CN118366563A - Indoor air quality rapid detection method based on artificial intelligence - Google Patents
Indoor air quality rapid detection method based on artificial intelligence Download PDFInfo
- Publication number
- CN118366563A CN118366563A CN202410788850.6A CN202410788850A CN118366563A CN 118366563 A CN118366563 A CN 118366563A CN 202410788850 A CN202410788850 A CN 202410788850A CN 118366563 A CN118366563 A CN 118366563A
- Authority
- CN
- China
- Prior art keywords
- intensity
- data set
- indoor air
- value
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 22
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 claims description 36
- 230000035772 mutation Effects 0.000 claims description 29
- 230000008569 process Effects 0.000 claims description 16
- 238000009499 grossing Methods 0.000 claims description 11
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000010521 absorption reaction Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 238000000638 solvent extraction Methods 0.000 claims 1
- 230000000875 corresponding effect Effects 0.000 description 72
- 238000001228 spectrum Methods 0.000 description 10
- 230000003044 adaptive effect Effects 0.000 description 9
- 230000008859 change Effects 0.000 description 7
- WSFSSNUMVMOOMR-UHFFFAOYSA-N Formaldehyde Chemical compound O=C WSFSSNUMVMOOMR-UHFFFAOYSA-N 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000002329 infrared spectrum Methods 0.000 description 3
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000011088 calibration curve Methods 0.000 description 1
- 239000001569 carbon dioxide Substances 0.000 description 1
- 229910002092 carbon dioxide Inorganic materials 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000010238 partial least squares regression Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/20—Identification of molecular entities, parts thereof or of chemical compositions
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3504—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing gases, e.g. multi-gas analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H21/00—Adaptive networks
- H03H21/0012—Digital adaptive filters
- H03H21/0043—Adaptive algorithms
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Crystallography & Structural Chemistry (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Immunology (AREA)
- Databases & Information Systems (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Pathology (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention relates to the technical field of air quality detection, in particular to an indoor air quality rapid detection method based on artificial intelligence. Firstly, determining local noise intensity expression weight and noise intensity based on a base line of intensity data in an indoor air spectrogram, and determining characteristic expression intensity based on the local noise intensity expression weight, a peak value in a curve of intensity and wavelength and a wavelength corresponding to the peak value; and then the noise intensity and the characteristic expression intensity are used for carrying out self-adaptive adjustment on the order of the SG filter to obtain an accurate self-adaptive order, and the self-adaptive order smoothes the indoor spectrogram, so that the indoor air spectrogram reflecting the characteristic of the indoor air components is more accurate, and the accuracy of judging the indoor air quality is improved.
Description
Technical Field
The invention relates to the technical field of air quality detection, in particular to an indoor air quality rapid detection method based on artificial intelligence.
Background
In indoor air quality detection, key parameters for judging the air quality include, but are not limited to, suspended particulate matters and carbon dioxide levels, and the freshness, formaldehyde content and the like of the indoor air are determined according to the parameters so as to judge the air quality. In general, the existing infrared spectrum technology is used for detecting and analyzing the air components and the corresponding concentrations, the components in the air can selectively absorb infrared rays with different wavelengths, transition of vibration energy level and rotation energy level in molecules is caused, and a spectral image of the air components can be obtained by detecting the condition that the infrared rays are absorbed. Because the intensities of characteristic peaks and corresponding wavelengths of various components with different concentrations are different, the concentrations of various components in the air can be analyzed through the characteristic of the absorption light of known components, so that the indoor air quality is judged, but because noise possibly causes random fluctuation of a base line in an infrared spectrogram, the base line is not smooth, the intensities of the characteristic peaks are further influenced, and therefore, the analysis result of the concentrations of various components in the air is particularly important for smoothing the spectrogram.
In the related technology, indoor air spectrum images are collected through an infrared spectrum technology, collected indoor spectrum data are subjected to smoothing processing by utilizing a Savitzky-Golay (SG) filter so as to weaken the influence of noise on characteristic peaks in a spectrogram, and the actual content of each component is judged according to the intensity of the characteristic peaks of different components in the spectrogram so as to judge the indoor air quality.
However, the size of the polynomial of the SG filter needs to be set manually at present, the excessive high order can lead to excessive smoothing of data in the spectrogram, partial data details are lost, the excessive low order can lead to excessive fitting of data in the spectrogram to distort, the degree of removing noise interference is insufficient, and the inaccuracy of the polynomial can influence the accuracy of the air quality judgment result.
Disclosure of Invention
In order to solve the problem that the order of the polynomial is inaccurate to influence the accuracy of an air quality judgment result, the invention aims to provide an artificial intelligence-based indoor air quality rapid detection method, and the adopted technical scheme is as follows:
the invention provides an artificial intelligence-based indoor air quality rapid detection method, which comprises the following steps:
Acquiring an indoor air spectrogram, wherein the indoor air spectrogram is a spectrogram in which infrared rays acquired by taking a frequency threshold value as acquisition frequency are absorbed by indoor air, the indoor air spectrogram comprises curves of intensity and wavelength of infrared light characteristic peaks of all indoor air components, and the abscissa of intensity data on the curves is the wavelength and the ordinate is the absorption intensity;
Acquiring a data base line of intensity data in the indoor air spectrogram, a peak value in a curve of intensity and wavelength and a wavelength corresponding to the peak value;
determining local noise intensity performance weights and noise intensities based on the data baselines;
Determining a characteristic representation intensity based on the local noise intensity representation weight, a peak value in the intensity-wavelength curve, and a wavelength corresponding to the peak value;
Square ratio operation is carried out on the noise intensity and the characteristic expression intensity, and an adaptive order is determined;
Smoothing the indoor air spectrogram through an SG filter with the order set as the self-adaptive order to obtain a smoothed indoor air spectrogram;
chemometric analysis is carried out on the smoothed indoor air spectrogram to obtain the components and the corresponding content of the indoor air;
And determining the indoor air quality according to the indoor air component standard content, the indoor air component and the corresponding content.
In some embodiments, the determining local noise intensity performance weights and noise intensities based on the data baselines comprises:
Determining an instantaneous baseline slope corresponding to each intensity data based on the data baseline;
Carrying out data set division on the instantaneous baseline slope corresponding to the intensity data to obtain a plurality of ordered data sets with overlapped parts;
Carrying out segmentation processing on each data set to obtain a head data set, a middle data set and a tail data set;
determining a degree of data fluctuation based on instantaneous baseline slopes in the leading dataset and the trailing dataset;
determining a local mutation level based on the middle dataset;
Determining a local noise intensity performance weight based on the degree of data fluctuation and the degree of local mutation;
The noise intensity is determined based on the local noise intensity performance weights for all data sets.
In some embodiments, the data set dividing the instantaneous baseline slope corresponding to the intensity data to obtain a plurality of ordered data sets with overlapping portions includes:
And constructing a window with a first length by taking an instantaneous baseline slope corresponding to any intensity data as a center, translating the window to the leftmost side of the curve of intensity and wavelength in the direction of decreasing the abscissa value, taking the instantaneous baseline slope corresponding to the intensity data in the window as a data set, taking the instantaneous baseline slope corresponding to the intensity data at the center position in the window as central instantaneous baseline slope data, moving the window along the direction of increasing the abscissa value until the window is moved to the leftmost side of the curve of intensity and wavelength, and sequentially obtaining a plurality of ordered data sets with overlapped parts and central instantaneous baseline slopes at the center position in each data set, wherein the number of the instantaneous baseline slopes in each data set is equal to the length of the window.
In some embodiments, the window includes a first local window, a second local window, and a third local window, where the lengths of the first local window and the third local window are the same and are both the second length, the length of the second local window is the third length, the third length is greater than the second length, the value of the second length is greater than or equal to a length threshold, the value of the first length is equal to a value obtained by adding two times the value of the second length to the value of the third length, and the segmentation processing is performed on each data set to obtain a header data set, a middle data set, and a tail data set, including:
And carrying out segmentation processing on each data set through the first local window, the second local window and the third local window to obtain a head data set corresponding to the first local window, a middle data set corresponding to the second local window and a tail data set corresponding to the third local window.
In some embodiments, determining the degree of data fluctuation based on the instantaneous baseline slope in the leading dataset and the trailing dataset, specifically includes:
calculating the mean and variance of the instantaneous baseline slopes in the leading dataset and the trailing dataset;
based on the average value and variance of the instantaneous baseline slopes in the head data set and the tail data set, obtaining a weighted slope average value of the head data set and a weighted slope average value of the tail data set;
Calculating the difference value of the weighted slope average value of the head data set and the weighted slope average value of the tail data set to obtain the data fluctuation degree;
Determining the degree of fluctuation of the data according to the following formula:
In the method, in the process of the invention, Indicating the degree of fluctuation of the data, a indicating the head data set, b indicating the tail data set,、Respectively represent the average value of the instantaneous baseline slope in the head data set and the average value of the instantaneous baseline slope in the tail data set,、Representing the variance of the instantaneous baseline slope in the leading dataset and the variance of the instantaneous baseline slope in the trailing dataset,、Respectively represent the weighted slope average of the header data set and the weighted slope average of the tail data set,The difference value of the weighted slope average value of the head data set and the weighted slope average value of the tail data set is used for representing the fluctuation degree of data, i represents the serial number of the central instantaneous baseline slope, the value is started from 1,Representing taking absolute value symbols.
In some embodiments, the central instantaneous baseline slope divides a middle data set in a data set corresponding to the central instantaneous baseline slope into a middle sub-data set one and a middle sub-data set two, and neither the middle sub-data set one nor the middle sub-data set two comprises the central instantaneous baseline slope, and the determining the local mutation degree based on the middle data set specifically comprises:
calculating the average value of the instantaneous baseline slope in the first middle sub-data set, the average value of the instantaneous baseline slope in the second middle sub-data set and the standard deviation of the instantaneous baseline slope in the middle data set;
Determining a local mutation level based on the mean of the instantaneous baseline slopes in the first middle sub-dataset, the mean of the instantaneous baseline slopes in the second middle sub-dataset, and the standard deviation of the instantaneous baseline slopes in the middle dataset according to the following formula:
In the method, in the process of the invention, Represents the local mutation degree, c represents the middle sub-data set I, d represents the middle sub-data set II, e represents the middle data set,Representing the mean of the instantaneous baseline slope in the middle sub-dataset,Representing the mean of the instantaneous baseline slope in the middle sub-dataset two,The standard deviation of the instantaneous baseline slope in the middle data set is represented, i represents the serial number of the instantaneous baseline slope of the center, the value is from 1,Representing taking absolute value symbols.
In some embodiments, the local noise intensity performance weight is determined based on the degree of data fluctuation and the degree of local mutation according to the following formula:
In the method, in the process of the invention, Representing the local noise intensity performance weight,The degree of local mutation is indicated,Indicating the degree of fluctuation of the data, i indicates the sequence number of the central instantaneous baseline slope, the value is from 1, and tanh indicates the hyperbolic tangent function.
In some embodiments, determining a characteristic performance intensity based on a local noise intensity performance weight, a peak in a curve of the intensity versus wavelength, and a wavelength corresponding to the peak, comprises:
Determining a peak value and a wavelength corresponding to the peak value based on the intensity data in the indoor air spectrogram;
Determining the maximum intensity of the peak and the half-peak width of the peak based on the peak value in the intensity and wavelength curve, wherein the half-peak width is the difference value between two intersection points, which are intersected by a straight line parallel to an abscissa axis and the intensity and wavelength curve near the maximum intensity of the peak, of the half-value of the maximum intensity of the peak, and the intersection points, which are parallel to an ordinate axis, of the abscissa axis, the two intersection points, and the intensity and wavelength curve, and an area is formed by the intersection points, wherein the intensity and wavelength curve in the area contains a plurality of intensity values;
determining characteristic values and characteristic peaks at wavelengths corresponding to the peak values in each region based on half-peak widths, a plurality of intensity values and local noise intensity expression weights of the peak values in each region, and further obtaining all characteristic values and characteristic peaks of each indoor air spectrogram;
and determining the characteristic expression intensity based on all characteristic values and characteristic peaks of each indoor air spectrogram.
In some embodiments, the characteristic value at the wavelength corresponding to the peak value in each region is determined based on the half-peak width of the peak value in each region, a plurality of intensity values and the local noise intensity expression weight according to the following formula, so as to obtain all the characteristic values of each indoor air spectrogram;
Where j represents the label of the peak in the intensity versus wavelength curve, Is the characteristic value at the wavelength corresponding to the jth peak,A half-width corresponding to the jth peak, n representing the number of intensity values in the region, r representing the index of the intensity values in the region,A value representing the value of the r-th intensity in said region,Representing the local noise intensity performance weight corresponding to the r-th intensity value in the region,A weighted value representing the r-th intensity value within the region,Representing the average intensity of r intensity values within the region.
In some embodiments, the characteristic performance intensity is determined based on all characteristic values and characteristic peaks of each of the indoor air spectrograms according to the following formula:
In the method, in the process of the invention, Representing the characteristic expression intensity, f representing the sequence number of the wavelength corresponding to the position of the characteristic peak in the indoor air spectrogram, wherein the sequence numbers are the same when the wavelength corresponding to the position of the characteristic peak in the indoor air spectrogram is the same, N representing the total number of times of the characteristic peak in all the indoor air spectrograms,Representing the number of times a characteristic peak at the same wavelength appears in all indoor air spectrograms, exp () represents an exponential normalization function, x represents the number of the indoor air spectrograms, M represents the number of sheets of the indoor air spectrograms,Representing the characteristic value at the same wavelength in the x-th indoor air spectrogram,Representing the average of the eigenvalues at the same wavelength in all indoor air spectrograms.
The invention has the following beneficial effects:
According to the artificial intelligence-based indoor air quality rapid detection method provided by the invention, based on the baseline of intensity data in an indoor air spectrogram, the local noise intensity expression weight and the noise intensity are determined, and the local noise intensity expression weight and the noise intensity can reflect the interference degree of noise in the indoor air spectrogram, so that the influence condition of the real intensity data in the indoor air spectrogram by the noise is integrally mastered; based on the local noise intensity expression weight and the intensity and the wavelength corresponding to the peak value in the curve of the intensity and the wavelength, determining the characteristic expression intensity, wherein the obtained characteristic expression intensity considers local noise interference, so that the expression condition of the characteristic peak of each indoor air component in an indoor air spectrogram can be accurately determined; the square ratio operation is carried out on the noise intensity and the characteristic expression intensity, the determined self-adaptive order gives consideration to the noise interference and the characteristic intensity, the problem of smoothing data details under the condition that the noise interference is large and the characteristic intensity is large and the set order is large is avoided, and the problem of insufficient removal degree and data distortion of the noise interference under the condition that the noise interference is large and the characteristic intensity is small and the set order is small can also be avoided; the SG filter with the self-adaptive order is used for carrying out smoothing treatment on the data in the indoor air spectrogram to obtain the indoor air spectrogram which accurately reflects the characteristics of the indoor air components, thereby improving the accuracy of indoor air quality judgment. According to the invention, the order of the SG filter is adaptively adjusted through the noise intensity and the characteristic expression intensity, so that an accurate adaptive order is obtained, and the adaptive order enables an indoor air spectrogram reflecting the characteristic of the indoor air component to be more accurate, thereby improving the accuracy of judging the indoor air quality.
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 flowchart of a method for quickly detecting indoor air quality based on artificial intelligence according to an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
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 following specifically describes a specific scheme of the indoor air quality rapid detection method based on artificial intelligence provided by the invention with reference to the accompanying drawings.
Fig. 1 is a flowchart of a method for quickly detecting indoor air quality based on artificial intelligence according to an embodiment of the invention. Referring to fig. 1, the method comprises the steps of:
Step 101, obtaining an indoor air spectrogram, wherein the indoor air spectrogram is a spectrogram in which infrared rays acquired by taking a threshold frequency as an acquisition frequency are absorbed by indoor air, the indoor air spectrogram comprises curves of intensity and wavelength of infrared light characteristic peaks of indoor air components, the abscissa of intensity data on the curves is the wavelength, and the ordinate is the absorption intensity.
The indoor air spectrogram is the basis for detecting the indoor air quality. In some embodiments, the indoor air spectrum is acquired by installing an infrared spectrum acquisition device in the enclosed space.
Due to being affected by ambient light, in some embodiments, the threshold frequency may be 1 acquisition every 1 h. And under the condition of ensuring that the indoor air circulation is negligible within one day, collecting the indoor air spectrogram 1 time every 1 hour, wherein the number M of the obtained indoor air spectrogram is 24.
In some embodiments, the collected indoor air spectrogram is preprocessed to eliminate extraneous information in the image, recover useful real information, enhance the detectability of useful information and maximally simplify data.
In some implementations, preprocessing may include graying, geometric transformation, and image enhancement.
Step 102, acquiring a data base line of intensity data in an indoor air spectrogram, a peak value in a curve of intensity and wavelength and a wavelength corresponding to the peak value.
When the spectrograms of various components in indoor air are collected, the various components usually show characteristic peaks in the spectrograms, wherein the peak value of the characteristic peaks is the peak value in the curve of intensity and wavelength, the peak value corresponds to the wavelength in the curve, namely, the peak value of the characteristic peak of a certain indoor component under the wavelength, the width and the height of the characteristic peak are related to the component quality and the actual concentration of the characteristic peak, and the data base line is a base line of the curve for measuring the intensity and the wavelength of the infrared light characteristic peak of each indoor air component, but noise can cause random fluctuation in the middle of the indoor air spectrogram.
In some embodiments, a data baseline for intensity data in an indoor air spectrum graph may be obtained by a wavelet transform algorithm.
Step 103, determining local noise intensity expression weights and noise intensities based on the data baselines.
In some embodiments, the specific implementation of step 103 includes:
(1) Based on the data baselines, an instantaneous baseline slope is determined for each intensity data.
Each intensity data has an instantaneous baseline slope corresponding to it, which can characterize the smoothness of the data baseline, the less smooth the baseline, the more abrupt the instantaneous baseline slope, which is manifested as a sudden change in the baseline, i.e., the baseline is mutated at a certain data point, and the instantaneous baseline slopes of other data points on either side of the mutation are similar.
(2) And carrying out data set division on the instantaneous baseline slope corresponding to the intensity data to obtain a plurality of ordered data sets with overlapped parts.
And carrying out data set division on the instantaneous baseline slope corresponding to the intensity data to obtain a plurality of ordered data sets with overlapped parts, so that the local data fluctuation degree in the indoor spectrogram can be conveniently determined later.
In some embodiments, the data set is divided into a plurality of ordered data sets with overlapping portions by specifically: and constructing a window with a first length by taking an instantaneous baseline slope corresponding to any intensity data as a center, translating the window to the leftmost side of a curve with intensity and wavelength in the direction of decreasing the abscissa value, taking the instantaneous baseline slope corresponding to the intensity data in the window as a data set, taking the instantaneous baseline slope corresponding to the intensity data at the central position in the window as central instantaneous baseline slope data, moving the window along the direction of increasing the abscissa value until the window moves to the leftmost side of the curve with intensity and wavelength, and sequentially obtaining a plurality of ordered data sets with overlapping parts and central instantaneous baseline slopes at the central position in each data set, wherein the number of the instantaneous baseline slopes in each data set is equal to the length of the window.
Through the window with the first length, the data sets are divided for the instantaneous baseline slope corresponding to the intensity data, so that a plurality of ordered data sets with overlapping parts and a central instantaneous baseline slope of each data set at the central position are obtained, the ordered data sets can bear continuous local change information in an indoor spectrogram, and the fluctuation degree of data near the intensity data corresponding to the central instantaneous baseline slope can be mastered comprehensively and orderly.
(3) And carrying out segmentation processing on each data set to obtain a head data set, a middle data set and a tail data set.
In some embodiments, the window includes a first local window, a second local window, and a third local window, the lengths of the first local window and the third local window are the same, and are both the second length, the length of the second local window is the third length, the third length is greater than the second length, the value of the second length is greater than or equal to a length threshold, and the value of the first length is equal to the value of the second length which is twice the value of the first length and the value of the third length added, and the method includes the following steps of:
And carrying out segmentation processing on each data set through the first local window, the second local window and the third local window to obtain a head data set corresponding to the first local window, a middle data set corresponding to the second local window and a tail data set corresponding to the third local window.
The third length is larger than the second length, so that the number of the instantaneous baseline slopes in the middle data set is larger than the number of the instantaneous baseline slopes in the head data set and the number of the instantaneous baseline slopes in the tail data set, the instantaneous baseline slope data in the middle data set is richer, and the subsequent calculation of the fluctuation degree of the data is facilitated.
To obtain the local data fluctuation degree in the indoor spectrogram, each data set needs to be segmented by using three local windows.
In some embodiments, the length threshold may be 3, so that the length threshold may be set to ensure that the number of data divided in the header data set and the tail data set of each data set is suitable for determining the local data fluctuation degree.
In some embodiments, the first length may be 11, the second length may be 3, and the third length may be 5. So configured, the total length of the window is 11, the number of instantaneous baseline slope data in each dataset is 11, the length of the first partial window may be 3, the length of the second partial window may be 5, and the length of the third partial window may be 3, dividing each dataset into a header dataset comprising 3 instantaneous baseline slope data, a middle dataset comprising 5 instantaneous baseline slope data, and a tail dataset comprising 3 instantaneous baseline slope data.
(4) The degree of data fluctuation is determined based on the instantaneous baseline slope in the leading data set and the trailing data set.
In some embodiments, determining the degree of data fluctuation based on the instantaneous baseline slope in the leading dataset and the trailing dataset, specifically includes: calculating the mean and variance of the instantaneous baseline slopes in the header data set and the trailer data set; based on the average value and variance of the instantaneous baseline slopes in the head data set and the tail data set, obtaining a weighted slope average value of the head data set and a weighted slope average value of the tail data set; calculating the difference between the weighted slope average value of the head data set and the weighted slope average value of the tail data set to obtain the data fluctuation degree, wherein the difference between the weighted slope average value of the head data set and the weighted slope average value of the tail data set can reflect the instantaneous baseline slope difference near the intensity data corresponding to the central instantaneous baseline slope, and the possibility of mutation at the intensity data corresponding to the central instantaneous baseline slope is evaluated, so that the data fluctuation degree is obtained, and the data fluctuation degree is represented by utilizing the difference between the weighted slope average value of the head data set and the weighted slope average value of the tail data set; determining the degree of fluctuation of the data according to the following formula:
In the method, in the process of the invention, Indicating the degree of fluctuation of the data, a indicating the head data set, b indicating the tail data set,、Respectively represent the average value of the instantaneous baseline slope in the head data set and the average value of the instantaneous baseline slope in the tail data set,、Representing the variance of the instantaneous baseline slope in the leading dataset and the variance of the instantaneous baseline slope in the trailing dataset,、Respectively represent the weighted slope average of the header data set and the weighted slope average of the tail data set,The difference value of the weighted slope average value of the head data set and the weighted slope average value of the tail data set is used for representing the fluctuation degree of data, i represents the serial number of the central instantaneous baseline slope, the value is started from 1,Representing taking absolute value symbols.
The degree of data fluctuation reflects the local continuity degree of the intensity data corresponding to the data set, namely the possibility of local mutation near the intensity data corresponding to the central instantaneous baseline slope, the smaller the degree of data fluctuation is, the more similar the instantaneous baseline slope near the intensity data corresponding to the central instantaneous baseline slope is, the smaller the possibility of mutation is generated at the intensity data corresponding to the central instantaneous baseline slope, which can be judged as the change of the actual intensity data, the larger the degree of data fluctuation is, the larger the instantaneous baseline slope difference near the intensity data corresponding to the central instantaneous baseline slope is, the larger the possibility of mutation is generated at the intensity data corresponding to the central instantaneous baseline slope is, and the change of the intensity data corresponding to the data set can be judged as the data fluctuation caused by noise.
(5) Based on the central dataset, the local degree of mutation is determined.
In some embodiments, the central instantaneous baseline slope divides a middle data set in the data set corresponding to the central instantaneous baseline slope into a middle sub-data set I and a middle sub-data set II, and neither the middle sub-data set I nor the middle sub-data set II comprises the central instantaneous baseline slope, and the local mutation degree is determined based on the middle data set, which specifically comprises: calculating the average value of the instantaneous baseline slope in the first middle sub-data set, the average value of the instantaneous baseline slope in the second middle sub-data set and the standard deviation of the instantaneous baseline slope in the middle data set; determining the local mutation level based on the mean of the instantaneous baseline slopes in the first middle sub-dataset, the mean of the instantaneous baseline slopes in the second middle sub-dataset, and the standard deviation of the instantaneous baseline slopes in the middle dataset according to the following formula:
In the method, in the process of the invention, Represents the local mutation degree, c represents the middle sub-data set I, d represents the middle sub-data set II, e represents the middle data set,Representing the mean of the instantaneous baseline slope in the middle sub-dataset,Representing the mean of the instantaneous baseline slope in the middle sub-dataset two,The standard deviation of the instantaneous baseline slope in the middle data set is represented, i represents the serial number of the instantaneous baseline slope of the center, the value is from 1,Representing taking absolute value symbols.
Because the abrupt change on the base line is expressed as abrupt increase and then abrupt decrease, the first middle sub-data set and the second middle sub-data set are the data positioned on both sides of the central instantaneous base line slope, and the characteristic expression near the intensity data corresponding to the central instantaneous base line slope is more obvious when the characteristic expression is closer to the central instantaneous base line slope, the difference value between the average value of the instantaneous base line slope in the first middle sub-data set on both sides of the central instantaneous base line slope and the average value of the instantaneous base line slope in the second middle sub-data set is calculated firstlyThe difference value can reflect the high degree of local mutation of the central instantaneous baseline slope, the standard deviation of the instantaneous baseline slope in the middle data set reflects the fluctuation of the instantaneous baseline slope in the middle data set e, the larger the standard deviation is, the higher the possibility of noise occurrence of the intensity data corresponding to the instantaneous baseline slope in the middle data set is,And (3) withMultiplying to obtain local mutation degreeThereby more accurately reflecting the local degree of abrupt change near the intensity data corresponding to the slope of the instantaneous baseline near the center.
(6) A local noise intensity performance weight is determined based on the degree of data fluctuation and the degree of local mutation.
In some embodiments, the local noise intensity performance weight is determined based on the degree of data fluctuation and the degree of local mutation according to the following formula:
In the method, in the process of the invention, Representing the local noise intensity performance weight,The degree of local mutation is indicated,Indicating the degree of fluctuation of the data, i indicates the sequence number of the central instantaneous baseline slope, the value is from 1, and tanh indicates the hyperbolic tangent function.
And normalizing the ratio of the fluctuation degree and the local mutation degree of the data by using a hyperbolic tangent function to obtain local noise intensity expression weight, wherein the local noise intensity expression weight reflects the interference degree of local noise in the indoor air spectrogram, and the larger the local noise intensity expression weight is, the larger the interference degree of the local noise in the indoor air spectrogram is reflected.
(7) The noise intensity is determined based on the local noise intensity performance weights for all data sets.
In some embodiments, the local noise intensity performance weights of all the data sets are calculated to obtain the noise intensity, thereby grasping the noise interference degree of the whole indoor air spectrogram.
Step 104, determining the characteristic expression intensity based on the local noise intensity expression weight, the peak value in the curve of the intensity and the wavelength corresponding to the peak value.
It should be noted that, for the whole indoor air spectrogram, in the process of denoising the whole indoor air spectrogram by using the SG filter, the actual intensity data of each component can generate the same local characteristic peak as noise, if only the whole noise intensity is considered to set the polynomial order of the SG filter, the polynomial order is too large to cause data in the spectrogram to be excessively fitted, the data detail is smoothed and data distortion is caused, so that the noise intensity in the indoor air spectrogram needs to be considered and estimated, and meanwhile, the performance condition of the characteristic peak of each component in the whole indoor air spectrogram needs to be judged, so that the influence of a single factor on the order self-adaption is avoided. The actual intensity data of various air components are expressed as peaks in the indoor air spectrogram, and meanwhile, the air flow capacity of the indoor space is negligible, so that the variation of the intensity in the spectrogram and the peak value in the wavelength curve is smaller for the same component, and the variation of the intensity value is larger for noise; meanwhile, the indoor air spectrogram can be influenced by natural light to a certain extent, and noise generated by the indoor air spectrogram can also generate changes with different degrees in the process of detecting multiple times, so that the actual numerical fluctuation condition of all the intensity data can be combined, namely, the numerical influence of noise interference on the intensity data is taken into consideration, and the characteristic self intensity of various air components in the indoor air spectrogram is judged.
In some embodiments, determining the characteristic performance intensity based on the local noise intensity performance weight, the peak in the intensity versus wavelength curve, and the wavelength corresponding to the peak, includes:
(1) And determining the peak value and the wavelength corresponding to the peak value based on the intensity data in the indoor air spectrogram.
In some embodiments, the peak value of the intensity data in the indoor air spectrogram is obtained by a peak detection algorithm, and the position where the peak value occurs, i.e., the wavelength corresponding to the peak value, is recorded so as to analyze what air component the peak value is represented by.
(2) And determining the maximum intensity of the peak and the half-peak width of the peak based on the peak value in the curve of the intensity and the wavelength, wherein the half-peak width is the difference between two intersection points of the intensity and the wavelength curve, which are at the position of a straight line parallel to the axis of abscissa and the vicinity of the maximum intensity of the peak, the axis of abscissa, the line parallel to the axis of ordinate, the intensity and the wavelength curve, and the intensity and the wavelength curve in the region contain a plurality of intensity values.
The half-peak width of the peak is determined so as to determine the characteristic influence range represented by the characteristic value of the indoor air spectrogram.
(3) And determining the characteristic value and the characteristic peak at the wavelength corresponding to the peak value in each region based on the half-peak width of the peak value in each region, a plurality of intensity values and the local noise intensity expression weight, so as to obtain all the characteristic values and the characteristic peaks of each indoor air spectrogram.
The half peak width of the peak value in each area, a plurality of intensity values and local noise intensity expression weights are combined, and the obtained characteristic values give consideration to noise interference and the intensity of the characteristic, so that the obtained characteristic values reflect characteristic peaks of various air components in an indoor air spectrogram more accurately.
(4) The characteristic expression intensity is determined based on all the characteristic values and characteristic peaks of each indoor air spectrogram.
In some embodiments, the characteristic value at the wavelength corresponding to the peak value in each region is determined based on the half-peak width of the peak value in each region, a plurality of intensity values and the local noise intensity expression weight according to the following formula, so as to obtain all the characteristic values of each indoor air spectrogram;
Where j represents the label of the peak in the intensity versus wavelength curve, Is the characteristic value at the wavelength corresponding to the jth peak,A half-width corresponding to the jth peak, n representing the number of intensity values in the region, r representing the index of the intensity values in the region,A value representing the value of the r-th intensity in said region,Representing the local noise intensity performance weight corresponding to the r-th intensity value in the region,A weighted value representing the r-th intensity value within the region,Representing the average intensity of r intensity values within the region.
Half-width corresponding to jth peakThe larger the spectrum is, the half-width corresponding to the jth peak in the indoor spectrum is illustratedThe greater the wavelength range encompassed by the intensity versus wavelength curve within; the larger the weighted value of the r-th intensity value in the region is, the higher the expression degree of the intensity data of the air component corresponding to the r-th intensity value in the region is; the larger the average intensity of r intensity values in the region, the higher the expression degree of the intensity data of the air component in the region; characteristic value at wavelength corresponding to jth peakThe larger the characteristic of the air component at the wavelength corresponding to the peak, the more pronounced the performance.
To determine the characteristic performance intensity of the indoor air component at the same wavelength in all of the collected indoor spectrograms, in some embodiments, the characteristic performance intensity is determined based on all of the characteristic values and characteristic peaks of each of the indoor air spectrograms according to the following formula:
In the method, in the process of the invention, Representing the characteristic expression intensity, f representing the sequence number of the wavelength corresponding to the position of the characteristic peak in the indoor air spectrogram, wherein the sequence numbers are the same when the wavelength corresponding to the position of the characteristic peak in the indoor air spectrogram is the same, N representing the total number of times of the characteristic peak in all the indoor air spectrograms,Representing the number of times a characteristic peak at the same wavelength appears in all indoor air spectrograms, exp () represents an exponential normalization function, x represents the number of the indoor air spectrograms, M represents the number of sheets of the indoor air spectrograms,Representing the characteristic value at the same wavelength in the x-th indoor air spectrogram,Representing the average of the eigenvalues at the same wavelength in all indoor air spectrograms.
The smaller the standard deviation is, the less the characteristic peak of the air component at the wavelength is affected by the change of time and environmental factors, the more obvious the characteristic performance is, the higher the characteristic performance intensity is, and then the negative correlation mapping and normalization processing is carried out by using an exponential function. The number of times the characteristic peak at the same wavelength appears in all indoor air spectrogramsThe larger the spectrum of the indoor air, the higher the frequency of occurrence of the characteristic peak in all indoor air spectra at a certain wavelength, and the higher the probability that the intensity data of the wavelength is expressed as the indoor air component, and the utilization is further carried outThe standard deviation is weighted and characteristic peaks appearing in all indoor air spectrums are traversed, so that characteristic expression intensity is obtained, local noise interference is considered in the obtained characteristic expression intensity, and therefore the expression situation of the characteristic peaks of each indoor air component in an indoor air spectrum graph can be accurately determined.
And 105, performing square ratio operation on the noise intensity and the characteristic expression intensity, and determining the self-adaptive order.
The self-adaptive order determined by the method gives consideration to the noise interference and the intensity of the feature, so that the problem of smoothing data details under the condition that the noise interference is large and the intensity of the feature is large and the set order is large is avoided, and the problem of insufficient removal degree of the noise interference and data distortion under the condition that the noise interference is large and the intensity of the feature is small and the set order is small is also avoided.
In some embodiments, the adaptive order is determined by square ratio operation based on noise intensity and feature expression intensity according to the following formula:
In the method, in the process of the invention, The number of the adaptive orders is represented,The local noise intensity representing all data sets represents the average of the weights, i.e. the noise intensity,Representing the characteristic expression intensity, C represents the value of the conventionally set order.
It should be noted that, when the noise interference is large and the intensity of the feature itself is large, the smoothing is not required to be performed to an excessive extent, that is, the polynomial order is not required to be excessive, and when the noise interference is large and the intensity of the feature itself is small, the noise interference needs to be filtered to a greater extent, that is, the polynomial order needs to be adjusted to be high.
The numerical value of the conventional set order is weighted and regulated through the square ratio of the noise intensity and the characteristic expression intensity, and the noise interference and the characteristic intensity are considered, so that the determined self-adaptive order has self-adaptability.
In some embodiments, the conventionally set degree C may have a value of 3, and the third degree is a polynomial degree that is relatively conventional and facilitates adaptive adjustment.
In some embodiments, since the square ratio of the noise intensity and the characteristic performance intensity may be a fraction, when the square ratio is a fraction, it is rounded, and rounded to obtain the final adaptive order.
And 106, smoothing the indoor air spectrogram through an SG filter with the order set as the self-adaptive order to obtain a smoothed indoor air spectrogram.
And smoothing the data in the indoor air spectrogram through the SG filter with the self-adaptive order to obtain the indoor air spectrogram accurately reflecting the characteristic of the indoor air components.
And 107, performing chemometric analysis on the smoothed indoor air spectrogram to obtain the components and the corresponding content of the indoor air.
The chemometric analysis can determine various air components and the corresponding contents thereof in the indoor spectrogram so as to provide a numerical basis for the evaluation of air quality, and the smoothed indoor air spectrogram can enable the contents of the various air components to be more accurate.
In some embodiments, a calibration curve or model is created using known levels of standard air components, which can be correlated to the intensity versus wavelength curve in the indoor spectrogram with known levels to determine the composition and corresponding levels of indoor air.
In some embodiments, the chemometric analysis may be principal component analysis, partial least squares regression, or the like.
And 108, determining the indoor air quality according to the indoor air component standard content, the indoor air component and the corresponding content.
The smoothed indoor air spectrogram can enable the obtained content of various air components to be more accurate, and the indoor air component standard content is compared with the indoor air component content obtained through analysis to obtain a judgment result of the indoor air quality, so that the accuracy of the determined indoor air quality is improved.
In summary, according to the artificial intelligence-based indoor air quality rapid detection method provided by the invention, the local noise intensity expression weight and the noise intensity are determined based on the base line of the intensity data in the indoor air spectrogram, the interference degree of the noise in the indoor air spectrogram is mastered, and the influence condition of the real intensity data in the indoor air spectrogram by the noise is mastered as a whole; based on the local noise intensity expression weight and the intensity and the wavelength corresponding to the peak value in the curve of the intensity and the wavelength, the characteristic expression intensity is determined, so that the obtained characteristic expression intensity is regulated by the local noise interference, and the characteristic peak expression condition of each determined indoor air component in the indoor air spectrogram can be more accurate; the square ratio operation is carried out on the noise intensity and the characteristic expression intensity, the determined self-adaptive order gives consideration to the noise interference and the characteristic intensity, the problem that data details are smoothed out under the condition that the noise interference is large and the characteristic intensity is large and the set order is large is avoided, and the problem that the noise interference removing degree is insufficient and the data distortion is avoided under the condition that the noise interference is large and the characteristic intensity is small and the set order is small is also avoided. According to the invention, the order of the SG filter is adaptively adjusted through the noise intensity and the characteristic expression intensity, so that an accurate adaptive order is obtained, and the adaptive order smoothes the indoor spectrogram, so that the indoor air spectrogram reflecting the characteristic of the indoor air components is more accurate, and the accuracy of judging the indoor air quality 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 (10)
1. The indoor air quality rapid detection method based on artificial intelligence is characterized by comprising the following steps of:
Acquiring an indoor air spectrogram, wherein the indoor air spectrogram is a spectrogram in which infrared rays acquired by taking a frequency threshold value as acquisition frequency are absorbed by indoor air, the indoor air spectrogram comprises curves of intensity and wavelength of infrared light characteristic peaks of all indoor air components, and the abscissa of intensity data on the curves is the wavelength and the ordinate is the absorption intensity;
Acquiring a data base line of intensity data in the indoor air spectrogram, a peak value in a curve of intensity and wavelength and a wavelength corresponding to the peak value;
determining local noise intensity performance weights and noise intensities based on the data baselines;
Determining a characteristic representation intensity based on the local noise intensity representation weight, a peak value in the intensity-wavelength curve, and a wavelength corresponding to the peak value;
Square ratio operation is carried out on the noise intensity and the characteristic expression intensity, and self-adaptive order is determined;
Smoothing the indoor air spectrogram through an SG filter with the order set as the self-adaptive order to obtain a smoothed indoor air spectrogram;
chemometric analysis is carried out on the smoothed indoor air spectrogram to obtain the components and the corresponding content of the indoor air;
And determining the indoor air quality according to the indoor air component standard content, the indoor air component and the corresponding content.
2. The rapid indoor air quality detection method based on artificial intelligence of claim 1, wherein the determining local noise intensity expression weights and noise intensities based on the data base line comprises:
Determining an instantaneous baseline slope corresponding to each intensity data based on the data baseline;
Carrying out data set division on the instantaneous baseline slope corresponding to the intensity data to obtain a plurality of ordered data sets with overlapped parts;
Carrying out segmentation processing on each data set to obtain a head data set, a middle data set and a tail data set;
determining a degree of data fluctuation based on instantaneous baseline slopes in the leading dataset and the trailing dataset;
determining a local mutation level based on the middle dataset;
Determining a local noise intensity performance weight based on the degree of data fluctuation and the degree of local mutation;
The noise intensity is determined based on the local noise intensity performance weights for all data sets.
3. The method for rapid indoor air quality detection based on artificial intelligence of claim 2, wherein the partitioning of the data set of instantaneous baseline slopes corresponding to the intensity data to obtain a plurality of ordered data sets with overlapping portions comprises:
And constructing a window with a first length by taking an instantaneous baseline slope corresponding to any intensity data as a center, translating the window to the leftmost side of the curve of intensity and wavelength in the direction of decreasing the abscissa value, taking the instantaneous baseline slope corresponding to the intensity data in the window as a data set, taking the instantaneous baseline slope corresponding to the intensity data at the center position in the window as central instantaneous baseline slope data, moving the window along the direction of increasing the abscissa value until the window is moved to the leftmost side of the curve of intensity and wavelength, and sequentially obtaining a plurality of ordered data sets with overlapped parts and central instantaneous baseline slopes at the center position in each data set, wherein the number of the instantaneous baseline slopes in each data set is equal to the length of the window.
4. The indoor air quality rapid detection method based on artificial intelligence according to claim 3, wherein the window comprises a first local window, a second local window and a third local window, the lengths of the first local window and the third local window are the same, the lengths of the second local window are the third length, the third length is greater than the second length, the value of the second length is greater than or equal to a length threshold, the value of the first length is equal to the value of the second length which is twice the value of the third length, and the value added by the value of the second length and the value of the third length is the value, and the segmentation processing is performed on each data set to obtain a head data set, a middle data set and a tail data set, which comprises:
And carrying out segmentation processing on each data set through the first local window, the second local window and the third local window to obtain a head data set corresponding to the first local window, a middle data set corresponding to the second local window and a tail data set corresponding to the third local window.
5. The indoor air quality rapid detection method based on artificial intelligence according to claim 3, wherein determining the degree of data fluctuation based on the instantaneous baseline slope in the header data set and the tail data set, specifically comprises:
calculating the mean and variance of the instantaneous baseline slopes in the leading dataset and the trailing dataset;
based on the average value and variance of the instantaneous baseline slopes in the head data set and the tail data set, obtaining a weighted slope average value of the head data set and a weighted slope average value of the tail data set;
Calculating the difference value of the weighted slope average value of the head data set and the weighted slope average value of the tail data set to obtain the data fluctuation degree;
Determining the degree of fluctuation of the data according to the following formula:
In the method, in the process of the invention, Indicating the degree of fluctuation of the data, a indicating the head data set, b indicating the tail data set,、Respectively represent the average value of the instantaneous baseline slope in the head data set and the average value of the instantaneous baseline slope in the tail data set,、Representing the variance of the instantaneous baseline slope in the leading dataset and the variance of the instantaneous baseline slope in the trailing dataset,、Respectively represent the weighted slope average of the header data set and the weighted slope average of the tail data set,The difference value of the weighted slope average value of the head data set and the weighted slope average value of the tail data set is used for representing the fluctuation degree of data, i represents the serial number of the central instantaneous baseline slope, the value is started from 1,Representing taking absolute value symbols.
6. The rapid indoor air quality detection method based on artificial intelligence of claim 5, wherein the central instantaneous baseline slope divides a middle data set in a data set corresponding to the central instantaneous baseline slope into a middle sub-data set one and a middle sub-data set two, and neither the middle sub-data set one nor the middle sub-data set two comprises the central instantaneous baseline slope, and the determining the local mutation degree based on the middle data set specifically comprises:
calculating the average value of the instantaneous baseline slope in the first middle sub-data set, the average value of the instantaneous baseline slope in the second middle sub-data set and the standard deviation of the instantaneous baseline slope in the middle data set;
Determining a local mutation level based on the mean of the instantaneous baseline slopes in the first middle sub-dataset, the mean of the instantaneous baseline slopes in the second middle sub-dataset, and the standard deviation of the instantaneous baseline slopes in the middle dataset according to the following formula:
In the method, in the process of the invention, Represents the local mutation degree, c represents the middle sub-data set I, d represents the middle sub-data set II, e represents the middle data set,Representing the mean of the instantaneous baseline slope in the middle sub-dataset,Representing the mean of the instantaneous baseline slope in the middle sub-dataset two,The standard deviation of the instantaneous baseline slope in the middle data set is represented, i represents the serial number of the instantaneous baseline slope of the center, the value is from 1,Representing taking absolute value symbols.
7. The rapid indoor air quality detection method based on artificial intelligence of claim 6, wherein the local noise intensity expression weight is determined based on the data fluctuation degree and the local mutation degree according to the following formula:
In the method, in the process of the invention, Representing the local noise intensity performance weight,The degree of local mutation is indicated,Indicating the degree of fluctuation of the data, i indicates the sequence number of the central instantaneous baseline slope, the value is from 1, and tanh indicates the hyperbolic tangent function.
8. The method of claim 3, wherein determining a characteristic performance intensity based on the local noise intensity performance weight, a peak in a curve of intensity versus wavelength, and a wavelength corresponding to the peak comprises:
Determining a peak value and a wavelength corresponding to the peak value based on the intensity data in the indoor air spectrogram;
Determining the maximum intensity of the peak and the half-peak width of the peak based on the peak value in the intensity and wavelength curve, wherein the half-peak width is the difference value between two intersection points, which are intersected by a straight line parallel to an abscissa axis and the intensity and wavelength curve near the maximum intensity of the peak, of the half-value of the maximum intensity of the peak, and the intersection points, which are parallel to an ordinate axis, of the abscissa axis, the two intersection points, and the intensity and wavelength curve, and an area is formed by the intersection points, wherein the intensity and wavelength curve in the area contains a plurality of intensity values;
determining characteristic values and characteristic peaks at wavelengths corresponding to the peak values in each region based on half-peak widths, a plurality of intensity values and local noise intensity expression weights of the peak values in each region, and further obtaining all characteristic values and characteristic peaks of each indoor air spectrogram;
and determining the characteristic expression intensity based on all characteristic values and characteristic peaks of each indoor air spectrogram.
9. The rapid indoor air quality detection method based on artificial intelligence according to claim 8, wherein the characteristic value at the wavelength corresponding to the peak value in each region is determined based on the half-peak width of the peak value in each region, a plurality of intensity values and local noise intensity expression weights according to the following formula, so as to obtain all characteristic values of each indoor air spectrogram;
Where j represents the label of the peak in the intensity versus wavelength curve, Is the characteristic value at the wavelength corresponding to the jth peak,A half-width corresponding to the jth peak, n representing the number of intensity values in the region, r representing the index of the intensity values in the region,A value representing the value of the r-th intensity in said region,Representing the local noise intensity performance weight corresponding to the r-th intensity value in the region,A weighted value representing the r-th intensity value within the region,Representing the average intensity of r intensity values within the region.
10. The rapid indoor air quality detection method based on artificial intelligence of claim 9, wherein the characteristic expression intensity is determined based on all characteristic values and characteristic peaks of each indoor air spectrogram according to the following formula:
In the method, in the process of the invention, Representing the characteristic expression intensity, f representing the sequence number of the wavelength corresponding to the position of the characteristic peak in the indoor air spectrogram, wherein the sequence numbers are the same when the wavelength corresponding to the position of the characteristic peak in the indoor air spectrogram is the same, N representing the total number of times of the characteristic peak in all the indoor air spectrograms,Representing the number of times a characteristic peak at the same wavelength appears in all indoor air spectrograms, exp () represents an exponential normalization function, x represents the number of the indoor air spectrograms, M represents the number of sheets of the indoor air spectrograms,Representing the characteristic value at the same wavelength in the x-th indoor air spectrogram,Representing the average of the eigenvalues at the same wavelength in all indoor air spectrograms.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410788850.6A CN118366563B (en) | 2024-06-19 | 2024-06-19 | Indoor air quality rapid detection method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410788850.6A CN118366563B (en) | 2024-06-19 | 2024-06-19 | Indoor air quality rapid detection method based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118366563A true CN118366563A (en) | 2024-07-19 |
| CN118366563B CN118366563B (en) | 2024-09-13 |
Family
ID=91882205
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410788850.6A Active CN118366563B (en) | 2024-06-19 | 2024-06-19 | Indoor air quality rapid detection method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118366563B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011106527A2 (en) * | 2010-02-24 | 2011-09-01 | Ampersand International, Inc. | Method for noise filtering based on confidence interval evaluation |
| CN105005978A (en) * | 2015-07-15 | 2015-10-28 | 天津大学 | Spectrum real-time filtering method based on Savitzky-Golay filter parameter optimization |
| WO2017153726A1 (en) * | 2016-03-07 | 2017-09-14 | Micromass Uk Limited | Spectrometric analysis |
| CN108844939A (en) * | 2018-03-14 | 2018-11-20 | 西安电子科技大学 | Raman spectrum based on asymmetric weighted least-squares detects baseline correction method |
| CN112666097A (en) * | 2020-12-30 | 2021-04-16 | 中船重工安谱（湖北）仪器有限公司 | Fourier infrared polynomial iterative smoothing filtering baseline correction method |
| CN115359847A (en) * | 2022-08-09 | 2022-11-18 | 国科温州研究院(温州生物材料与工程研究所) | Peak searching algorithm for proteomics series mass spectrogram |
| US20220383979A1 (en) * | 2020-02-10 | 2022-12-01 | Zhejiang Digena Diagnostic Technology Co., Ltd. | Nucleic acid mass spectrum numerical processing method |
| US20230018735A1 (en) * | 2019-12-11 | 2023-01-19 | Thermo Fisher Scientific (Bremen) Gmbh | Processing optical spectra |
| CN117030654A (en) * | 2023-10-10 | 2023-11-10 | 武汉怡特环保科技有限公司 | Method for measuring concentration of trace-level sulfur dioxide in air |
| CN117629967A (en) * | 2024-01-26 | 2024-03-01 | 江苏省南京环境监测中心 | Quick soil layer pollution detection method based on spectral analysis |
-
2024
- 2024-06-19 CN CN202410788850.6A patent/CN118366563B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011106527A2 (en) * | 2010-02-24 | 2011-09-01 | Ampersand International, Inc. | Method for noise filtering based on confidence interval evaluation |
| CN105005978A (en) * | 2015-07-15 | 2015-10-28 | 天津大学 | Spectrum real-time filtering method based on Savitzky-Golay filter parameter optimization |
| WO2017153726A1 (en) * | 2016-03-07 | 2017-09-14 | Micromass Uk Limited | Spectrometric analysis |
| CN108844939A (en) * | 2018-03-14 | 2018-11-20 | 西安电子科技大学 | Raman spectrum based on asymmetric weighted least-squares detects baseline correction method |
| US20230018735A1 (en) * | 2019-12-11 | 2023-01-19 | Thermo Fisher Scientific (Bremen) Gmbh | Processing optical spectra |
| US20220383979A1 (en) * | 2020-02-10 | 2022-12-01 | Zhejiang Digena Diagnostic Technology Co., Ltd. | Nucleic acid mass spectrum numerical processing method |
| CN112666097A (en) * | 2020-12-30 | 2021-04-16 | 中船重工安谱（湖北）仪器有限公司 | Fourier infrared polynomial iterative smoothing filtering baseline correction method |
| CN115359847A (en) * | 2022-08-09 | 2022-11-18 | 国科温州研究院(温州生物材料与工程研究所) | Peak searching algorithm for proteomics series mass spectrogram |
| CN117030654A (en) * | 2023-10-10 | 2023-11-10 | 武汉怡特环保科技有限公司 | Method for measuring concentration of trace-level sulfur dioxide in air |
| CN117629967A (en) * | 2024-01-26 | 2024-03-01 | 江苏省南京环境监测中心 | Quick soil layer pollution detection method based on spectral analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118366563B (en) | 2024-09-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116660198B (en) | Carbon nanomaterial slurry performance test method based on spectral analysis | |
| CN111089856B (en) | Post-processing method for extracting Raman spectrum weak signal | |
| CN117589741B (en) | Indocyanine green intelligent detection method based on optical characteristics | |
| CN116818739A (en) | Indocyanine green detection method based on optics | |
| CN116698680B (en) | Automatic monitoring method and system for biological aerosol | |
| CN117368141B (en) | Perchlorate wastewater concentration intelligent detection method based on artificial intelligence | |
| CN109991181B (en) | Adaptive surface absorption spectrum analysis method, system, storage medium and device | |
| CN105005978A (en) | Spectrum real-time filtering method based on Savitzky-Golay filter parameter optimization | |
| CN115221927A (en) | Ultraviolet-visible spectrum dissolved organic carbon detection method | |
| CN118366563B (en) | Indoor air quality rapid detection method based on artificial intelligence | |
| CN114186596B (en) | Multi-window identification method and device for spectrogram peaks and electronic equipment | |
| CN113076692B (en) | Method for inverting nitrogen content of leaf | |
| CN117609872A (en) | Method for improving accuracy of LIBS (laser induced breakdown spectroscopy) detection on content of calcium element in carbonate rock | |
| CN117541626A (en) | Remote sensing image change detection method based on deep learning | |
| CN117556245A (en) | Method for detecting filtered impurities in tetramethylammonium hydroxide production | |
| CN114062306B (en) | Near infrared spectrum data segmentation preprocessing method | |
| CN113361610B (en) | Intelligent identification method and system for wine production place | |
| CN115146219A (en) | Raman spectrum spike removing method and system based on derivative spectrum and cubic smoothing | |
| CN118032695B (en) | Nitrogen-sulfur in-situ measurement method and system based on differential ultraviolet spectrum technology | |
| CN112229817A (en) | Method for establishing soda saline-alkali soil heavy metal quantitative inversion model | |
| CN118549378B (en) | Quick rice starch quality detection method for rice grouting period | |
| CN115389440A (en) | Method and device for predicting content of components in sample | |
| CN112697746A (en) | Spectrum information-based sweet potato powder alum content detection method | |
| CN117929356B (en) | LIBS quantitative analysis method based on Gaussian process regression | |
| CN117825601B (en) | Method for measuring sulfur dioxide in food |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |