CN116359974B - Method for processing uranium radioactive pollution detection data in pipeline - Google Patents

Abstract

The invention relates to the technical field of data processing, in particular to a method for processing uranium radioactivity pollution detection data in a pipeline. According to the method for selecting the optimal filtering window for each data point, noise can be accurately filtered, effective data can be prevented from being filtered, and the uranium radioactive pollution detection precision is improved.

Description

Method for processing uranium radioactive pollution detection data in pipeline

Technical Field

The invention relates to the technical field of data processing, in particular to a method for processing uranium radioactive contamination detection data in a pipeline.

Background

With the continuous development of the nuclear industry, a part of nuclear industry facilities enter a retirement or impending retirement stage, and a nuclear pipeline is an important nuclear industry facility in the nuclear industry, so the detection of radioactive pollution in the retired nuclear pipeline is a key technical problem for implementing retirement of the nuclear industry facilities. In the traditional technology, the detection method of the pipeline inner core pollution generally generates alpha decay when uranium radioactivity pollutes mainly in the nuclear industry technology by an ionization method, so that the detection of the uranium radioactivity pollution in the pipeline can be completed by detecting the alpha pollution on the inner wall of the pipeline, but the pipeline length measured by the ionization method is limited in range, is easily influenced by beta particles and detection environment, has poor capability of locating the specific uranium radioactivity pollution position, and has larger limitation. According to the prior art, the alpha energy spectrogram corresponding to the pipeline can be detected by the pipeline robot and the alpha energy spectrometer, and the detection treatment of uranium radioactive pollution in the pipeline is completed according to the data spectral line of the alpha energy spectrogram.

However, when the pipeline robot carries the radiation detector to acquire the alpha energy spectrum, the finally obtained alpha energy spectrum is seriously affected by noise due to the factors of radionuclide decay, inherent statistical fluctuation of the alpha energy spectrometer, alpha ray scattering, electronic system noise and the like, so that the phenomena of false peaks, weak peaks covering and the like appear on spectral lines corresponding to the alpha energy spectrum. Therefore, when the existing spectral line noise removing method is adopted, the phenomena of weak peak leakage, incapability of distinguishing heavy peaks with very close distance and the like caused by improper selection of the size of a filtering window further cause the problems of inaccurate noise filtering, incorrect filtering of effective data and the like, so that the noise removing effect is poor, and the uranium radioactive pollution detection precision is reduced.

Disclosure of Invention

In order to solve the problems that when the existing spectral line noise removing method removes alpha energy spectrogram noise, the phenomena of weak peak leakage, incapability of distinguishing heavy peaks with very close distance and the like caused by improper selection of the size of a filtering window, and poor denoising effect caused by the problems of inaccurate noise filtering, incorrect effective data filtering and the like, the invention aims to provide a uranium radioactive pollution detection data processing method in a pipeline, which comprises the following specific steps:

the invention provides a method for processing uranium radioactive pollution detection data in a pipeline, which comprises the following steps:

acquiring alpha spectrum lines of each region of the inner wall of the uranium radioactivity pipeline to be detected;

dividing each alpha spectrogram spectral line into at least two spectral line segments, obtaining a first noise degree of each data point according to the position distribution characteristics of all data points on each spectral line segment, screening each data point according to the first noise degree to obtain a marked data point, and obtaining a second noise degree of all data points according to the position distribution characteristics of the marked data points on each spectral line segment;

obtaining similar alpha spectrogram spectral lines of each alpha spectrogram spectral line, obtaining data point noise degree change characteristics according to the difference of the second noise degree and the first noise degree of each data point, and obtaining final noise degree according to the noise degree change characteristics of the data points and the difference characteristics between the alpha spectrogram spectral lines where the data points are positioned and the similar alpha spectrogram spectral lines; screening each data point according to the final noise degree to obtain standard data points, obtaining a smooth degree according to the final noise degree of the data points on each spectrum segment and the position distribution characteristics of the standard data points, obtaining preselected filter windows with different sizes according to the smooth degree of each spectrum segment, and selecting an optimal filter window corresponding to each data point according to the final noise degree distribution characteristics of the data points in each preselected filter window of each spectrum segment and the data value change characteristics of the filtered data points;

And (3) filtering by using an optimal filter window according to each data point to complete denoising of each alpha spectrogram spectral line, and detecting uranium radioactive pollution according to an alpha spectrogram corresponding to the denoised alpha spectrogram spectral line.

Further, the method for acquiring the first noise level includes:

performing straight line fitting on the data points on the target spectrum line segment to obtain a corresponding fitting straight line;

if the target data point is not the end point, calculating the angle difference between the vector direction angle of the two adjacent data points of the target data point on the target spectrum line segment and the direction angle of the fitting straight line; obtaining a first noise degree of the target data point according to the product of the angle difference and the distance between the target data point and the fitting straight line; the vectors include a front vector that points from a previous data point to a target data point and a rear vector that points from the target data point to a subsequent data point;

if the target data point is an endpoint, calculating the angle difference between the direction angle of the vector between the adjacent data points of the target data point on the target spectrum line segment and the direction angle of the fitting straight line; obtaining a first noise degree of the target data point according to the product of the two times of the angle difference and the distance from the target data point to the fitting straight line; when the target data point is the starting end point, the vector points from the target data point to the next data point; when the target data point is a termination endpoint, the vector points from the previous data point to the target data point;

Changing the target data point obtains a first noise level of all data points on the target spectrum line segment, and changing the target spectrum line segment obtains a first noise level of all data points on all spectrum line segments.

Further, the method for acquiring the second noise level includes:

performing straight line fitting on the marker data points on the target spectrum line segment to obtain corresponding improved fitting straight lines, calculating improved angle differences between the direction angles of vectors between two adjacent data points of the target data points on the target spectrum line segment and the direction angles of the improved fitting straight lines, and obtaining second noise degrees of the target data points according to products of the improved angle differences and distances between the target data points and the improved fitting straight lines;

changing the target data point to obtain a second noise level of all the data points on the target spectrum line segment, and changing the target spectrum line segment to obtain a second noise level of all the data points on the target spectrum line segment.

Further, the method for obtaining the final noise level includes:

obtaining a first noise degree and a second noise degree of a target data point on a spectrum line of a target alpha spectrogram, and calculating the difference between the first noise degree and the second noise degree of the target data point to obtain a noise degree change characteristic value;

Obtaining all similar alpha spectrogram spectral lines of the target alpha spectrogram spectral line, calculating the curve similarity of the target alpha spectrogram spectral line and each similar alpha spectrogram spectral line, counting similar data points corresponding to target data points of the target alpha spectrogram spectral line in each similar alpha spectrogram spectral line, obtaining second noise degree of each similar data point, calculating the difference of the second noise degree between the target data point and each similar data point to obtain similar noise difference, and obtaining a random noise influence value of the target data point according to the sum of products of the similar noise difference corresponding to each similar data point of the target data point, the second noise degree and the curve similarity corresponding to the similar alpha spectrogram spectral line;

obtaining the final noise degree of the target data point according to the sum value of the noise degree change characteristic value and the random noise influence value corresponding to the target data point; and changing the target data point to obtain the final noise degree of each data point on the target alpha spectrogram spectral line, and changing the target alpha spectrogram spectral line to obtain the final noise degree of all the data points.

Further, the method for obtaining the smoothness degree comprises the following steps:

and performing straight line fitting on standard data points on the target spectrum line segment to obtain a corresponding final fitting straight line, calculating the distance variance from all the data points on the target spectrum line segment to the final fitting straight line, calculating the accumulated sum of products between the distance from each data point on the target spectrum line segment to the final fitting straight line and the corresponding final noise degree to obtain the accumulated sum, and performing negative correlation mapping on the product of the distance variance and the accumulated sum to obtain the smoothness degree of the target spectrum line segment.

Further, the method for acquiring the pre-selection filtering window comprises the following steps:

normalizing the smoothness of the target spectrum line segment to obtain a normalized smoothness value; when the normalized smoothness degree value is greater than or equal to a preset smoothness degree threshold value, downward rounding the product of the difference value between the value 1 and the normalized smoothness degree value and a preset adjusting parameter to obtain a preselected maximum filter window size value, wherein the maximum filter window size value is a positive odd number; when the normalized smoothness degree value is smaller than a preset smoothness degree threshold value, upwardly rounding the product of the difference value between the value 2 and the normalized smoothness degree value and a preset adjusting parameter to obtain a preselected maximum filter window size value, wherein the maximum filter window size value is a positive odd number;

screening all positive odd numbers in the preselected maximum filter window size values to obtain all preselected filter window size values of the target spectrum line segment; obtaining all pre-selection filter windows according to the size values of the pre-selection filter windows, wherein the pre-selection filter windows are square windows;

changing the target spectral line segment results in all pre-selected filter windows for each spectral line segment.

Further, the method for obtaining the optimal filtering window comprises the following steps:

Counting all pre-selection filter windows and all data points of a target spectrum line segment, calculating Euclidean distance between a target data point and a nearest extremum point to obtain extremum influence degree, selecting a target pre-selection filter window for filtering by taking the target data point as a center, counting data value differences of the target data points before and after filtering, and carrying out negative correlation mapping on products of the extremum influence degree and the data value differences to obtain a corresponding filtering interference characteristic value when the target data point selects the target pre-selection filter window for filtering;

obtaining a window size value of a target preselection filtering window, carrying out negative correlation mapping on final noise degrees of all data points in the target preselection filtering window, accumulating to obtain final noise degree distribution characteristic values, calculating a product of the corresponding interference characteristic values and the final noise degree distribution characteristic values when the target data points are selected for filtering of the target preselection filtering window, and calculating a ratio of the product to the window size value to obtain a preferable value of the target data points for selecting the target preselection filtering window;

changing a target pre-selection filter window to obtain preferred values of all pre-selection filter windows corresponding to the target spectrum line segment where the target data point is located, and selecting the pre-selection filter window with the maximum preferred value as an optimal filter window of the target data point;

Changing the target data points to obtain the optimal filter window of all the data points in the target spectrum line segment, and changing the target spectrum line segment to obtain all the data pointsAn optimal filtering window for all data points in the spectral line of the spectrogram.

Further, the method for acquiring the difference of the data values of the target data points before and after filtering comprises the following steps:

obtaining a data value before a target data point passes through a target pre-selection filter window, normalizing the final noise degree of each data point in the target pre-selection filter window to obtain normalized final noise degree, calculating the mean value of the product of the mapping value of the normalized final noise degree negative correlation mapping of each data point in the target pre-selection filter window and the corresponding data value, obtaining the data value after the target data point passes through the target pre-selection filter window, and calculating the difference of the data values before and after the target data point passes through the target pre-selection filter window to obtain the data value difference.

Further, the method for acquiring the spectrum segment comprises the following steps:

and obtaining extreme points of the target alpha spectrogram spectral lines, dividing the target alpha spectrogram spectral lines into at least two spectral line segments by taking the extreme points as interval points, and changing the target alpha spectrogram spectral lines to obtain all spectral line segments of each alpha spectrogram spectral line.

The invention has the following beneficial effects:

considering that local nuclear pollution has certain spreading characteristics in the inner wall of a pipeline and alpha energy spectrograms reflect the types and contents of radionuclides, when uranium radioactive pollution occurs in a certain position of the inner wall of the pipeline, the acquired alpha energy spectrograms generally have certain similarity characteristics, so that the embodiment of the invention obtains the final noise degree according to the noise degree change characteristics of data points and the difference characteristics between the spectrum lines of the alpha energy spectrograms and the spectrum lines of the alpha energy spectrograms of the same type, accurately represents that each data point is influenced by noise through the final noise degree, and improves the accuracy in the subsequent selection of a filtering window. And further adaptively selecting the optimal filtering window with the best filtering effect of each data point according to the smoothness degree of each spectrum segment and the final noise degree of each data point. Because the selection of the optimal filter window refers to the final noise degree representing the influence of accurate noise and the change characteristic of the data value of the filtered data point of each pre-selected filter window, the invention can accurately filter noise and avoid filtering effective data when denoising the data based on the optimal filter window, and improves the detection precision of uranium radioactive contamination.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a method for processing uranium radioactive contamination detection data in a pipeline according to an embodiment of the present invention;

fig. 2 is a diagram of an in-pipe uranium contamination automatic detection device for a circular array ZnS (Ag) coated plastic scintillator according to an embodiment of the present invention.

Detailed Description

In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description refers to the specific implementation, structure, characteristics and effects of the method for processing uranium radioactive contamination detection data in a pipeline according to the invention by combining the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

The invention provides a specific scheme of a method for processing uranium radioactive contamination detection data in a pipeline, which is specifically described below with reference to the accompanying drawings.

Referring to fig. 1, a flowchart of a method for processing uranium radioactivity pollution detection data in a pipeline according to an embodiment of the present invention is shown, where the method includes:

step S1: and acquiring an alpha spectrum line of each region of the inner wall of the uranium radioactivity pipeline to be detected.

The invention aims to provide a method for processing uranium radioactive contamination detection data in a pipeline, which is used for accurately removing noise in a data spectral line by processing the data spectral line and improving the detection precision of detected data. The embodiment of the invention aims at the main implementation scenes that: the pipeline robot carries the radiation detector, the rotary positioning system and the decontamination device to directly enter the pipeline, and the collected alpha energy spectrogram is used for measuring and positioning uranium radioactive pollution. It should be noted that, the data processing method provided by the embodiment of the invention is applicable to all implementation scenes similar to the distribution characteristics of uranium radioactivity on the surface of a pipeline, and the embodiment of the invention is mainly used for analyzing and processing uranium radioactivity pollution data. Therefore, firstly, acquiring alpha spectrum lines of each region of the inner wall of the uranium radioactivity pipeline to be detected, and specifically:

Referring to fig. 2, a diagram of an automatic detection device for uranium contamination in a pipeline of an annular array ZnS (Ag) coated plastic scintillator according to an embodiment of the present invention is shown, where the device shown in fig. 2 mainly includes a detergent storage tank, a detergent spray nozzle, an α -energy spectrum measuring instrument, a pipeline robot, and three pollution detectors, where an acquisition tool for acquiring an α -energy spectrum is the pollution detector, and included angles between the three detectors are the same, when the device stays in a certain area inside the pipeline, the three pollution detectors integrally rotate 8 times, and each rotation is 15 degrees to acquire an α -energy spectrum corresponding to all surfaces of an entire annular inner wall of the corresponding area, and further an α -energy spectrum of each area is acquired through movement of the pipeline robot. According to the embodiment of the invention, the alpha energy spectrum of each area of the inner wall of the uranium radioactivity pipeline to be detected is obtained through the annular array ZnS (Ag) coated plastic scintillator pipeline inner uranium contamination automatic detection device shown in fig. 2, the energy of alpha particles of the alpha energy spectrum, namely the pulse amplitude, is taken as an abscissa, the alpha particle number, namely the counting rate, is taken as an ordinate, and a coordinate system is established and the corresponding alpha energy spectrum spectral line is fitted.

Step S2: dividing each alpha spectrogram spectral line into at least two spectral line segments, obtaining a first noise degree of each data point according to the position distribution characteristics of all data points on each spectral line segment, screening each data point according to the first noise degree to obtain a marked data point, and obtaining a second noise degree of all data points according to the position distribution characteristics of the marked data points on each spectral line segment.

Considering that noise tends to be random, there is a certain difference between the data value of the data point affected by noise on the alpha spectrogram spectral line and the real data value, and the greater the degree of influence by noise, the greater the difference between the data value and the real data value. Further, in order to make the characteristic shown by each data point on the α spectrum line affected by noise clearer, the entire α spectrum line is divided into a plurality of spectrum line segments, if the data at a certain position is affected by noise, the numerical value of the data point at a corresponding position will change, and further the position on the curve will also change, and according to the situation that each data point obtained primarily from the position distribution of the data point on each spectrum line segment is affected by noise. According to the embodiment of the invention, each alpha spectrogram spectral line is divided into at least two spectral line segments, and the first noise degree of each data point is obtained according to the position distribution characteristics of all data points on each spectral line segment.

Preferably, the method for acquiring the spectrum segment comprises the following steps:

and obtaining extreme points of spectral lines of the target alpha spectrogram, wherein the extreme points are wave peak points and wave valley points on the spectral lines. It should be noted that, the method for obtaining the extremum point of the spectral line is well known in the art, and is not further limited and described herein.

Dividing the target alpha spectrogram spectral line into at least two spectral line segments by taking the extreme point as an interval point, and changing the target alpha spectrogram spectral line to obtain all spectral line segments of each alpha spectrogram spectral line. The alpha spectrogram spectral line is divided according to the extreme points, so that certain trend exists in the data values of the data points on each divided spectral line segment, and the corresponding noise influence degree of each data point is more obvious when the data points are influenced by noise.

Preferably, the method for acquiring the first noise level includes:

and carrying out straight line fitting on the data points on the target spectrum line segment to obtain corresponding fitting straight lines. In the embodiment of the invention, the corresponding fitting straight line is obtained through the least square method according to the data points on the target spectral line segment, the fitting straight line can be closer to the spectral line segment without noise through the least square method, and the overall error is smaller. It should be noted that, the least square method is a prior art well known to those skilled in the art, and is not further defined and described herein.

If the target data point is not an endpoint:

the angular difference between the directional angle of the vector between two adjacent data points of the target data point on the target spectral line segment and the directional angle of the fitted line is calculated. The deviation degree of the target data point relative to the fitting straight line can be represented in the aspect of the angle difference, and when the angle difference corresponding to the target data point is larger, the greater the deviation degree of the target data point is, the more likely the data point is noise, namely the greater the first noise degree is.

Obtaining a first noise degree of the target data point according to the product of the angle difference and the distance from the target data point to the fitting straight line; changing the target data point obtains a first noise level of all data points on the target spectrum line segment, and changing the target spectrum line segment obtains a first noise level of all data points on all spectrum line segments. Combining the distance of the target data point to the fitted line on the basis of the angle difference can enable the characterization of the first noise level of the target data point to be more accurate. The greater the deviation of the target data point, the more likely it is that the data point will be noisy, i.e., the greater the first noise level, when the target data point is further from the fitted line.

The first noise level acquisition method when the target data point is not an endpoint is expressed as:

wherein, the liquid crystal display device comprises a liquid crystal display device,data pointsIs used to determine the first noise level of the (c) signal,data pointsData points adjacent to the latter on the corresponding spectral line segmentVector betweenThe corresponding direction angle is used for the direction of the light beam,data pointsData points adjacent to the previous one on the corresponding spectral line segmentVector betweenThe corresponding direction angle is used for the direction of the light beam,data pointsCorresponding simulationStraight line of fusionIs provided with a plurality of directional angles,data points To a corresponding fitting straight lineIs a distance of (3).Data pointsCorresponding angle differences, overall characterization data pointsDegree of deviation in the angular direction with respect to the fitted straight line; data pointsThe greater the corresponding angle difference is,the larger the corresponding data pointThe greater the first noise level.Characterization data pointsDegree of deviation in distance from the fitted line, as data pointsThe farther the distance to the corresponding fitted line, the description data pointIs the first noise of (1)The greater the degree.

If the target data point is the endpoint:

calculating the angle difference between the direction angle of the vector between the adjacent data points of the target data point on the target spectrum line segment and the direction angle of the fitting straight line; and obtaining the first noise degree of the target data point according to the product of the two times of the angle difference and the distance between the target data point and the fitting straight line. When the target data point is the starting end point, the vector points from the target data point to the next data point; when the target data point is a termination endpoint, the vector points from the previous data point to the target data point. It should be noted that, the principle of acquiring the first noise level when the target data point is the endpoint and the target data point is not the endpoint is the same, and further description is omitted herein.

Since the first noise level is obtained based on the position distribution characteristics of all the data points during the acquisition, the corresponding first noise level is affected by the position distribution characteristics of the noise data points, so that the corresponding first noise level is not accurate enough, and therefore, the obtained first noise level needs to be further optimized. According to the embodiment of the invention, each data point is screened according to the first noise level to obtain a marked data point, the marked data point is used as a reference, and the second noise level of all the data points is obtained according to the position distribution characteristics of the marked data point on each spectrum line segment.

And taking the data points with the first noise degree smaller than the preset noise degree threshold value as marker data points, and taking the data points with the first noise degree larger than or equal to the preset noise degree threshold value as suspected noise data points. When the first noise degree of the data point is larger than or equal to a preset noise degree threshold value, the deviation degree of the corresponding data point relative to the overall numerical trend on the corresponding spectrum line segment is larger, and the possibility of corresponding noise is larger; when the first noise level of the data point is smaller than or equal to the preset noise level threshold, the deviation degree of the corresponding data point relative to the whole numerical value area on the spectrum line segment is smaller, and the probability of the corresponding data point being noise is smaller. Therefore, the first noise level is further improved according to the position distribution characteristics of the marker data points on each spectrum segment, and the improved second noise level can be more accurate. In the embodiment of the invention, the preset noise level threshold is set to 0.8. It should be noted that, the implementer may select the preset noise level threshold according to the specific implementation situation, which will not be further described herein.

Preferably, the method for acquiring the second noise level includes:

and performing straight line fitting on the marker data points on the target spectrum line segment to obtain corresponding improved fitting straight lines. The improved fitting line is closer to the spectral line segment when noise is absent than the fitting line before improvement, and data points with higher noise probability are eliminated in the fitting process, and the data points are less affected by noise. According to the embodiment of the invention, the corresponding improved fitting straight line is obtained by performing straight line fitting on the marker data points on the target spectrum line segment through a least square method.

And calculating an improved angle difference between the direction angle of the vector of the target data point between two adjacent data points on the target spectrum line segment and the direction angle of the improved fitting straight line, and obtaining a second noise degree of the target data point according to the product of the improved angle difference and the distance from the target data point to the improved fitting straight line. Changing the target data point to obtain a second noise level of all the data points on the target spectrum line segment, and changing the target spectrum line segment to obtain a second noise level of all the data points on the target spectrum line segment. It should be noted that, the second noise level obtaining process only changes the position of the fitting straight line relative to the first noise level obtaining process, and there is no difference in the obtaining method, and the meaning of each parameter in the second noise level obtaining process is not further described herein.

Step S3: obtaining similar alpha spectrogram spectral lines of each alpha spectrogram spectral line, obtaining data point noise degree change characteristics according to the difference of the second noise degree and the first noise degree of each data point, and obtaining final noise degree according to the noise degree change characteristics of the data points and the difference characteristics between the alpha spectrogram spectral lines where the data points are positioned and the similar alpha spectrogram spectral lines; screening each data point according to the final noise degree to obtain standard data points, obtaining the smoothness degree according to the final noise degree of the data points on each spectrum segment and the position distribution characteristics of the standard data points, obtaining preselected filter windows with different sizes according to the smoothness degree of each spectrum segment, and selecting the optimal filter window corresponding to each data point according to the final noise degree distribution characteristics of the data points in each preselected filter window of each spectrum segment and the data value change characteristics of the filtered data points.

The first noise level and the second noise level of the data points are obtained through the step S2, but the first noise level and the second noise level of all the data points on each alpha spectrogram spectral line are obtained according to the data points, and the accurate noise level of each data point affected by random noise cannot be accurately obtained. Considering that the uranium radioactive pollution detected by the embodiment of the invention has certain spreading characteristics, namely when the uranium radioactive pollution occurs on the inner wall of the pipeline, the alpha energy spectrogram reflects the types and the contents of radionuclides, all the similar alpha energy spectrograms acquired at the position corresponding to the occurrence of the uranium radioactive pollution generally have certain similarity characteristics, and the corresponding similarity is generally high when the uranium radioactive pollution is not influenced by noise. Therefore, according to the embodiment of the invention, the real noise degree of each data point affected by random noise can be obtained by combining the first noise degree and the second noise degree of each data according to the similarity characteristics between each alpha spectrogram and the corresponding similar alpha spectrograms.

Firstly, the similar alpha spectrum lines of each alpha spectrum line need to be obtained. Since the embodiment of the invention collects the alpha energy spectrogram through the annular array ZnS (Ag) coated plastic scintillator in-pipeline uranium contamination automatic detection device shown in fig. 2, when each region of the inner wall of the pipeline is detected, 8 alpha energy spectrograms can be obtained by each detector, namely 24 alpha energy spectrograms are corresponding to each region. Further, 8 alpha energy spectrograms obtained by each detector in each region are used as similar energy spectrograms, namely, each alpha energy spectrogram corresponding to the embodiment of the invention can obtain 7 similar alpha energy spectrograms, and each corresponding alpha energy spectrogram spectral line can obtain 7 similar alpha energy spectrogram spectral lines.

Further, according to the difference between the second noise degree and the first noise degree of each data point, the noise degree change characteristics of the data points are obtained, and according to the noise degree change characteristics of the data points and the difference characteristics between the alpha spectrogram spectral lines where the data points are located and the alpha spectrogram spectral lines of the same kind, the final noise degree is obtained.

Preferably, the method for obtaining the final noise level includes:

and obtaining the first noise degree and the second noise degree of the target data point on the spectrum line of the target alpha spectrogram, and calculating the difference between the first noise degree and the second noise degree of the target data point to obtain a noise degree change characteristic value. The larger the noise level change characteristic value, the lower the accuracy corresponding to the first noise level. For normal data points which are not affected by noise, the corresponding first noise degree and second noise degree are smaller, and the corresponding noise degree change characteristic value is smaller; for misjudging noise points, namely, misjudging the noise points as normal points and data points of which the normal points are misjudged as noise points according to the first noise degree, the corresponding noise degree change characteristic value is larger; for judging the correct noise point, namely, the data point which can be judged as the noise point according to the first noise degree and the second noise degree, the corresponding first noise degree and second noise degree are larger, and the noise change characteristic value is smaller. Further, in the subsequent process, the final noise degree is calculated by adding the noise degree change characteristic value, so that the obtained final noise degree contains the data point misjudgment information, and the misjudgment condition of the data point is effectively reduced.

And obtaining all similar alpha spectrum lines of the target alpha spectrum line, and calculating the curve similarity between the target alpha spectrum line and each similar alpha spectrum line. In the calculation process of the final noise degree, the curve similarity is added, so that the influence of the alpha spectrogram spectral lines with lower similarity on the final noise degree in the similar alpha spectrogram spectral lines can be reduced, and the calculation error is reduced. It should be noted that, the calculation of the similarity of curves between the spectral lines is well known in the art, and is not further limited and described herein.

And counting similar data points in the similar alpha spectrogram spectral lines at positions corresponding to the target data points of the target alpha spectrogram spectral lines, obtaining the second noise degree of each similar data point, and calculating the difference of the second noise degree between the target data point and each similar data point to obtain similar noise difference. Because the noise has randomness, when the target data point on the spectrum line of the target alpha spectrogram is a noise point, the data point at the corresponding position on the spectrum line of the same kind alpha spectrogram is often a normal point, and the corresponding noise degree has a difference, namely, when the difference between the target data point and the data point at the corresponding position on each spectrum line of the same kind alpha spectrogram is larger, the noise difference of the same kind is larger, which means that the influence of random noise on the target data point is larger.

Obtaining a random noise influence value of the target data point according to the accumulated sum of products of the similar noise difference corresponding to each similar data point of the target data point, the second noise degree and the curve similarity corresponding to the spectrum line of the similar alpha spectrogram; the random noise influence value introduces a difference between a target data point and a data point at a corresponding position on the spectrum line of the alpha spectrogram of the same type on the basis of the second noise degree. The influence of random noise is reduced through the difference between the alpha spectrogram spectral lines with certain similar characteristics to the alpha spectrogram spectral lines where the target data points are located, so that the calculated final noise degree represents the real noise degree more accurately. And when the curve similarity of the similar alpha spectrogram spectral lines corresponding to the alpha spectrogram spectral lines where the target data points are positioned is higher, the similar noise difference is larger, and the second noise degree is larger, the noise affected degree of the target data points is larger, namely the random noise affected value is larger.

Obtaining the final noise degree of the target data point according to the sum value of the noise degree change characteristic value and the random noise influence value corresponding to the target data point; and changing the target data point to obtain the final noise degree of each data point on the target alpha spectrogram spectral line, and changing the target alpha spectrogram spectral line to obtain the final noise degree of all the data points. The final noise degree is obtained through the noise degree change characteristic value and the random noise influence value corresponding to the target data point, so that the false judgment condition of the data point can be effectively reduced, and meanwhile, the representation of the real noise degree is more accurate.

The method for obtaining the final noise level is expressed as the following formula:

wherein, the liquid crystal display device comprises a liquid crystal display device,as the final noise level of the target data point,alpha spectrum line of alpha spectrogram where target data point isThe first degree of noise obtained above is then,alpha spectrum line of alpha spectrogram where target data point isThe second degree of noise obtained above is used,for alpha spectrum linesAnd correspond to the firstCurve similarity between spectrum lines of the same class of alpha spectrograms,for the alpha spectrum line at which the target data point is locatedCorresponding firstAnd a second noise level of the data points at corresponding positions on the spectrum lines of the same class alpha spectrogram.Is the alpha energy spectrum line of the target pixel pointCorresponding number of alpha spectrum lines of the same kind, in the embodiment of the invention, the alpha spectrum line where the target pixel point is locatedNumber of corresponding alpha profile lines of the same class7.

Thus, the final noise degree of all the data points is obtained. And further acquiring the interference degree of the corresponding spectral line segment by noise according to the final noise degree of all the data points on each spectral line segment, and selecting a filtering window with a proper size according to the interference degree of the noise of each spectral line segment for filtering to finish the denoising of the spectral line. According to the embodiment of the invention, each data point is screened according to the final noise degree to obtain the standard data point, and the smoothness degree is obtained according to the final noise degree of the standard data point on each spectrum line segment and the position distribution characteristics of the standard data point. The purpose of obtaining the smoothness of each spectral line segment is to select a proper filtering window for filtering, so that the denoising effect is improved, and effective data are prevented from being filtered or real noise data are prevented from being omitted. When the smoothing degree of the corresponding spectrum line segment is larger, the interference caused by noise is smaller, and a smaller window is selected for filtering when filtering is carried out; when the smoothing degree corresponding to the spectral line segment is smaller, the interference caused by noise is larger, and a larger window is selected for filtering when filtering is performed so as to weaken the interference of the noise. The embodiment of the invention obtains the preselected filter windows with different sizes according to the smoothness degree of each spectrum segment.

In the embodiment of the invention, the data points with the final noise degree smaller than the preset noise threshold value are recorded as standard data points, the standard data points are points which are less or not affected by noise, the data value trend represented by the position distribution characteristics of the standard data points on the corresponding spectral line segments can be maximally close to the real data value trend, and the interference degree of noise on each spectral line segment is further represented according to the calculated corresponding smoothness degree.

Preferably, the method for acquiring the smoothness degree includes:

and performing straight line fitting on the standard data points on the target spectrum line segment to obtain a corresponding final fitting straight line. According to the embodiment of the invention, the standard data points on the target spectrum line segment are subjected to line fitting by a least square method to obtain a final fitting line. The straight line corresponding to the final fitting straight line can be maximally close to the spectral line segment without noise influence, and can represent the standard of the data value corresponding to each data point on the spectral line segment.

And calculating the distance variance from all data points on the target spectrum line segment to the final fitting straight line. The deviation of the data value of each data point from the real data value can be represented by the distance from each data point to the final fitting straight line on the spectrum line segment, and the larger the corresponding distance of the data point is, the smoother the spectrum line is. The integral deviation distance condition of the data points on the spectrum line segment can be represented through the distance variance, and the larger the distance variance of the data points on the integral spectrum line segment is, the smaller the spectrum line segment is close to linearity, and the smaller the smoothing degree of the corresponding spectrum line segment is.

And calculating an accumulated sum of products between the distance from each data point on the target spectrum segment to the final fitting straight line and the corresponding final noise degree to obtain the accumulated sum, and carrying out negative correlation mapping on the product of the distance variance and the accumulated sum to obtain the smoothness degree of the target spectrum segment. The accumulated sum can represent the degree of influence of noise on the whole data point on the target spectrum segment, and further represents the smoothness degree of the target spectrum segment through the product of the distance variance of the data point and the degree of influence of noise on the whole data point, so that the degree of influence of noise on the target spectrum segment represented by the obtained smoothness degree is more accurate. The larger the corresponding final noise degree of the data points on the target spectrum line segment is, the longer the distance from the final fitting straight line is, the larger the distance variance is, and the more serious the influence of noise on the data points on the target spectrum line segment is, the smaller the smoothness degree is.

The method for obtaining the smoothness is expressed as follows:

wherein, the liquid crystal display device comprises a liquid crystal display device,for spectral line segmentsA corresponding degree of smoothness is provided for the user,for spectral line segmentsFirst, theThe final noise level of the individual data points,for spectral line segmentsFirst, theData point to spectral line segmentThe distance between the corresponding final fit lines, For spectral line segmentsThe variance of the distances of all data points to the corresponding final fit line,for spectral line segmentsThe number of data points to be added,is an exponential function with a base of natural constant.

Preferably, the method for acquiring the pre-selection filter window comprises the following steps:

and normalizing the smoothness of the target spectrum line segment to obtain a normalized smoothness value. The normalization aims to unify the numerical values of subsequent calculation, and facilitates the selection of a preset smoothness threshold value and a preset adjustment parameter.

When the normalized smoothness degree value is greater than or equal to a preset smoothness degree threshold value, the product of the difference value between the value 1 and the normalized smoothness degree value and a preset adjusting parameter is downwards rounded to obtain a preselected maximum filter window size value, wherein the maximum filter window size value is a positive odd number; and when the normalized smoothness degree value is smaller than a preset smoothness degree threshold value, upwardly rounding the product of the difference value between the value 2 and the normalized smoothness degree value and a preset adjusting parameter to obtain a preselected maximum filter window size value, wherein the maximum filter window size value is a positive odd number. According to different corresponding values of the smoothness values, smaller filter windows are selected for spectral line segments with larger smoothness values, and larger filter windows are selected for spectral line segments with smaller smoothness values. It should be noted that, the shape of all the filtering windows related to the present invention is only square, i.e. the length and width of the corresponding windows are consistent, i.e. the corresponding window sizes are all n×n.

The method for obtaining the preselected maximum filter window size value is expressed in terms of a formula:

wherein, the liquid crystal display device comprises a liquid crystal display device,for spectral line segmentsThe corresponding pre-selected maximum filter window size,for spectral line segmentsCorresponding normalized smoothness values;for the preset smoothness threshold, the preset smoothness threshold is set to 0.6 in the embodiment of the present invention;for the preset adjustment parameter, the preset adjustment parameter is set to 15 in the embodiment of the present invention.To take the odd integer function down on the value in brackets,to take an odd integer up to the value in brackets.

Screening all positive odd numbers in all values smaller than or equal to the preselected maximum filter window size value to obtain all preselected filter window size values of the target spectrum line segment; obtaining all pre-selection filter windows according to the size values of the pre-selection filter windows, wherein the pre-selection filter windows are square windows; changing the target spectral line segment results in all pre-selected filter windows for each spectral line segment. For example, when a line segmentCorresponding preselected maximum filter window size valueWhen the size is 9, the size of the preselected maximum filtering window is 9 multiplied by 9, and the corresponding spectral line segmentIs 3 x 3, 5 x 5, 7 x 7 and 9 x 9. The preselection filter window of each spectrum segment is calculated, the selection of the size of the local optimal filter window corresponding to the data point on each spectrum segment can be limited in a certain range, and the denoising effect of the optimal filter window calculated in the subsequent process is improved while the calculated amount is reduced.

After all pre-selected filter windows corresponding to each spectral line segment are obtained, further analyzing data points in each spectral line segment to obtain an optimal filter window of each data point, and denoising the whole alpha spectrogram spectral line according to the optimal filter window of each data point to achieve the aim of improving the denoising effect of the alpha spectrogram spectral line. Considering that the denoising effect of the filtering windows with different sizes on each data point is different and different data points are affected by noise, the embodiment of the invention selects the optimal filtering window corresponding to each data point according to the final noise degree distribution characteristic of the data points in each pre-selected filtering window of each spectrum line segment and the data value change characteristic of the filtered data points.

Preferably, the method for acquiring the optimal filtering window includes:

and counting all pre-selected filter windows and all data points of the target spectrum line segment, and calculating the Euclidean distance between the target data point and the nearest extremum point to obtain the extremum influence degree. And selecting a target pre-selection filtering window for filtering by taking the target data point as the center, counting the data value difference of the target data points before and after filtering, and carrying out negative correlation mapping on the product of the extreme value influence degree and the data value difference to obtain a corresponding filtering interference characteristic value when the target data point is selected for filtering by the target pre-selection filtering window.

The extremum influence degree can represent the influence degree of each data point on the wave crest and the wave trough when the data points are filtered through the filtering window, and further, the negative correlation mapping is carried out according to the product of the extremum influence degree and the data value difference, so that the variation difference of the data points close to the extremum point before and after the filtering is reduced when the obtained window with high optimal value is filtered, the characteristics of the wave crest and the wave trough on the spectrum line of the alpha spectrogram are reserved, and the deformation phenomenon caused by the interference of the filtering on the wave crest and the wave trough is reduced. The greater the filter interference characteristic value, the greater the preferred value of the corresponding target preselected filter window.

Preferably, the method for acquiring the difference of the data values of the target data points before and after filtering comprises the following steps:

the method comprises the steps of obtaining a data value before a target data point passes through a target pre-selection filter window, normalizing the final noise degree of each data point in the target pre-selection filter window to obtain normalized final noise degree, calculating the mean value of the product of the mapping value of the normalized final noise degree of each data point in the target pre-selection filter window and the corresponding data value, obtaining the data value after the target data point passes through the target pre-selection filter window, and calculating the difference of the data values before and after the target data point passes through the target pre-selection filter window to obtain the difference of the data values. According to the embodiment of the invention, the degree of influence of noise on each data point in the filtering window is considered when the data points are filtered through the filtering window, and the data points with high noise degree have small influence on the filtered data value based on the final noise degree of the data points in the filtering window as the weight of the data value when the average filtering is carried out, so that the influence degree of noise on the data value filtered through the filtering window is reduced to a certain extent, and the denoising effect of the spectral line of the integral alpha spectrogram is improved.

The method for acquiring the data value filtered by the filter window is expressed as the following formula:

wherein, the liquid crystal display device comprises a liquid crystal display device,for data values filtered through the target filter window,for the number of data points within the target filter window,filter the inside of window for the targetThe normalized final noise level corresponding to the data point,filter the inside of window for the targetData values corresponding to the data points.

Obtaining a window size value of a target pre-selection filtering window, and accumulating the final noise degree negative correlation mapping of all data points in the target pre-selection filtering window to obtain a final noise degree distribution characteristic value; the final noise degree distribution characteristic value can represent the distribution condition of the final noise degree of all noise points in the target pre-selection filtering window, and the embodiment of the invention enables the final noise degree to be inversely proportional to the optimal value of the target pre-selection filtering window through the method of accumulation after negative correlation mapping, and when the noise degree in the target pre-selection filtering window is larger, the filtering effect of the window is poorer. Because the more severely the data points in the target pre-selected filter window are affected by noise, the more the filtered data values of the corresponding data points are affected by noise, the less trustworthy the corresponding data values.

Calculating the product of the corresponding interference characteristic value and the final noise degree distribution characteristic value when the target data point selects the target preselection filtering window for filtering, and calculating the ratio of the product to the window size value to obtain the optimal value of the target data point selecting the target preselection filtering window; finally, the optimal value is limited through the size of the target preselection filtering window, and the window size of the target preselection filtering window is taken as a denominator in the embodiment of the invention, so that the window size and the optimal value are in inverse relation in consideration of the fact that the alpha spectrogram spectral line after filtering is deformed when the target preselection filtering window is overlarge, and further the follow-up analysis is inaccurate.

The method of obtaining the preferred value of the preselected filter window is expressed in terms of the equation:

wherein, the liquid crystal display device comprises a liquid crystal display device,to be based on data pointsPreselecting the filter window for the center of the filter windowIs used as a reference to the preferred value of (c),data pointsThe euclidean distance between the alpha spectrum line and the nearest extreme point,data pointsThrough a pre-selection filter windowIs used to filter the data values before the filtering,data pointsThrough a pre-selection filter windowIs used to determine the data value of the data,for preselecting filter windowsThe corresponding window size value is used to determine,to be based on data pointsPreselecting the filter window for the center of the filter window Inner firstThe final noise level of the individual data points,to be based on data pointsPreselecting the filter window for the center of the filter windowThe number of internal data points;in the embodiment of the invention, the preset parameters are the preset parameters0.01, in order to prevent denominator from being 0;is an exponential function with a base of natural constant. It should be noted that, the method for calculating the euclidean distance is known in the prior art by those skilled in the art, and is not further limited and described herein.

Changing a target pre-selection filter window to obtain preferred values of all pre-selection filter windows corresponding to the target spectrum line segment where the target data point is located, and selecting the pre-selection filter window with the maximum preferred value as an optimal filter window of the target data point; changing the target data points to obtain the optimal filter windows of all the data points in the target spectrum line segment, and changing the target spectrum line segment to obtain the optimal filter windows of all the data points in all the alpha spectrogram spectral lines.

Step S4: and (3) filtering by using an optimal filter window according to each data point to complete denoising of each alpha spectrogram spectral line, and detecting uranium radioactive pollution according to an alpha spectrogram corresponding to the denoised alpha spectrogram spectral line.

After the optimal filter window of each data point of the alpha spectrogram spectral line is obtained, the optimal filter window is selected according to each data point to filter so as to complete denoising of each alpha spectrogram spectral line, and uranium radioactive pollution detection is carried out according to the alpha spectrogram corresponding to the denoised alpha spectrogram spectral line. According to the embodiment of the invention, the data points in the alpha spectrum collected by the detector are denoised through the corresponding optimal filter windows, the denoised alpha spectrum is converted into the denoised alpha spectrum, and uranium contamination identification is carried out on the denoised alpha spectrum to carry out uranium radioactive contamination detection. It should be noted that, the process of uranium contamination identification according to the α spectrogram and α spectrogram spectral line conversion of the α spectrogram according to the embodiment of the present invention is known to those skilled in the art, and is not further limited and described herein.

The present invention has been completed.

In summary, the invention analyzes the alpha spectrogram, segments the spectral lines, obtains the final noise degree based on the data point position distribution characteristics on different segments of spectral lines and the difference between the similar alpha spectrograms of the corresponding alpha spectrogram, obtains the preselected filter window according to the distribution condition of the final noise degree of the data points on each spectral line segment, selects the optimal filter window of each data according to the preselected filter window when filtering different data points, and completes the denoising of each alpha spectrogram spectral line by selecting the optimal filter window according to each data point. According to the method for selecting the optimal filtering window for each data point, noise can be accurately filtered, effective data can be prevented from being filtered, and the uranium radioactive pollution detection precision is improved.

It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.

Claims (2)

1. A method for processing uranium radioactive contamination detection data in a pipeline, the method comprising:

acquiring alpha spectrum lines of each region of the inner wall of the uranium radioactivity pipeline to be detected;

dividing each alpha spectrogram spectral line into at least two spectral line segments, obtaining a first noise degree of each data point according to the position distribution characteristics of all data points on each spectral line segment, screening each data point according to the first noise degree to obtain a marked data point, and obtaining a second noise degree of all data points according to the position distribution characteristics of the marked data points on each spectral line segment;

obtaining similar alpha spectrogram spectral lines of each alpha spectrogram spectral line, obtaining data point noise degree change characteristics according to the difference of the second noise degree and the first noise degree of each data point, and obtaining final noise degree according to the noise degree change characteristics of the data points and the difference characteristics between the alpha spectrogram spectral lines where the data points are positioned and the similar alpha spectrogram spectral lines; screening each data point according to the final noise degree to obtain standard data points, obtaining a smooth degree according to the final noise degree of the data points on each spectrum segment and the position distribution characteristics of the standard data points, obtaining preselected filter windows with different sizes according to the smooth degree of each spectrum segment, and selecting an optimal filter window corresponding to each data point according to the final noise degree distribution characteristics of the data points in each preselected filter window of each spectrum segment and the data value change characteristics of the filtered data points;

According to each data point, an optimal filter window is selected for filtering to complete denoising of each alpha spectrum line, and uranium radioactive pollution detection is carried out according to an alpha spectrum corresponding to the denoised alpha spectrum line;

the method for acquiring the spectrum line segment comprises the following steps:

obtaining extreme points of the target alpha spectrogram spectral lines, dividing the target alpha spectrogram spectral lines into at least two spectral line segments by taking the extreme points as interval points, and changing the target alpha spectrogram spectral lines to obtain all spectral line segments of each alpha spectrogram spectral line;

the method for acquiring the first noise degree comprises the following steps:

performing straight line fitting on the data points on the target spectrum line segment to obtain a corresponding fitting straight line;

if the target data point is not the end point, calculating the angle difference between the vector direction angle of the two adjacent data points of the target data point on the target spectrum line segment and the direction angle of the fitting straight line; obtaining a first noise degree of the target data point according to the product of the angle difference and the distance between the target data point and the fitting straight line; the vectors include a front vector that points from a previous data point to a target data point and a rear vector that points from the target data point to a subsequent data point;

If the target data point is an endpoint, calculating the angle difference between the direction angle of the vector between the adjacent data points of the target data point on the target spectrum line segment and the direction angle of the fitting straight line; obtaining a first noise degree of the target data point according to the product of the two times of the angle difference and the distance from the target data point to the fitting straight line; when the target data point is the starting end point, the vector points from the target data point to the next data point; when the target data point is a termination endpoint, the vector points from the previous data point to the target data point;

changing the target data points to obtain first noise degrees of all the data points on the target spectrum line segment, and changing the target spectrum line segment to obtain first noise degrees of all the data points on all the spectrum line segment;

the second noise degree obtaining method comprises the following steps:

performing straight line fitting on the marker data points on the target spectrum line segment to obtain corresponding improved fitting straight lines, calculating improved angle differences between the direction angles of vectors between two adjacent data points of the target data points on the target spectrum line segment and the direction angles of the improved fitting straight lines, and obtaining second noise degrees of the target data points according to products of the improved angle differences and distances between the target data points and the improved fitting straight lines;

Changing the target data point to obtain second noise degrees of all the data points on the target spectrum line segment, and changing the target spectrum line segment to obtain second noise degrees of all the data points on all the spectrum line segment;

the method for acquiring the final noise degree comprises the following steps:

obtaining a first noise degree and a second noise degree of a target data point on a spectrum line of a target alpha spectrogram, and calculating the difference between the first noise degree and the second noise degree of the target data point to obtain a noise degree change characteristic value;

obtaining all similar alpha spectrogram spectral lines of the target alpha spectrogram spectral line, calculating the curve similarity of the target alpha spectrogram spectral line and each similar alpha spectrogram spectral line, counting similar data points corresponding to target data points of the target alpha spectrogram spectral line in each similar alpha spectrogram spectral line, obtaining second noise degree of each similar data point, calculating the difference of the second noise degree between the target data point and each similar data point to obtain similar noise difference, and obtaining a random noise influence value of the target data point according to the sum of products of the similar noise difference corresponding to each similar data point of the target data point, the second noise degree and the curve similarity corresponding to the similar alpha spectrogram spectral line;

Obtaining the final noise degree of the target data point according to the sum value of the noise degree change characteristic value and the random noise influence value corresponding to the target data point; changing the target data points to obtain the final noise degree of each data point on the spectrum line of the target alpha spectrogram, and changing the spectrum line of the target alpha spectrogram to obtain the final noise degree of all the data points;

the method for acquiring the smoothness comprises the following steps:

performing straight line fitting on standard data points on a target spectrum line segment to obtain a corresponding final fitting straight line, calculating the distance variance from all data points on the target spectrum line segment to the final fitting straight line, calculating the accumulated sum of products between the distance from each data point on the target spectrum line segment to the final fitting straight line and the corresponding final noise degree to obtain the accumulated sum, and performing negative correlation mapping on the product of the distance variance and the accumulated sum to obtain the smoothness degree of the target spectrum line segment;

the method for acquiring the pre-selection filtering window comprises the following steps:

normalizing the smoothness of the target spectrum line segment to obtain a normalized smoothness value; when the normalized smoothness degree value is greater than or equal to a preset smoothness degree threshold value, downward rounding the product of the difference value between the value 1 and the normalized smoothness degree value and a preset adjusting parameter to obtain a preselected maximum filter window size value, wherein the maximum filter window size value is a positive odd number; when the normalized smoothness degree value is smaller than a preset smoothness degree threshold value, upwardly rounding the product of the difference value between the value 2 and the normalized smoothness degree value and a preset adjusting parameter to obtain a preselected maximum filter window size value, wherein the maximum filter window size value is a positive odd number;

Screening all positive odd numbers in the preselected maximum filter window size values to obtain all preselected filter window size values of the target spectrum line segment; obtaining all pre-selection filter windows according to the size values of the pre-selection filter windows, wherein the pre-selection filter windows are square windows;

changing the target spectrum line segment to obtain all preselected filtering windows of each spectrum line segment;

the method for acquiring the optimal filtering window comprises the following steps:

counting all pre-selection filter windows and all data points of a target spectrum line segment, calculating Euclidean distance between a target data point and a nearest extremum point to obtain extremum influence degree, selecting a target pre-selection filter window for filtering by taking the target data point as a center, counting data value differences of the target data points before and after filtering, and carrying out negative correlation mapping on products of the extremum influence degree and the data value differences to obtain a corresponding filtering interference characteristic value when the target data point selects the target pre-selection filter window for filtering;

obtaining a window size value of a target preselection filtering window, carrying out negative correlation mapping on final noise degrees of all data points in the target preselection filtering window, accumulating to obtain final noise degree distribution characteristic values, calculating a product of the corresponding interference characteristic values and the final noise degree distribution characteristic values when the target data points are selected for filtering of the target preselection filtering window, and calculating a ratio of the product to the window size value to obtain a preferable value of the target data points for selecting the target preselection filtering window;

Changing a target pre-selection filter window to obtain preferred values of all pre-selection filter windows corresponding to the target spectrum line segment where the target data point is located, and selecting the pre-selection filter window with the maximum preferred value as an optimal filter window of the target data point;

changing the target data point to obtain optimal filtering of all data points in the target spectrum line segmentWindow, changing target spectrum line segment to obtain allAn optimal filtering window for all data points in the spectral line of the spectrogram.

2. The method for processing uranium radioactive contamination detection data in a pipeline according to claim 1, wherein the method for acquiring a difference in data values of target data points before and after filtering includes:

obtaining a data value before a target data point passes through a target pre-selection filter window, normalizing the final noise degree of each data point in the target pre-selection filter window to obtain normalized final noise degree, calculating the mean value of the product of the mapping value of the normalized final noise degree negative correlation mapping of each data point in the target pre-selection filter window and the corresponding data value, obtaining the data value after the target data point passes through the target pre-selection filter window, and calculating the difference of the data values before and after the target data point passes through the target pre-selection filter window to obtain the data value difference.