WO2013064035A1

WO2013064035A1 - Denoising method and device for optical signal

Info

Publication number: WO2013064035A1
Application number: PCT/CN2012/083605
Authority: WO
Inventors: 代勇; 郁琦
Original assignee: 深圳迈瑞生物医疗电子股份有限公司
Priority date: 2011-10-31
Filing date: 2012-10-26
Publication date: 2013-05-10
Also published as: CN103090886A; CN103090886B

Abstract

A denoising method for an optical signal, including the following steps: receiving an optical signal, and generating a simulated optical pulse signal (S01); converting the optical pulse signal to a digital optical signal (S02); acquiring at least two kinds of pulse characteristic information of each pulse in the digital optical signal (S03); generating a 2D or more than 2D scatter diagram for the digital optical signal taking the at least two kinds of pulse characteristic information as a coordinate axis (S04); removing the pulse data corresponding to a noise data point in the digital optical signal according to the feature that the distribution area of the valid data points and that of the noise digital points are separated relatively in the scatter diagram of the digital optical signal and generating a valid optical signal (S05); and storing the valid optical signal (S06). Also provided is a corresponding denoising device for an optical signal. The denoising method and device for an optical signal employ multi-dimensional denoising, and by way of experiment verification, as compared to the conventional art, the denoising method and device for an optical signal can reach better denoising effects, thus making the subsequent detection results more accurate.

Description

Optical signal denoising method and device

[Technical Field]

The present invention relates to the field of data processing, and in particular, to a method and apparatus for denoising an optical signal.

【Background technique】

In blood cell analyzers using fluorescence and scattering methods, fluorescent and scatter signals are extracted for identification and analysis of various blood cells. The fluorescence signal is very weak and is easily interfered by various factors, such as background noise, electrical noise, non-specific fluorescent signals, etc. (the so-called non-specific fluorescent signal is relative to the fluorescent signal emitted by the substance to be tested, it is other substances in the blood. The resulting fluorescent signal, also known as noise, has a significant impact on the accuracy of fluorescence signal analysis. Especially when measuring platelets, the signal generated by platelets is weak, and the influence of noise and interference signals is particularly serious. Similarly, scattered signals can also be affected by noise.

Current blood cell analyzers using fluorescence and scattering methods often use improved circuit design, use expensive devices to reduce noise, or design complex filters to improve system signal-to-noise ratio, or improve signal recognition thresholds to remove noise. influences. The shortcomings and disadvantages of these methods are that the circuit noise reduction method will increase the design and manufacturing cost, and the noise reduction degree is limited. Although the recognition threshold can remove noise and interference to some extent, it will reduce the sensitivity and influence of the measurement. The accuracy of the results.

[Summary of the Invention]

Based on this, it is necessary to provide a more accurate and efficient method and apparatus for denoising optical signals.

A method for denoising an optical signal, the embodiment of the present invention takes an optical signal as an example, and includes the following steps:

Step S01, receiving an optical signal to generate an analog optical pulse signal;

Step S02, converting the optical pulse signal into a digital optical signal;

Step S03, acquiring at least two types of pulse characteristic information of each pulse in the digital optical signal;

Step S04, generating a scattergram of the digital optical signal of two or more dimensions with the at least two kinds of pulse characteristic information as coordinate axes;

Step S05, according to the feature that the effective data point and the noise data point distribution area in the scattergram of the digital optical signal are relatively separated, the pulse data corresponding to the noise data point in the digital optical signal is removed, and an effective optical signal is generated;

Step S06, storing the effective optical signal.

In a preferred embodiment, the step S05 is to remove pulse data in the digital light signal corresponding to the preset denoising template according to the preset denoising template; the denoising template is The scatter plot of the digital light signal corresponds to a two-dimensional or two-dimensional boundary region having a boundary value corresponding to the noise data point distribution region.

In a preferred embodiment, the preset denoising templates have at least two, corresponding to the background detection mode and the blood sample detection mode.

In a preferred embodiment, the step S05 includes the following steps:

Obtaining a value value of a data point within a preset range in the scattergram of the digital light signal;

Comparing the quantity value with a preset value, when it is less than the preset value, determining that the background detection mode outputs a first mode signal; when it is greater than or equal to the preset value, determining that the Blood sample detection mode, outputting a second mode signal;

Denoising is performed by calling the corresponding denoising template according to the first mode signal and the second mode signal.

In a preferred embodiment, the at least two kinds of pulse characteristic information are a two-two combination or a three-three combination in a peak, a relative peak, a pulse width, a pulse area, a pulse generation time, and a pulse interval time.

In a preferred embodiment, the pulse characteristic information acquired in the step S03 includes a relative peak value and a pulse width; and the step S04 is to generate the two-dimensional digital light signal with the relative peak value and the pulse width as coordinate axes. Scatter plot.

In a preferred embodiment, the step S05 includes the following steps:

Identifying one large and one small data point distribution area in the scattergram of the digital light signal;

The pulse data corresponding to the data points in the smaller distribution area is removed.

In a preferred embodiment, the optical signal is a fluorescent signal.

A denoising device for an optical signal, comprising: a photosensor for receiving an optical signal to generate an analog optical pulse signal; and an analog to digital conversion circuit for using the simulated light The pulse signal is converted into a digital optical signal; the denoising device of the optical signal further comprises: a pulse extraction module, a pulse information combination module, a denoising module and a storage module,

The pulse extraction module is configured to acquire at least two types of pulse characteristic information of each pulse in the digital optical signal;

The pulse information combination module is configured to generate a scattergram of the digital optical signal of two or more dimensions with at least two kinds of pulse characteristic information as coordinate axes;

The denoising module is configured to remove pulse data corresponding to the noise data point in the digital optical signal according to a feature that the effective data point and the noise data point distribution area in the scattergram of the digital optical signal are relatively separated, and generate effective light. signal;

The storage module is configured to store the effective optical signal.

In a preferred embodiment, the denoising module includes: a denoising template memory and a denoising unit, wherein the denoising template memory is configured to store at least one preset corresponding to a scattergram of the digital optical signal a two-dimensional or two-dimensional denoising template having a boundary value corresponding to the noise data point distribution area; the denoising unit is configured to invoke the denoising template in the denoising template memory to remove the The pulse data in the region corresponding to the denoising template in the digital optical signal.

In a preferred embodiment, at least two types of the denoising templates are preset in the denoising template memory, corresponding to the background detection mode and the blood sample detection mode.

In a preferred embodiment, the denoising module further includes: a pulse information statistic unit and a comparison unit, wherein the pulse information statistic unit is configured to analyze and obtain data in a preset range in the scattergram of the digital optical signal. The number of points;

The comparing unit is configured to compare the quantity value with a preset value, and when it is less than the preset value, determine that the background detecting mode outputs a first mode signal, when the preset value is greater than or equal to the preset value. Determining that the blood sample detection mode outputs a second mode signal;

The denoising unit is configured to invoke the corresponding denoising template in the denoising template memory according to the first mode signal and the second mode signal, and remove the denoising template from the digital optical signal. Corresponding to the pulse data in the area.

In a preferred embodiment, the denoising module includes: an image recognition unit and a noise processing unit, wherein the image recognition unit is configured to identify one large and one small data points in the scattergram of the digital light signal a distribution area; the noise processing unit is configured to remove pulse data corresponding to data points in the smaller distribution area.

In a preferred embodiment, the at least two kinds of pulse characteristic information acquired by the pulse extraction module are two or two combinations in a peak, a relative peak, a pulse width, a pulse area, a pulse generation time, and a pulse interval time. Three or three combinations.

In a preferred embodiment, the pulse extraction module includes: a pulse identification unit, a peak buffer, a baseline buffer, a pulse width buffer, and a relative peak calculation unit.

The pulse identification unit is configured to acquire peak information, baseline information, and pulse width information of each pulse in the digital optical signal, and store the information in the peak buffer, the baseline buffer, and the pulse width buffer, respectively. ;

The relative peak calculating unit is configured to calculate relative peak information according to the peak information in the peak buffer and the baseline buffer and the baseline information;

The pulse information combining module is configured to generate a scattergram of the two-dimensional digital light signals with relative peaks and pulse widths as coordinate axes.

In a preferred embodiment, the optical signal is a fluorescent signal.

The above method and device for denoising an optical signal are effective data points and noises in a multi-dimensional (two-dimensional and above) scattergram composed of a plurality of (two or more) pulse characteristic information of a digital optical signal obtained by experimental research. The data point distribution area is relatively separated, and the two-dimensional or multi-dimensional denoising is performed. The traditional technology only uses a single threshold filtering method (ie, horizontal/vertical one-cutting method) to denoise, and the denoising effect is that either more noise signals can be removed or more effective data can be removed. . Through experimental verification, compared with the conventional technology, the optical signal denoising method and device of the present invention can achieve better denoising effect, thereby making the subsequent detection result more accurate.

[Description of the Drawings]

1 is a flow chart showing the steps of a method for denoising an optical signal according to an embodiment;

2 is a schematic diagram of a two-dimensional scattergram of a digital fluorescent signal;

3 is a schematic diagram of a denoising template corresponding to a two-dimensional scattergram of the digital fluorescent signal shown in FIG. 2;

Figure 4 is a two-dimensional scatter plot of the digital fluorescent signal and the digital scattered signal of the background before denoising;

Figure 5 is a two-dimensional scatter plot of the digital fluorescent signal and the digital scattered signal of the background after denoising;

Figure 6 is a two-dimensional scatter plot of the digital fluorescent signal and the digital scattered signal of the blood sample before denoising;

7 is a two-dimensional scattergram of a digital fluorescent signal and a digital scattered signal of a blood sample after denoising;

8 is a functional block diagram of a denoising device for an optical signal according to an embodiment;

9 is a functional block diagram of a denoising module of another embodiment.

【detailed description】

In order to solve the problem that the detection result is inaccurate due to noise during the fluorescence detection process, a denoising method and device for improving the accuracy of the fluorescence detection result are proposed.

As shown in FIG. 1 , it is a flow chart of a method for denoising an optical signal according to an embodiment, and includes the following steps:

In step S01, a fluorescent signal is received to generate a simulated fluorescent pulse signal.

In step S02, the fluorescent pulse signal is converted into a digital fluorescent signal.

Step S03, acquiring at least two kinds of pulse characteristic information of each pulse in the digital fluorescent signal.

The pulse characteristic information refers to information such as a peak value, a relative peak value, a pulse width, a pulse area, a time at which a pulse is generated, and a time interval at a pulse interval. The relative peak refers to the peak value of the actually obtained peak relative to the baseline of the signal. The at least two kinds of pulse characteristic information acquired in the step S03 may be any two-two combination or three-three combination of the above-mentioned various pulse characteristic information.

In step S04, a scattergram of two-dimensional or two-dimensional digital fluorescent signals with at least two types of pulse characteristic information as coordinate axes is generated.

If two kinds of pulse characteristic information are used as the coordinate axes, a two-dimensional scattergram of the digital fluorescent signal is generated.

If three kinds of pulse characteristic information are used as the coordinate axes, a three-dimensional scattergram of the digital fluorescent signal is generated. And so on. As shown in FIG. 2, it is a schematic diagram of a two-dimensional scattergram of a digital fluorescent signal with a pulse width as a horizontal axis and a relative peak as a vertical axis obtained by experiments. In Figure 2, the distribution of data points presents one large and one small two separate data areas: a larger data area 10 and a smaller data area 12. Moreover, it is obtained through experimental analysis that the smaller data area 12 is distributed with noise data points, and the larger data area 10 is distributed with valid data points. In reality, the two data areas do not necessarily have the absolute separation as shown in Figure 2, there will be a small part of the adhesion, but basically separated.

Step S05, according to the feature that the effective data point and the noise data point distribution area in the scattergram of the digital fluorescence signal are relatively separated, the pulse data corresponding to the noise data point in the digital fluorescent signal is removed, and an effective fluorescent signal is generated. The effective fluorescent signal is the denoised digital fluorescent signal. The relative separation refers to: in actual cases, the area where the effective data points and the noise data points are distributed is not absolutely separated, and there is a small part of adhesion, but basically in two separate areas.

Step S06, storing the effective fluorescent signal.

Still taking FIG. 2 as an example, step S05 is to remove the pulse data corresponding to the noise data points distributed in the smaller data area 12. After obtaining the scattergram of the digital fluorescent signal and knowing that the effective data point and the noise data point distribution area are relatively separated, there are many methods for removing the pulse data corresponding to the noise data point, such as by using a preset denoising template. Denoising and image processing are used for denoising.

In the above step S05, if denoising is performed by using a preset template, at least one two-dimensional or two-dimensional denoising template having a boundary value corresponding to the scattergram of the digital fluorescent signal may be preset in advance, the boundary. The value corresponds to the noise data point distribution area. Step S05 is to remove the pulse data in the corresponding region of the digital fluorescence signal corresponding to the preset denoising template according to the preset denoising template.

In a preferred embodiment, two different denoising templates can be preset and used in different situations. For example, the blood cell analyzer includes two types of background detection mode and blood sample detection mode. The background detection mode refers to the empty machine test when the blood sample is not placed to check whether the instrument is clean. The blood sample detection mode refers to the detection mode in which the blood sample to be tested is placed. The preset two different denoising templates respectively correspond to the two detection modes.

As shown in FIG. 3, it is a schematic diagram of a denoising template corresponding to the two-dimensional scattergram of the digital fluorescent signal shown in FIG. 2. The boundary area 20 and the border area 22 shown in FIG. 3 represent two different denoising templates preset. If denoising according to the boundary region 20, please refer to FIG. 2 at the same time, that is, the data points (corresponding pulse data) in the region corresponding to the boundary region 20 in FIG. 2 are removed, that is, the dotted line 14 and the coordinate axis in FIG. 2 are removed. Data points within a region, and these data points are noise data points.

In this case, different denoising templates are set for the two detection modes because: the distribution range of the noise data points in the two detection modes obtained according to the experiment is different, and accordingly, corresponding denoising templates are set for different detection modes to achieve More accurate denoising effect.

In order to automatically select different denoising templates according to different detection modes, the present invention provides an automatic identification detection mode method, as follows:

First, the number of data points in the preset range in the scattergram of the digital fluorescence signal obtained in step S04 is analyzed. Continuing with FIG. 2 as an example, the preset range may be: a range defined by a relative peak > 512 and a pulse width > 10, and the specific values are obtained experimentally.

And comparing the quantity value with the preset value, and when less than the preset value (such as 50), determining the background detection mode, and outputting the first mode signal. In step S05, the denoising template corresponding to the background detection mode is called according to the first mode signal to perform denoising. When the preset value is greater than or equal to, the blood sample detection mode is determined, and the second mode signal is output. Step S05 is to perform denoising according to the demodulation template corresponding to the blood sample detection mode according to the second mode signal.

When a blood cell analyzer performs background and blood sample detection, it is usually a two-dimensional scattergram that establishes a digital fluorescent signal and a digital scattered signal to analyze the sample. The digital scatter signal is obtained by analog-to-digital conversion of a forward pulse signal generated by a scatter signal.

Please also refer to FIG. 4 and FIG. 5, which are two-dimensional scatter plots based on digital fluorescence signals and digital scatter signals before and after denoising in actual background detection. The horizontal axis is the digital fluorescent signal FL and the vertical axis is the digital scatter signal. FSC. It can be clearly seen that there are more data points in the lower left corner of Figure 4, and all data points are noise data points in the background detection. Figure 5 shows the effect of denoising, where the data points are significantly reduced, that is, the noise data is substantially cleared.

Please also refer to Figure 6 and Figure 7, which are two-dimensional scatter plots based on digital fluorescence signals and digital scatter signals before and after denoising, the horizontal axis is the digital fluorescence signal FL and the vertical axis is the digital scatter signal FSC. . The data points in the box shown in Figure 6 are the corresponding data points of the platelets in the blood sample. It is obvious that there are a large number of data points on the left side of the box. (Traditionally, the left side of the box is directly placed on this figure. The data points are all "cut off", but doing so will cut off useful platelet signals at the same time, resulting in reduced accuracy of the results) and most of these data points are noise data points. In the analysis of platelets, since the noise data points and the data points corresponding to the platelets are very close, these noise data points will be calculated as the data points of the platelets, resulting in the detected platelet content being much higher than the actual platelet content, severely distorted. . Next, look at Figure 7, which is the result of denoising. The data points in the box are the data points corresponding to the platelets. The data points around the box are basically cleared, so that the platelet test results in the subsequent blood samples are more accurate.

In order to further verify the effects of the present invention, it is also detected by an impedance method while performing optical detection. The platelet value obtained by the impedance sample using the impedance method was 0 in the background detection mode and 81 × 10 ⁹ /L in the blood sample detection mode. The data measured before and after denoising by optical detection are shown in the following table:

		Before denoising	After denoising
Optical platelet measurement (10 ⁹ /L)	Background detection mode	twenty one	1
Optical platelet measurement (10 ⁹ /L)	Blood sample detection mode	141	78

It can be clearly seen that before denoising, the detection value of the optical platelet reaches about twice the detected value of the impedance platelet, which is severely distorted; after denoising, the optical platelet detection value returns to the normal level, and the impedance method Platelet detection values are very close.

If the above step S05 is performed by image processing for denoising, the following steps are included:

First, a large and a small data point distribution area in the scattergram of the digital fluorescent signal obtained in step S04 is identified.

The pulse data corresponding to the data points in the smaller distribution area is then removed.

This method does not need to preset the boundary of the fixed data point to be removed, but removes the data in the smaller distribution area by means of image recognition, and the purpose and effect of removing noise can also be achieved.

In addition, the present invention provides a denoising device for the above-mentioned optical signal denoising method with an optical signal corresponding thereto.

As shown in FIG. 8, it is a functional block diagram of the optical signal denoising device 30 of an embodiment, including: a photo sensor 301, a signal conditioning circuit 302, an analog to digital conversion circuit 303, a pulse extraction module 304, and a combination of pulse information. Module 305, denoising module 306, and storage module 307.

The photosensor 301 is configured to receive a fluorescent signal to generate a simulated fluorescent pulse signal.

The signal conditioning circuit 302 is configured to amplify, filter, and limit the fluorescent pulse signals.

An analog to digital conversion circuit 303 is used to convert the fluorescent pulse signal into a digital fluorescent signal.

The pulse extraction module 304 is configured to acquire at least two types of pulse characteristic information of each pulse in the digital fluorescent signal.

The pulse characteristic information refers to information such as a peak value, a relative peak value, a pulse width, a pulse area, a timing at which a pulse is generated, and a time interval of a pulse interval. The relative peak refers to the peak value of the actually obtained peak relative to the baseline of the signal. The acquired at least two kinds of pulse characteristic information may be any two-two combination or three-three combination of the various pulse characteristic information enumerated above.

The pulse information combination module 305 is configured to generate a scattergram of a two-dimensional or two-dimensional digital fluorescent signal having at least two types of pulse characteristic information as coordinate axes.

The denoising module 306 is configured to remove the pulse data corresponding to the noise data points in the digital fluorescent signal according to the characteristics of the relative separation of the effective data points and the noise data point distribution regions in the scattergram of the digital fluorescent signal, to generate an effective fluorescent signal.

The storage module 307 is configured to store the effective fluorescent signal. The effective fluorescent signal is the denoised digital fluorescent signal.

In this embodiment, the pulse extraction module 304 further includes: a pulse identification unit 341, a peak buffer 342, a baseline buffer 343, a pulse width buffer 344, and a relative peak calculation unit 345.

The pulse identification unit 341 is configured to acquire peak information, baseline information, and pulse width information of each pulse in the digital fluorescent signal, and store them in the peak buffer 342, the baseline buffer 343, and the pulse width buffer 344, respectively.

The relative peak calculation unit 345 is configured to calculate relative peak information based on the peak information and the baseline information in the peak buffer 342 and the baseline buffer 343.

At this time, the pulse information combining module 305 is used to generate a scattergram of the two-dimensional digital fluorescent signal with the relative peak and pulse width as the coordinate axes.

In this embodiment, the denoising module 306 further includes: a pulse information counting unit 361, a comparing unit 362, a denoising template memory 363, and a denoising unit 364.

The denoising template memory 363 is configured to store at least one predetermined two-dimensional or two-dimensional denoising template having a boundary value corresponding to a scattergram of the digital fluorescent signal, the boundary value corresponding to the noise data point distribution area. In a preferred embodiment, the denoising template memory 363 stores two denoising templates corresponding to the background detection mode and the blood sample detection mode.

The pulse information statistics unit 361 is configured to analyze the number of data points in the scatter plot of the digital fluorescence signal within a preset range. Taking the two-dimensional scattergram of the relative peak and pulse width as described above as an example, the preset range may be a range defined by a relative peak > 512 and a pulse width > 10.

The comparison unit 362 is configured to compare the quantity value with a preset value, and when less than the preset value (such as 50), determine a background detection mode, and output a first mode signal; when the preset is greater than or equal to the preset When the value is determined, the blood sample detection mode is determined, and the second mode signal is output.

The denoising unit 364 is configured to invoke a corresponding denoising template in the denoising template memory 363 according to the first mode signal and the second mode signal to remove pulse data in the corresponding region of the digital fluorescence signal and the denoising template.

Please refer to FIG. 9 , which is a functional block diagram of the denoising module 406 of another embodiment, which includes an image recognition unit 461 and a noise processing unit 462 .

The image recognition unit 461 is configured to recognize a large one and a small data point distribution area in the scattergram of the digital fluorescent signal.

Because, through multiple experimental tests, the effective data points and the noise data points in the scatter plot of the digital fluorescent signal are respectively distributed in two different sizes, and the noise data point distribution area is small. Based on this feature, the present invention proposes the use of image processing to remove noise.

The noise processing unit 462 is configured to remove pulse data corresponding to data points in the smaller distribution area.

In summary, the optical signal denoising method and apparatus of the present invention are multi-dimensional (two-dimensional and above) scatters composed of multiple (two or more) pulse characteristic information of the digital fluorescent signal obtained by experimental research. In the figure, the effective data points and the noise data point distribution areas are relatively separated, and the two-dimensional or multi-dimensional denoising is performed. That is, after the multi-dimensional scattergram of the digital fluorescent signal is established, the noise data can be conveniently and accurately removed, such as the preset denoising template method and image processing method provided in the present case.

The traditional technology only uses a single threshold filtering method (ie, horizontal/vertical one-cutting method) to denoise, and the denoising effect is that either more noise signals can be removed or more effective data can be removed. .

Through experimental verification, compared with the conventional technology, the optical signal denoising method and device of the present invention can achieve better denoising effect, thereby making the subsequent detection result more accurate.

The above method and device for denoising the optical signal are only exemplified by the fluorescent signal, and other optical signals, how to scatter the signal, etc., can be denoised by using the denoising method and device of the above optical signal.

The above-mentioned embodiments are merely illustrative of several embodiments of the present invention, and the description thereof is more specific and detailed, but is not to be construed as limiting the scope of the invention. It should be noted that a number of variations and modifications may be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of the invention should be determined by the appended claims.

Claims

A method for denoising an optical signal, comprising the steps of:

Step S01, receiving an optical signal to generate an analog optical pulse signal;

Step S02, converting the optical pulse signal into a digital optical signal;

Step S03, acquiring at least two types of pulse characteristic information of each pulse in the digital optical signal;

Step S04, generating a scattergram of the digital optical signal of two or more dimensions with the at least two kinds of pulse characteristic information as coordinate axes;

Step S05, according to the feature that the effective data point and the noise data point distribution area in the scattergram of the digital optical signal are relatively separated, the pulse data corresponding to the noise data point in the digital optical signal is removed, and an effective optical signal is generated;

Step S06, storing the effective optical signal.
The method for denoising an optical signal according to claim 1 , wherein the step S05 is to remove a region of the digital optical signal corresponding to the preset denoising template according to a preset denoising template. Pulse data; the denoising template is a two-dimensional or two-dimensional or more boundary region provided with a boundary value corresponding to a scattergram of the digital optical signal, and the boundary value corresponds to a noise data point distribution region.
The optical signal denoising method according to claim 2, wherein the preset denoising templates are at least two, corresponding to a background detection mode and a blood sample detection mode, respectively.
The optical signal denoising method according to claim 3, wherein the step S05 comprises the following steps:

Obtaining a value value of a data point within a preset range in the scattergram of the digital light signal;

Comparing the quantity value with a preset value, when it is less than the preset value, determining that the background detection mode outputs a first mode signal; when it is greater than or equal to the preset value, determining that the Blood sample detection mode, outputting a second mode signal;

Denoising is performed by calling the corresponding denoising template according to the first mode signal and the second mode signal.
The optical signal denoising method according to claim 1, wherein the at least two kinds of pulse characteristic information are in a peak, a relative peak, a pulse width, a pulse area, a pulse generation time, and a pulse interval time. Two or two combinations or three or three combinations.
The optical signal denoising method according to claim 1, wherein the pulse characteristic information acquired in the step S03 comprises a relative peak value and a pulse width; and the step S04 is to generate the relative peak value and the pulse width as coordinates. A two-dimensional scatter plot of the digital light signal of the axis.
The optical signal denoising method according to claim 1, wherein the step S05 comprises the following steps:

Identifying one large and one small data point distribution area in the scattergram of the digital light signal;

The pulse data corresponding to the data points in the smaller distribution area is removed.
The optical signal denoising method according to claim 1, wherein the optical signal is a fluorescent signal.
A denoising device for an optical signal, comprising: a photosensor for receiving an optical signal to generate an analog optical pulse signal; and an analog to digital conversion circuit for using the simulated light The pulse signal is converted into a digital light signal; wherein the optical signal denoising device further comprises: a pulse extraction module, a pulse information combination module, a denoising module and a storage module,

The pulse extraction module is configured to acquire at least two types of pulse characteristic information of each pulse in the digital optical signal;

The pulse information combination module is configured to generate a scattergram of the digital optical signal of two or more dimensions with at least two kinds of pulse characteristic information as coordinate axes;

The denoising module is configured to remove pulse data corresponding to the noise data point in the digital optical signal according to a feature that the effective data point and the noise data point distribution area in the scattergram of the digital optical signal are relatively separated, and generate effective light. signal;

The storage module is configured to store the effective optical signal.
The optical signal denoising apparatus according to claim 9, wherein the denoising module comprises: a denoising template memory and a denoising unit, wherein the denoising template memory is configured to store at least one preset a two-dimensional or two-dimensional denoising template having a boundary value corresponding to a scattergram of the digital optical signal, wherein the boundary value corresponds to a noise data point distribution area; and the denoising unit is configured to invoke the denoising template The denoising template in the memory removes pulse data in a region corresponding to the denoising template in the digital optical signal.
The optical signal denoising apparatus according to claim 10, wherein at least two of the denoising templates are pre-set in the denoising template memory, corresponding to a background detection mode and a blood sample detection mode.
The optical signal denoising apparatus according to claim 11, wherein the denoising module further comprises: a pulse information statistical unit and a comparing unit, wherein the pulse information statistical unit is configured to analyze the obtained digital optical signal The number of data points in the scatter plot within the preset range;

The comparing unit is configured to compare the quantity value with a preset value, and when it is less than the preset value, determine that the background detecting mode outputs a first mode signal, when the preset value is greater than or equal to the preset value. Determining that the blood sample detection mode outputs a second mode signal;

The denoising unit is configured to invoke the corresponding denoising template in the denoising template memory according to the first mode signal and the second mode signal, and remove the denoising template from the digital optical signal. Corresponding to the pulse data in the area.
The optical signal denoising apparatus according to claim 9, wherein the denoising module comprises: an image recognition unit and a noise processing unit, wherein the image recognition unit is configured to identify a scatter of the digital optical signal One large and one small data point distribution area in the figure; the noise processing unit is used to remove pulse data corresponding to data points in a small distribution area.
The optical signal denoising apparatus according to claim 9, wherein the at least two kinds of pulse characteristic information acquired by the pulse extraction module are at a peak, a relative peak, a pulse width, a pulse area, and a pulse generation time. And two or two combinations of three or three in the time interval of the pulse interval.
The optical signal denoising apparatus according to claim 9, wherein the pulse extraction module comprises: a pulse identification unit, a peak buffer, a baseline buffer, a pulse width buffer, and a relative peak calculation unit,

The pulse identification unit is configured to acquire peak information, baseline information, and pulse width information of each pulse in the digital optical signal, and store the information in the peak buffer, the baseline buffer, and the pulse width buffer, respectively. ;

The relative peak calculating unit is configured to calculate relative peak information according to the peak information in the peak buffer and the baseline buffer and the baseline information;

The pulse information combining module is configured to generate a scattergram of the two-dimensional digital light signals with relative peaks and pulse widths as coordinate axes.
The optical signal denoising apparatus according to claim 8, wherein the optical signal is a fluorescent signal.