1 e +H1 e ), the amplitude is smaller, and it can be seen that noise is effectively reduced.

[0039] In addition, as shown in Fig. 7, the simulation results for mass measurement on a conveyor were similar to those for mass measurement on a ship shown in Fig. 3, and the Hilbert transformer (H1 e +H2 e ) is transformed into a Hilbert transform signal by the Hilbert transformer (H1 e +H1 e ), the amplitude is smaller and it can be seen that the noise has been reduced. e +H1 e The amplitude of the signal transformed by the Hilbert transformer (H1 e +H2 e The amplitude of the signal transformed by the Hilbert transformer (H1 e +H1 e ), the Hilbert transformer (H1e +H2 e ) has a smaller amplitude than the Hilbert transformed signal.

[0040] As shown in FIG. 8, the simulation result of mass measurement on a conveyor is similar to the simulation result of mass measurement on a ship shown in FIG. 3, and the first Hilbert transformer H1 o and the second Hilbert transformer of odd order H2 o A Hilbert transformer (H1 o +H2 o ), the Hilbert transformer (H1 o +H1 o ), the amplitude is smaller, and it can be seen that noise is effectively reduced.

[0041] In addition, as shown in Fig. 9, the simulation results for mass measurement on a conveyor are similar to the simulation results for mass measurement on a ship shown in Fig. 3, in that the Hilbert transformer with odd orders (H1 o +H2 o ) is transformed into a Hilbert transform signal by the Hilbert transformer (H1 o +H1 o ), the amplitude is smaller and it can be seen that the noise has been reduced. o +H1 o The amplitude of the signal transformed by the Hilbert transformer (H1 o +H2 o The amplitude of the signal transformed by the Hilbert transformer (H1 o +H1 o ), the Hilbert transformer (H1 o +H2 o ) has a smaller amplitude than the Hilbert transformed signal.

[0042] In addition, as shown in Fig. 10, the simulation results for mass measurement on a conveyor also show that the amplitude of the odd signal is smaller than that of the even signal, similar to the simulation results for mass measurement on a ship shown in Fig. 5. This is also clear from the results in Figs. 7 and 9.

[0043] Each functional unit included in the noise reduction device 10 is configured, for example, with an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), etc., but is not limited thereto. For example, the noise reduction device 10 may be configured with a computer including a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), a non-volatile memory, and an input / output interface (I / O), etc. In this case, the noise reduction program may be stored in the ROM in advance, and the CPU may read the noise reduction program from the ROM and execute the noise reduction process, thereby realizing each functional unit. In addition, the noise reduction program may be stored in a non-volatile non-transitory recording medium, or distributed via a network, and appropriately installed to be stored in a non-volatile memory, etc.

[0044] Examples of non-volatile non-transient recording media include CD-ROMs (Compact Disc Read Only Memory), magneto-optical disks, HDDs (Hard Disk Drives), DVD-ROMs (Digital Versatile Disc Read Only Memory), flash memories, memory cards, and the like.

[0045] Next, as an operation of this embodiment, the noise reduction process executed by the noise reduction device 10 will be described with reference to the flowchart shown in Fig. 2. Note that the noise reduction process in Fig. 2 is repeatedly executed while the input signal x(n) is being input.

[0046] In step S100, a first Hilbert transform is performed on an input signal x(n) to generate a first Hilbert transform signal h1(n).

[0047] In step S101, a second Hilbert transform is performed on the first Hilbert transform signal h1(n) generated in step S100 to generate a second Hilbert transform signal h2(n).

[0048] In step S102, a delayed signal d(n) is generated by delaying the input signal x(n) by an order that is 1 / 2 the total order N of the order N1 of the first Hilbert transform processing in step S100 and the order N2 of the second Hilbert transform processing in step S101.

[0049] The processes in steps S100 and S101 and the process in step S102 are executed in parallel.

[0050] In step S103, the second Hilbert transform signal h2(n) generated in step S101 and the delayed signal d(n) generated in step S102 are added together to generate and output a sum signal y(n).

[0051] In this manner, in this embodiment, the first Hilbert transformer H1 and the second Hilbert transformer H2 are configured in cascade connection, so that an output signal y(n) can be obtained in which the noise is reduced while suppressing the amplitude from increasing relative to the input signal x(n).

[0052] Moreover, by making the orders of the first Hilbert transformer H1 and the second Hilbert transformer H2 odd numbers, it is possible to pass signals having normalized angular frequencies near 1.

[0053] Note that the present embodiment does not limit the invention according to the claims, and not all of the combinations of features described in the embodiment are necessarily essential to the solution of the invention. The above-mentioned embodiment includes inventions at various stages, and various inventions are extracted by combining the multiple components disclosed. Even if some components are deleted from all the components shown in the embodiment, as long as the effect can be obtained, the configuration from which these some components are deleted can be extracted as an invention. [Explanation of symbols]

[0054] 10 Noise reduction device H1 First Hilbert transformer H2 Second Hilbert transformer AD adder DL Delay Unit

Claims

1. a first Hilbert transformer for performing a Hilbert transform on an input signal; a second Hilbert transformer that is cascade-connected with the first Hilbert transformer and performs Hilbert transform on the first Hilbert transform signal output from the first Hilbert transformer, the second Hilbert transformer having inverse characteristics to the first Hilbert transformer; a delay unit that delays the input signal in accordance with the order of the first Hilbert transformer and the order of the second Hilbert transformer; an adder that adds the delayed signal output from the delay unit and the second Hilbert transform signal output from the second Hilbert transformer; A noise reduction device comprising:

2. The order of the first Hilbert transformer and the order of the second Hilbert transformer are odd orders.

2. The noise reduction device according to claim 1.

3. The computer converting the input signal into a first Hilbert transform signal by a first Hilbert transform process; converting the first Hilbert transform signal into a second Hilbert transform signal by a second Hilbert transform process having an inverse characteristic to the first Hilbert transform process; generating a delayed signal by delaying the input signal in accordance with an order of the first Hilbert transform processing and an order of the second Hilbert transform processing; adding the delayed signal and the second Hilbert transform signal; The noise reduction method according to claim 1,

4. On the computer, converting the input signal into a first Hilbert transform signal by a first Hilbert transform process; converting the first Hilbert transform signal into a second Hilbert transform signal by a second Hilbert transform process having an inverse characteristic to the first Hilbert transform process; generating a delayed signal by delaying the input signal in accordance with an order of the first Hilbert transform processing and an order of the second Hilbert transform processing; adding the delayed signal and the second Hilbert transform signal; A noise reduction program that performs a process including the steps of:

Images