JP2006284528A - Signal processing mechanism - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To remove a noise component of a digital detection noise obtained by A/D-converting a detection signal detected in each of various measuring instruments. <P>SOLUTION: Only the noise component is reduced by constitution of detecting an increase and decrease of a detection signal value, and of judging whether a compression value provided by dividing a difference value between the detection signal values before and after with a preset numeral is added to an output-after-processed value output just before or not, or whether the detection signal value is output as it is or not, followed with execution thereof. The constitution of comparing the difference value with a prescribed threshold value may be possible to enhance more processing precision. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention is detected by the detectors of various measuring devices (collectively referred to herein as “measuring devices”) including fluorescent X-ray analyzers, mass spectrometers, gas chromatographs, and liquid chromatographs. The present invention relates to a technique for removing a noise component of a detection signal. In particular, the present invention relates to a technique for removing a noise component of a detection signal, such as a fluorescent X-ray analysis, in which a waveform of a signal detected by a detector has a stepped shape.

  The analog signals detected by the measuring device are sequentially sampled (digital numerical values) by the A / D converter at a predetermined sampling frequency. In this specification, an analog signal detected by the measurement device is referred to as a “detection signal”, and a detection signal sampled by A / D conversion is referred to as a “digital detection signal”. Each digital numerical value constituting the digital detection signal is referred to as a “detection signal value”.

  Since the digital detection signal contains various noises, noise processing is performed by various signal processing devices in order to extract desired data from the detection signals to the maximum. Various signal processing mechanisms have been devised and disclosed so far. For example, Patent Document 1 discloses a signal processing device that stabilizes the signal level of a stepped voltage pulse signal in fluorescent X-ray analysis and improves analysis accuracy. Is described.

An example of the configuration of a signal processing system in a conventional measuring apparatus is shown in FIG. In this signal processing system, the detection signal output from the detector is converted into a digital detection signal having a stepped waveform via a preamplifier and an A / D converter. Filtering data is generated by the digital detection signal passing through the digital filter unit, and then a threshold determination or the like based on a pulse detection threshold is performed by a multi-channel analyzer to generate a spectrum histogram.
JP 2005-55399 A ([0013])

  In the signal processing performed under the above configuration, when the noise in the digital detection signal after A / D conversion is large, especially when the time constant of the filter is small, it is originally detected by the influence of the noise. There has been a problem that a signal that should not be detected is erroneously detected as a pulse, but there has been no effective technique for solving this problem.

  The problem to be solved by the present invention is to reduce only the noise part of the digital detection signal after A / D conversion in the digital detection signal having a stepped shape, and to reduce the filter time constant, Even if the detection threshold value is made small, no false detection occurs.

The signal processing mechanism according to the present invention, which has been made to solve the above problems,
A signal processing mechanism for reducing noise of a digital detection signal detected by a detector of a measurement device and A / D converted,
An increase / decrease detection unit that outputs whether or not the detection signal value is continuously increased or decreased a predetermined number of times,
A difference value calculation unit that calculates a difference value between the detection signal values before and after consecutive;
A difference value compression unit that calculates a compression value by dividing the difference value by a preset number;
When the output of the increase / decrease detection unit is positive, the detection signal value is output as it is as a processed output value, and when the output is negative, the compressed value is added to the processed output value output immediately before. And a processing unit that outputs the added value as a processed output value.

Furthermore, the signal processing mechanism according to the present invention is another preferred embodiment,
An increase / decrease detection unit that outputs whether or not the detection signal value is continuously increased or decreased a predetermined number of times,
A difference value calculation unit that calculates a difference value between the detection signal values before and after consecutive;
A difference value compression unit that calculates a compression value by dividing the difference value by a preset number;
A threshold value comparing unit that compares the difference value with a preset threshold value and outputs whether the difference value is larger or not;
The output of the increase / decrease detection unit and the output of the threshold comparison unit are compared. When the output of the threshold comparison unit is positive, a positive value is always output regardless of the output of the increase / decrease detection unit, and the threshold comparison A process determination unit that outputs a positive or negative value based on the output of the increase / decrease detection unit when the output of the unit is NO,
When the output of the process determination unit is positive, the detection signal value is output as it is as a processed output value, and when the output is negative, the compressed value is added to the post-processed output value output immediately before. And a processing unit that outputs the added value as a post-processing output value.

  By using the signal processing mechanism of the present invention, even if the filter time constant is small, it can be made less susceptible to noise included in the signal, and the filtering data is generated by mistakenly exceeding the pulse detection threshold. False detection is less likely to occur. As a result, the throughput is improved and the pulse detection threshold can be set to a small value, so that even a low element can be detected and the analysis accuracy is improved.

  The signal processing mechanism of the present invention is provided between the A / D converter and the digital filter in the measuring apparatus. The signal processing mechanism may be configured by software, or may be configured by hardware such as a circuit. Moreover, the structure with which both were combined may be sufficient.

  The configuration of the signal processing mechanism of the present invention will be described with reference to FIG. First, individual detection signal values of the A / D converted digital detection signal are sequentially stored in the detection signal value shift register 1. When M detection signal values are held in the detection signal value shift register 1, the difference value calculation unit 2 sequentially calculates the difference value between the Mth detection signal value and the M−1th detection signal value. The The difference value compression unit 3 provided in the subsequent stage of the difference value calculation unit 2 calculates the compression difference value by dividing the difference value by a predetermined value L. This value L may be set freely including infinity, however, according to actual measurement conducted by the inventor, about 2 to 4 is usually appropriate.

  On the other hand, the increase / decrease detection unit 4 compares the first detection signal value with the second detection signal value stored in the detection signal value shift register 1 to determine whether the value is increasing or decreasing. The detected increase / decrease value (usually binary) is detected, and the detected increase / decrease value is stored in the detection result shift register 41. The number of increase / decrease detection values stored in the detection result shift register 41 is the same as the number M of detection signal values stored in the detection signal value shift register 1. Next, the increase / decrease detection unit 4 sequentially receives the increase / decrease detection values stored in the detection result shift register 41, and whether the increase / decrease detection values continue for a preset number of times, that is, the detection signal values continue for a predetermined number of times. It is determined whether it has increased or continuously decreased, and 1 is output if it has increased or decreased continuously a predetermined number of times, and 0 is output otherwise. The value of the number of times set in advance is not particularly limited, but since this value corresponds to a slew rate (not regarded as noise) time, it is necessary to set an appropriate value. According to the inventor's actual measurement, about 4 to 8 is appropriate.

  In the processing unit 5, when the value output from the increase / decrease detection unit 4 is 1, the selector 52 outputs the Mth detection signal value stored in the detection signal value shift register 1 as it is as an output value after processing. . When the value output from the increase / decrease detection unit 4 is 0, the selector 52 adds the compression difference value calculated by the difference value compression unit 3 to the post-processing output value output immediately before from the selector 52 Is output as an output value after processing. This means that the original noise component is compressed to 1 / L.

  FIG. 3 shows an example of the digital detection signal (upper stage) and the processed output signal (lower stage) input to the signal processing mechanism according to the present invention. Although the noise is reduced in the flat portion of the detection signal, it is determined that the step portion (original signal component) is not noise, and a waveform in which only the noise is reduced is obtained.

  Depending on the A / D converted digital detection signal, the detection signal value may not monotonously increase or decrease a predetermined number of times even if it is originally a stepped portion. Even in such a case, in order to correctly recognize the staircase, that is, to further improve the accuracy of the signal processing according to the present invention, as described below, the signal processing mechanism further includes a threshold comparison unit and a processing. A determination unit may be provided to perform processing based on a predetermined threshold value.

  FIG. 2 shows the configuration of a signal processing mechanism according to the present invention further provided with a determination function based on a threshold. The threshold value comparison unit 6 compares the difference value calculated by the difference value calculation unit 2 with a preset threshold value. When the threshold value is larger than the difference value, 1 is set. Outputs 0.

  The process determination unit 7 receives the value 0 or 1 output from the threshold comparison unit 6 and the value 0 or 1 output from the increase / decrease detection unit 4. When the output value of the threshold comparison unit 6 is 1, regardless of the output value of the increase / decrease detection unit 4, a value of 1 is output to the selector 52 of the processing unit 5. At this time, the selector 52 outputs the Mth detection signal value stored in the detection signal value shift register 1 as a processed output value. On the other hand, when the output value of the threshold comparison unit 6 is 0, the process determination unit 7 outputs the value 0 or 1 that is the output value of the increase / decrease detection unit 4, and the selector 52 determines the noise component based on the value Is processed (in the case of 0) or the detection signal value is passed through as it is (in the case of 1).

FIG. 4 shows a graph of an example of detection signal processing by a signal processing mechanism having a determination function based on a threshold value. From the waveform of the digital detection signal (A) and the waveform of the processed output signal (B), it can be seen that the noise component is reduced in the flat portion of the signal and the staircase portion is correctly recognized. It was also confirmed that the signal component of the processed output signal was correctly recognized as a pulse by the processing according to the present invention from the filter output wave (C) and the pulse (D) above the threshold.

The block diagram of the signal processing mechanism which concerns on this invention. The block diagram of the other signal processing mechanism which concerns on this invention. Digital detection signal before signal processing (upper stage) and output signal after processing (lower stage). (A) Digital detection signal, (B) Output signal after processing, (C) Filter output wave, (d) Graph of output pulse above threshold. The structural example of the signal processing system in the conventional measuring device.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Detection signal value shift register 2 ... Difference value calculation part 3 ... Difference value compression part 4 ... Increase / decrease detection part 41 ... Detection result shift register 5 ... Processing part 51 ... Adder 52 ... Selector 6 ... Threshold comparison part 7 ... Process judgment Part

Claims (3)

A signal processing mechanism for reducing noise of a digital detection signal detected by a detector of a measurement device and A / D converted,
An increase / decrease detection unit that outputs whether or not the detection signal value is continuously increased or decreased a predetermined number of times,
A difference value calculation unit that calculates a difference value between the detection signal values before and after consecutive;
A difference value compression unit that calculates a compression value by dividing the difference value by a preset number;
When the output of the increase / decrease detection unit is positive, the detection signal value is output as it is as a processed output value, and when the output is negative, the compressed value is added to the processed output value output immediately before. A processing unit that outputs the added value as an output value after processing;
A signal processing mechanism comprising: A signal processing mechanism that reduces noise of a detection signal after being detected by a detector of a measurement device and A / D converted,
An increase / decrease detection unit that outputs whether or not the detection signal value is continuously increased or decreased a predetermined number of times,
A difference value calculation unit that calculates a difference value between the detection signal values before and after consecutive;
A difference value compression unit that calculates a compression value by dividing the difference value by a preset number;
A threshold value comparing unit that compares the difference value with a preset threshold value and outputs whether the difference value is larger or not;
The output of the increase / decrease detection unit and the output of the threshold comparison unit are compared. When the output of the threshold comparison unit is positive, a positive value is always output regardless of the output of the increase / decrease detection unit, and the threshold comparison A process determination unit that outputs a positive or negative value based on the output of the increase / decrease detection unit when the output of the unit is NO,
When the output of the process determination unit is positive, the detection signal value is output as it is as a processed output value, and when the output is negative, the compressed value is added to the post-processed output value output immediately before. A processing unit that outputs the added value as an output value after processing;
A signal processing mechanism comprising: The signal processing mechanism according to claim 1, wherein the measurement device is a fluorescent X-ray analysis device.

Images