CN111366554B - Method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy - Google Patents

Method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy Download PDF

Info

Publication number
CN111366554B
CN111366554B CN202010452130.4A CN202010452130A CN111366554B CN 111366554 B CN111366554 B CN 111366554B CN 202010452130 A CN202010452130 A CN 202010452130A CN 111366554 B CN111366554 B CN 111366554B
Authority
CN
China
Prior art keywords
trace
sample
remaining time
infrared spectrum
total reflection
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.)
Active
Application number
CN202010452130.4A
Other languages
Chinese (zh)
Other versions
CN111366554A (en
Inventor
刘昇
叶珈菁
丁丁
林龚伟
张剑峰
朱宇翀
邢锦宏
王志体
朱金龙
刘畅
许如春
胡友辰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suzhou Public Security Bureau Gusu Branch
Original Assignee
Suzhou Public Security Bureau Gusu Branch
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Suzhou Public Security Bureau Gusu Branch filed Critical Suzhou Public Security Bureau Gusu Branch
Priority to CN202010452130.4A priority Critical patent/CN111366554B/en
Publication of CN111366554A publication Critical patent/CN111366554A/en
Application granted granted Critical
Publication of CN111366554B publication Critical patent/CN111366554B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G01N21/3577—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing liquids, e.g. polluted water
    • 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
    • G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
    • 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
    • G01N21/55—Specular reflectivity
    • G01N21/552—Attenuated total reflection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • 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
    • G01N2021/3595—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using FTIR

Abstract

A method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy comprises the steps of collecting infrared spectrum data of material detection traces, making sample traces, classifying samples, collecting a sample data set, establishing a corresponding relation model of the infrared spectrum data of the sample traces and the remaining time by using a neural network, verifying precision, inputting the infrared spectrum data of the material detection traces into a model meeting precision requirements, simulating the remaining time of the material detection traces, evaluating a simulation result, and expanding the sample data set if necessary; the method has good reproducibility, simple and convenient sample preparation, stable and reliable result and high detection precision, and meets the requirement of performing rapid nondestructive identification in the criminal investigation field.

Description

Method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy
Technical Field
The invention belongs to the technical field of criminal science, and particularly relates to a method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy.
Background
When the case is prospectively checked on site, the correct identification and judgment of the remaining time of the on-site traces (including blood traces, handwriting, sweat fingerprints and the like) is helpful for objectively reflecting the activity condition of on-site personnel at the moment of the case, besides the conventional extraction and detection of the traces such as fingerprints, footprints, DNA and the like and the record and detection of the basic form, color, thickness, distribution form and the like of the blood traces, and also can be a powerful basis for confirming the time of the case.
Limited by field conditions and the level of basic scientific research, when the case field is inspected and examined, the method has no deep research progress on the observation and judgment of the remaining time of the field trace, the research on the change rule of the cadaver tissue is concentrated on the forensic medicine inspection, and the method is rarely used for regularly exploring the in-vitro blood trace of the case field; most research directions of blood are medical, methods for judging the blood trace remaining time on site through change rules are few, the methods for researching the blood trace on site mostly need complicated experimental flows and expensive equipment, and most judgment methods have large errors.
The method for detecting the traces with damage comprises an electrophoresis technology, a high performance liquid chromatography technology, a gas chromatography, an oxygen electrode technology, a gene analysis technology and the like, and the trace amount is small, so that the research method for detecting the trace remaining time without loss is more beneficial to the storage of the trace, and the sample preparation and sampling can be repeatedly carried out. The nondestructive inspection method for trace remaining time includes a stereo microscope, a scanning electron microscope, an atomic force microscope, a near infrared reflection spectroscopy technology, an optical fiber spectroscopy technology, a hyperspectral photography method, an electron paramagnetic resonance technology and the like, however, the existing nondestructive inspection is mostly focused on experimental research.
Dengxiangjun and the like (research on relationship between blood mark infrared thermal imaging change and blood mark time [ J ]. forensic test 2015, (8): 72-74) adopt an infrared thermography technology and a computer image analysis technology to observe and calculate the blood mark temperature on five types of carriers such as glass, textile, A4 paper, ceramic tiles, wood blocks and the like, record and count the temperature condition of the blood marks on the five types of objects within 15 minutes per minute from sample preparation, and provide a better regression equation through regression analysis, but the investigation time is too short, and the space for further research and application is smaller.
Researches show that the attenuated total reflection infrared spectroscopy (ATR) can reveal the change of an infrared spectrogram caused by small change of an experimental sample and can meet the requirement of nondestructive rapid inspection of traces.
Huangping et al (applying FT infrared spectroscopy technology to infer death time [ J ]. J. China journal of forensic medicine 2011, (2): 104-109) apply KBr tablet method and ATR method to carry out Fourier infrared spectroscopy detection on left ventricular muscle, lung, liver, spleen, kidney cortex, skeletal muscle and cerebral cortex of human body and rat respectively, and select main absorption peaks in spectral data as research indexes to indicate the change rule of attenuated total reflection infrared spectroscopy for human cadaver in vitro tissues, but the judgment on death time and organ in vitro time is fuzzy and has certain limitation.
The neural network is a mathematical model for simulating a brain neural synapse connection structure to process information, signals between two nodes are endowed with weighted values, the built neural network is subjected to learning training for inputting a large number of signals, the weighting is adjusted to realize classification of the information, after a new signal is input to the neural network trained by learning, the neural network realizes automatic classification of the information according to the determined weighted values, and the neural network is the basis for parallel processing and large-scale parallel calculation of a large amount of information. The combination of the neural network and the spectroscopy is widely applied in various fields at present, and the fields of automatic license plate identification, liquor year identification, big data analysis and the like are all applied.
Zhang Bao Ju et al (BP neural network-based noninvasive detection [ J ] of red blood cell concentration in human blood spectroscopy and spectral analysis 2012, (9): 2508-2511) studied the BP neural network-based noninvasive detection method of red blood cell concentration in human blood, and used the spectral data of fresh blood as modeling data to predict the red blood cell concentration of fresh blood.
Chinese patent application CN110196233A and Master thesis blood trace detection research based on hyperspectral imaging technology disclose a blood trace time prediction method based on hyperspectral imaging, the method takes a hyperspectral bloodstain three-dimensional image cube as data support, extracts a spectral curve of a bloodstain sample, performs preprocessing matrixing, and constructs a time-spectral domain model related to bloodstain time prediction by using a nonlinear iteration and nonlinear fitting method, however, the model building process is complex, the sample is blood stain (dry sample) collected in a laboratory and stored at the constant temperature of 25 ℃, and only sample data of 1h, 24h and 30 days are collected to build the model, the reliability of the method is questionable, the blood samples (liquid samples, low-temperature storage and anticoagulant-containing samples) collected by a blood bank of a hospital are used for testing and checking, the sampling time is unreasonable, and the difference between the sample and the storage environment of a test material can obviously influence the accuracy of the experimental conclusion.
The traditional investigation and identification process of the criminal science technology is mature in reflecting the spatial relation problem of people and people, people and objects, and objects, but only can roughly realize the inference on the time relation, the analysis and identification of the remaining time of the field trace is always the difficult point of the work of the criminal science technology, and at present, the research of combining the attenuated total reflection infrared spectroscopy and the neural network for judging the remaining time of the trace is not reported.
Disclosure of Invention
The invention aims to provide a method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy (ATR-FTIR), which combines the traditional criminal scientific and technical research, spectroscopy and computer science, trains a neural network by using infrared spectrum data of a trace sample, and further judges the trace remaining time2>0.8, meeting the requirement of carrying out rapid nondestructive identification in the criminal investigation field.
In order to achieve the purpose, the invention provides the following technical scheme:
a method for measuring trace remaining time based on an attenuated total reflection infrared spectrum comprises the following steps:
1) collecting infrared spectrum data of material detection trace
The material detection trace comes from a case scene, the material detection trace comprises but is not limited to blood trace, handwriting, sweat fingerprint and the like, a fixed wave number range and resolution ratio are set, infrared spectrum data of the material detection trace are collected by using attenuated total reflection infrared spectroscopy,
2) simulating material-testing trace to make sample trace and classifying sample
And (3) preparing a sample trace by using the same material as the material detection trace carrier as the carrier, wherein the sample trace is prepared according to the following steps of 7: 3 proportion, randomly dividing the trace into a training sample trace and a verification sample trace,
3) collecting sample data set
Simulating the legacy environment of the material detection trace, using attenuated total reflection infrared spectroscopy, requiring the wave number range and resolution to be the same as the former, acquiring infrared spectrum data of sample traces with different legacy times at uniform time intervals, summarizing the infrared spectrum data and the corresponding legacy times into a sample data set, recording the maximum value of the legacy times of the sample traces, retaining the corresponding relation between the infrared spectrum data and the legacy times in the sample data set, dividing the sample data set into a training set and a verification set according to the training and verification purposes,
4) establishing a corresponding relation model of infrared spectrum data and the remaining time of the sample trace by using a neural network, verifying the precision, training the neural network by taking the infrared spectrum data and the remaining time in a training set as variables, establishing a corresponding relation model of the infrared spectrum data and the remaining time of the sample trace, and verifying the model precision by using a verification set, such as a judgment coefficient R2If the value is more than 0.8, the neural network is available, otherwise, the process from training the neural network to verifying the model precision by using the verification set is repeated until the judgment coefficient R2Is greater than 0.8 of the total weight of the rubber,
5) inputting the infrared spectrum data of the material detection trace into a model meeting the precision requirement, simulating the leaving time of the material detection trace, inputting the infrared spectrum data of the material detection trace into an available model to obtain the simulated leaving time of the material detection trace,
6) evaluation of remaining time of material trace
Comparing the simulated left time of the material detection trace with the maximum value of the left time of the sample trace, if the ratio is less than or equal to 0.9, evaluating the simulated left time of the material detection trace as the measured left time of the material detection trace, and if the ratio is greater than 0.9, expanding the sample data set, prolonging the left time of the sample trace, repeating the steps 3) to 5), comparing the simulated left time of the material detection trace after expanding the sample data set with the maximum value of the left time of the sample trace until the ratio is less than or equal to 0.9, evaluating the simulated left time of the material detection trace as the measured left time of the material detection trace, and finishing the measurement.
Preferably, in step 3), the sampling time interval of the different time trace samples is 24 hours, and the collection wave number range of the infrared spectrum is: 4000cm-1-600cm-1。
The technical scheme of the invention can also be expressed as follows:
a method for measuring trace remaining time based on an attenuated total reflection infrared spectrum comprises the following steps:
a. collecting infrared spectral data sets
Collecting infrared spectrum data of a material detection trace of a case sending site to be detected;
making sample traces simulating the field of case issuance at different leaving times, collecting infrared spectrums of the sample traces, and acquiring infrared absorption spectrum data to obtain an infrared spectrum data set;
the infrared spectrum data set comprises an infrared spectrum data set of the corresponding training set sample trace and an infrared spectrum data set of the corresponding verification set sample trace;
b. constructing neural networks
Constructing a neural network architecture by using the remaining time of each sample trace and infrared spectrum data as variables by using a data processing tool, setting the number of layers, selecting a training type and an error type, setting a transfer function of a hidden layer and an output layer as TANSILG, and setting the number of the hidden layers and the number of neurons;
setting the number of neurons and training parameters according to the selected training type, training the constructed neural network architecture by utilizing the infrared spectrum data set of the training set sample trace, then inputting the infrared spectrum data of the verification set trace sample, outputting a trace remaining time analog value of the verification set sample trace, and comparing the trace remaining time analog value with a true value of the corresponding verification set sample trace remaining time to obtain a judgment coefficient R2If the coefficient R is determined2>0.8, finishing construction, otherwise, training the neural network again until a judgment coefficient R2>0.8, obtaining a neural network;
c. detecting a sample to be tested
And inputting the collected infrared spectrum data of the sample trace of the to-be-tested case launching site into a neural network, and outputting the remaining time of the to-be-tested trace sample by the neural network.
Further, in step a), the sampling time intervals of the different remaining time sample traces are 24 hours.
Preferably, in step a), the infrared spectrum data extracted is data of peak positions and peak intensities of all absorption peaks in the whole or part of wavenumber range.
Further, in the step a), in the infrared spectrum of each trace sample, absorption peak data in a range from amide I, amide II, amide III, amide A, PO 2-antisymmetric stretching vibration to PO2 symmetric stretching vibration corresponding to wave number is extracted as an infrared spectrum data set.
Preferably, in step a), the infrared spectrum is collected over the following wavenumber ranges: 4000cm-1-600cm-1。
In the step c), the collection wave number range of the infrared absorption spectrum data of the trace sample to be detected is 1800cm-1-1300cm-1。
Preferably, in step b), the training type of the neural network is a quantitative conjugate gradient method, an L-M optimization algorithm or a gradient descent method with momentum.
In step b), the obtained neural network decision coefficient R2>0.9。
Preferably, the trace sample is a blood trace, handwriting, sweat stain or finger print.
The invention explores the change rule of the relative property of the in-vitro trace along with the time environment, expands the thinking of acquiring evidence and checking and identifying trace in site investigation on the basis of the traditional site investigation, carries out checking analysis and summary on the trace (such as blood trace, sweat trace or handwriting) left on the case site from the view point of trace physical evidence, and researches the change rule of the trace in-vitro or generated trace along with the time and environmental factors.
According to the invention, the attenuation total reflection infrared spectrum is combined with a neural network method to predict the remaining time of the trace, and the attenuation total reflection infrared spectrum technical method is used to collect the infrared spectrum of the trace, so that the basic characteristics of the trace can be clearly reflected. The infrared spectra collected at different times after the trace is left on the same substrate can reflect the basic rule that the infrared property of the trace on the substrate changes along with the time, and the rule is not transferred due to the change of the individual trace left.
In the invention, the absorption peaks of the amide I and the amide II are two peak values with highest peak intensity of the blood stain in an infrared spectrogram, the linear relation between the two peak values is good, and the wave number is 1536cm-1The intensity value of the amide I peak at (A) was 1650cm with respect to the wave number-1The intensity value of the amide I peak has a good linear relation, so that on one hand, the blood trace infrared spectrum data acquired by the attenuated total reflection infrared spectrum method has relatively stable and available conditions; on the other hand, the relative relationship between the two peaks is less influenced by factors such as uneven samples and experimental operation difference, and detailed research on key peak positions is facilitated.
Compared with the prior art, the invention has the following beneficial effects:
the method combines the traditional criminal science and technology research with spectroscopy and computer science, explores analysis and identification of the criminal science and technology trace leaving time, trains a neural network by taking the leaving time of sample traces and infrared spectrum data as variables, establishes a corresponding relation model of the infrared spectrum data and the leaving time of the sample traces, screens out a model with higher precision through verification, can realize nondestructive rapid detection of trace inspection materials, has controllable labor cost and later maintenance and consumable cost during use, has popularization conditions, and fills up the technical blank of trace sample leaving time detection in the technical field of criminal science and technology.
The invention trains the neural network by adopting the full-band spectrogram data, has higher measurement precision, can further improve the measurement precision by adjusting the range of key wave numbers, has obvious infrared absorption characteristic band of trace, good reproducibility, simple and convenient sample preparation and stable and reliable result.
Drawings
FIGS. 1-7 are the IR data spectra of samples D1YP1-1, D2YP1-1, D3YP1-1, D4YP1-1, D5YP1-1, D6YP1-1 and D7YP1-1 in this order in this invention.
Detailed Description
The present invention is further illustrated by the following specific examples.
The embodiment of the invention relates to a method for measuring trace remaining time based on an attenuated total reflection infrared spectrum, which comprises the following steps:
1. collecting infrared spectrum data of material detection trace
Collecting infrared spectrum data of a blood mark with terylene substrate chromatographic silica gel as a carrier, wherein the collection conditions are as follows:
an experimental instrument: bruker V70 fourier transform infrared spectrometer (with microscope accessory and ATR accessory);
scanning range: 4000cm-1-600cm-1；
Spatial resolution: 25 μm-1；
The test method comprises the following steps: attenuated total reflectance infrared spectroscopy (infrared microscope);
the scanning times are as follows: 32 times;
temperature: 21 ℃;
humidity: 40 percent.
Using an infrared microscope with ATR accessory at 4000cm-1-600cm-1The wavelength band of (a) is the scan range, as shown in table 1, wherein,"… …" indicates partial infrared spectral data not all listed in the preceding and following wavenumber ranges:
TABLE 1
Wave number Absorption peak intensity Wave number Absorption peak intensity Wave number Absorption peak intensity
4001.49232 0.00033 2985.20969 0.00054 1000.85519 0.03265
3999.56389 0.0009 2983.28126 0.00029 998.92676 0.03175
3997.63546 0.00132 …… …… 996.99833 0.03082
3995.70703 0.0021 2015.20938 0.00059 995.0699 0.02997
3993.7786 0.00269 2013.28095 0.00056 993.14147 0.02955
3991.85017 0.00241 2011.35252 0.00052 991.21304 0.02964
3989.92174 0.00166 2009.42409 0.00054 989.28461 0.02996
3987.99331 0.00109 2007.49566 0.00044 987.35618 0.02996
3986.06488 0.0009 2005.56723 0.00037 985.42775 0.02932
3984.13645 0.00104 2003.6388 0.00058 983.49932 0.02887
…… …… 2001.71037 0.00085 …… ……
3010.27928 0.00123 1999.78194 0.0008 622.8829 0.00852
3008.35085 0.00097 1997.85351 0.00051 620.95447 0.00642
3006.42242 0.00098 1995.92508 0.00017 619.02604 0.00503
3004.49399 0.00138 1993.99665 0 617.09761 0.006
3002.56556 0.00166 1992.06822 0.00027 615.16918 0.00765
3000.63713 0.00143 1990.13979 0.00053 613.24075 0.00557
2998.7087 0.00125 1988.21136 0.00051 611.31232 0.0018
2996.78027 0.00062 …… …… 609.38389 0
2994.85184 0 1010.49734 0.03978 607.45546 0.00402
2992.92341 0.00008 1008.56891 0.03851 605.52703 0.00985
2990.99498 0.00091 1006.64048 0.03707 603.5986 0.01127
2989.06655 0.00168 1004.71205 0.03539 601.67017 0.00616
2987.13812 0.00143 1002.78362 0.03381 599.74174 0
2. Simulating material-testing trace to make sample trace and classifying sample
Preparing a sample trace by using terylene substrate chromatographic silica gel as a carrier;
sampling: fresh venous blood samples provided by 10 volunteers are collected, 150 mu L of blood is dripped on a chromatographic silica gel sheet of a polyester substrate for each person to prepare sample blood traces, the sample blood traces are stood in a room with the temperature of 21 ℃ and the humidity of 40% for storage, 7 of the blood trace samples are used as training set samples, and 3 of the blood trace samples are used as verification set samples.
3. Collecting sample data set
And (3) placing 10 experimental samples in an experimental environment, timing after blood collection, wherein the sampling time is 1d after 24 hours after blood collection, and repeating until 7 d. Sampling is carried out at 1d, 2d, 3d, 4d, 5d, 6d and 7d after blood collection. During sampling, an experimental sample is placed on an objective table of an experimental instrument, 3 sampling points are marked under an infrared microscope, the infrared spectrum of the 3 sampling points is repeatedly sampled every day, and one blank sample on the surface of a chromatographic silica gel sheet of the polyester substrate is prepared. A maximum value of the leave-time of the sample trace 7 is recorded.
And recording sample data by using OPUS software, numbering the collected sample data, taking D1YP1-1 as an example, D is days, YP1-1 represents data on the first point on the sample No. 1, obtaining 210 parts of effective data in total, and dividing the effective data into 147 training set blood trace samples and 63 verification set blood trace samples according to the sample classification.
Respectively collecting 147 training set blood trace samples and 63 verification set blood trace samplesCollecting each sample in the wavenumber range of 4000cm-1-600cm-1And obtaining infrared spectrum data sets including an infrared spectrum data set (a training set for short) of the blood mark sample of the training set and an infrared spectrum data set (a verification set for short) of the blood mark sample of the verification set.
The infrared spectrogram of 210 blood mark samples is clear in the wave number range of important functional groups such as amide I, amide II, amide A and the like, and the infrared spectrogram of the sample No. 1 from day 1 to day 7 is shown in the attached figures 1-7 by taking the first point as an example.
The strongest peak values in the wavenumber range corresponding to the data of amide I, amide II, amide III, amide A, PO 2-antisymmetric stretching vibration and PO2 symmetric stretching vibration of each sample were extracted and considered, see Table 2.
TABLE 2
Sample name Amide I1650 Amide II 1536 Amide A3289 Amide III 1315 PO 2-antisymmetric telescopic vibration 1080 PO2 antisymmetric contraction vibration 1238
D1YP1-1 0.08089 0.05406 0.01598 0.00294 0.1304 0.03477
D1YP1-2 0.18241 0.11802 0.04344 0.00515 0.22214 0.04731
D1YP1-3 0.09355 0.05604 0.01979 0.00134 0.33532 0.02809
D1YP2-1 0.01106 0.01362 0.00363 0.00477 0.01466 0.00187
D1YP2-2 0.02197 0.01493 0.0051 0.00118 0.12493 0.01393
D1YP2-3 0.16845 0.1106 0.03129 0.00438 0.28947 0.02753
D1YP3-1 0.22003 0.15852 0.04593 0.00799 0.12556 0.00569
D1YP3-2 0.01013 0.01072 0.00367 0.0045 0.00742 0.02028
D1YP3-3 0.05871 0.0387 0.01103 0.00228 0.16031 0.02467
D1YP4-1 0.03004 0.02278 0.00688 0.00226 0.10585 0.03093
D1YP4-2 0.15073 0.0979 0.03321 0.00707 0.21574 0.05183
D1YP4-3 0.01859 0.0143 0.00263 0.00338 0.02942 0.01941
D1YP5-1 0.04717 0.03486 0.0053 0.00296 0.02087 0.01301
D1YP5-2 0.11054 0.06811 0.02179 0.00345 0.19702 0.04177
D1YP5-3 0.00666 0.00716 0.00214 0.00145 0.07659 0.02617
D1YP6-1 0.21577 0.17509 0.04426 0.031 0.02088 0.05368
D1YP6-2 0.05492 0.03784 0.00929 0.00122 0.14774 0.02496
D1YP6-3 0.02757 0.01927 0.00344 0.00109 0.11637 0.01164
D1YP7-1 0.04081 0.02379 0.0077 0.00024 0.22219 0.03846
D1YP7-2 0.02192 0.01451 0.00316 0.00036 0.16243 0.02503
D1YP7-3 0.21577 0.17509 0.04426 0.031 0.02088 0.05368
D1YP8-1 0.05492 0.03784 0.00929 0.00122 0.14774 0.02496
D1YP8-2 0.02757 0.01927 0.00344 0.00109 0.11637 0.01164
D1YP8-3 0.04081 0.02379 0.0077 0.00024 0.22219 0.03846
D1YP9-1 0.02192 0.01451 0.00316 0.00036 0.16243 0.02503
D1YP9-2 0.01081 0.0122 0.00287 0.00316 0.03168 0.00046
D1YP9-3 0.03723 0.02209 0.00674 0.00015 0.25687 0.02738
D1YP10-1 0.01301 0.01057 0.00311 0.00138 0.08874 0.02734
D1YP10-2 0.03412 0.02175 0.00469 0.0003 0.22982 0.04224
D1YP10-3 0.01707 0.01495 0.00387 0.00238 0.09031 0.03215
D2YP1-1 0.14099 0.09589 0.03454 0.00119 0.22462 0.01499
D2YP1-2 0.105 0.07966 0.02657 0.00343 0.08786 0.03049
D2YP1-3 0.0638 0.04679 0.01451 -0.00004 0.16789 0.01567
D2YP2-1 0.19096 0.13358 0.04617 0.00614 0.16077 0.03129
D2YP2-2 0.03523 0.0257 0.01158 0.00142 0.16833 0.03542
D2YP2-3 0.20601 0.17759 0.05072 0.03391 0.04814 0.05525
D2YP3-1 0.11147 0.0817 0.02686 0.004 0.10574 0.02959
D2YP3-2 0.01037 0.01042 0.00702 0.00076 0.02013 0.01037
D2YP3-3 0.06237 0.04308 0.01542 0.00189 0.15756 0.02826
D2YP4-1 0.04722 0.03337 0.01009 0.0007 0.18717 0.01154
D2YP4-2 0.07786 0.05543 0.02085 0.00163 0.0989 0.02065
D2YP4-3 0.19836 0.13065 0.0455 0.00797 0.25106 0.04816
D2YP5-1 0.01942 0.01451 0.00462 0.00095 0.0405 0.00752
D2YP5-2 0.04742 0.03801 0.00796 0.00388 0.02295 0.01432
D2YP5-3 0.05091 0.03441 0.00644 0.00081 0.1479 0.01042
D2YP6-1 0.06275 0.04736 0.01061 0.00232 0.05311 0.00309
D2YP6-2 0.06432 0.04094 0.01093 0.00142 0.26904 0.01901
D2YP6-3 0.0504 0.03368 0.00932 0.00013 0.22259 0.02952
D2YP7-1 0.04535 0.02795 0.00799 -0.00026 0.32813 0.04209
D2YP7-2 0.02254 0.01488 0.00604 0.00111 0.15686 0.03195
D2YP7-3 0.06275 0.04736 0.01061 0.00232 0.05311 0.00309
D2YP8-1 0.06432 0.04094 0.01093 0.00142 0.26904 0.01901
D2YP8-2 0.0504 0.03368 0.00932 0.00013 0.22259 0.02952
D2YP8-3 0.04535 0.02795 0.00799 -0.00026 0.32813 0.04209
D2YP9-1 0.02254 0.01488 0.00604 0.00111 0.15686 0.03195
D2YP9-2 0.08489 0.05132 0.02128 0.00063 0.32017 0.03696
D2YP9-3 0.03501 0.02431 0.00596 0.00036 0.23966 0.02576
D2YP10-1 0.02242 0.0158 0.00656 0.00043 0.1319 0.01342
D2YP10-2 0.03179 0.02284 0.00673 -0.00019 0.22954 0.02373
D2YP10-3 0.03256 0.02177 0.00857 0.00043 0.2301 0.03148
D3YP1-1 0.05401 0.03719 0.01132 0.00306 0.10351 0.01695
D3YP1-2 0.14379 0.0919 0.03807 0.00447 0.25191 0.0479
D3YP1-3 0.04599 0.0274 0.00919 0.00414 0.15445 0.03734
D3YP2-1 0.02212 0.01639 0.00735 0.0018 0.07849 0.00626
D3YP2-2 0.02939 0.01584 0.00744 0.00247 0.21956 0.04261
D3YP2-3 0.22217 0.18965 0.0513 0.05108 0.06586 0.07621
D3YP3-1 0.12576 0.1041 0.03022 0.01658 0.13519 0.02832
D3YP3-2 0.00588 0.01866 0.00316 0.01938 0.03328 0.04826
D3YP3-3 0.09794 0.07345 0.02738 0.01488 0.24178 0.05386
D3YP4-1 0.04119 0.03907 0.01143 0.01663 0.14282 0.04577
D3YP4-2 0.05268 0.04416 0.01483 0.01628 0.1113 0.04119
D3YP4-3 0.18374 0.11984 0.0397 0.00955 0.24753 0.03944
D3YP5-1 0.00709 0.00769 0.00223 0.00676 0.01087 0.0131
D3YP5-2 0.03435 0.02669 0.00659 0.00767 0.02317 0.02075
D3YP5-3 0.05818 0.03813 0.01041 0.00124 0.15935 0.00698
D3YP6-1 0.04345 0.03471 0.01156 0.00864 0.03397 0.02176
D3YP6-2 0.03527 0.02128 0.00805 0.00347 0.13724 0.03263
D3YP6-3 0.06817 0.04079 0.01063 0.00319 0.28102 0.04666
D3YP7-1 0.02951 0.01602 0.00588 0.00167 0.22565 0.03591
D3YP7-2 0.0391 0.02164 0.00749 0.00138 0.25287 0.04196
D3YP7-3 0.04345 0.03471 0.01156 0.00864 0.03397 0.02176
D3YP8-1 0.03527 0.02128 0.00805 0.00347 0.13724 0.03263
D3YP8-2 0.06817 0.04079 0.01063 0.00319 0.28102 0.04666
D3YP8-3 0.02951 0.01602 0.00588 0.00167 0.22565 0.03591
D3YP9-1 0.0391 0.02164 0.00749 0.00138 0.25287 0.04196
D3YP9-2 0.07131 0.04436 0.01572 0.00092 0.25308 0.02301
D3YP9-3 0.03165 0.01882 0.00376 0.00022 0.35489 0.01367
D3YP10-1 0.0065 0.0048 0.00289 0.00258 0.04835 0.01818
D3YP10-2 0.04318 0.02143 0.00506 0.00068 0.37513 0.03453
D3YP10-3 0.03607 0.02148 0.00874 0.00118 0.18703 0.03476
D4YP1-1 0.0239 0.01951 0.00763 0.00095 0.04035 0.00524
D4YP1-2 0.115 0.07917 0.02549 0.00156 0.14182 0.01966
D4YP1-3 0.05883 0.03798 0.00868 0.00196 0.23941 0.01007
D4YP2-1 0.02724 0.02068 0.00667 0.00144 0.0758 0.00737
D4YP2-2 0.01509 0.01295 0.0052 0.00099 0.0759 0.0091
D4YP2-3 0.19176 0.12844 0.0407 0.00264 0.20994 0.02328
D4YP3-1 0.08068 0.06284 0.01677 0.00244 0.07657 0.00753
D4YP3-2 0.00944 0.01029 0.00571 0.00194 0.00702 0.00914
D4YP3-3 0.10097 0.06426 0.0192 0.00163 0.25192 0.02816
D4YP4-1 0.05803 0.03907 0.01143 0.00139 0.1496 0.0248
D4YP4-2 0.03347 0.02179 0.00472 0.00103 0.25085 0.0182
D4YP4-3 0.13959 0.09677 0.01815 0.00256 0.40862 0.0082
D4YP5-1 0.16087 0.10958 0.02664 0.00249 0.28953 0.02241
D4YP5-2 0.16171 0.09837 0.0315 0.00211 0.27407 0.02358
D4YP5-3 0.14848 0.0941 0.02838 0.00165 0.25105 0.03131
D4YP6-1 0.15228 0.10442 0.0311 0.00279 0.16577 0.0209
D4YP6-2 0.05589 0.03289 0.00838 0.00094 0.20004 0.02251
D4YP6-3 0.08265 0.05007 0.00961 0.00114 0.2776 0.02019
D4YP7-1 0.0587 0.03248 0.00937 0.00066 0.3017 0.02693
D4YP7-2 0.05713 0.03143 0.00761 0.00034 0.32705 0.03402
D4YP7-3 0.15228 0.10442 0.0311 0.00279 0.16577 0.0209
D4YP8-1 0.05589 0.03289 0.00838 0.00094 0.20004 0.02251
D4YP8-2 0.08265 0.05007 0.00961 0.00114 0.2776 0.02019
D4YP8-3 0.0587 0.03248 0.00937 0.00066 0.3017 0.02693
D4YP9-1 0.05713 0.03143 0.00761 0.00034 0.32705 0.03402
D4YP9-2 0.03859 0.02051 0.00686 0.00005 0.25089 0.02736
D4YP9-3 0.02302 0.01483 0.004 0.00075 0.19234 0.01555
D4YP10-1 0.02623 0.01584 0.0035 0.00028 0.23632 0.0252
D4YP10-2 0.03019 0.01852 0.00428 0.00083 0.18727 0.03251
D4YP10-3 0.01042 0.00876 0.00286 0.00044 0.07556 0.01226
D5YP1-1 0.07723 0.05588 0.02856 0.0126 0.12958 0.02034
D5YP1-2 0.09295 0.06794 0.03664 0.0183 0.2128 0.04516
D5YP1-3 0.05697 0.04745 0.02115 0.0275 0.17975 0.03629
D5YP2-1 0.07045 0.04795 0.02713 0.00916 0.18077 0.02758
D5YP2-2 0.02231 0.02252 0.01208 0.02171 0.02524 0.04046
D5YP2-3 0.15494 0.13254 0.06704 0.05114 0.22358 0.09948
D5YP3-1 0.21006 0.15889 0.07658 0.037 0.22867 0.06468
D5YP3-2 0.18802 0.138 0.06494 0.04148 0.21518 0.07512
D5YP3-3 0.11732 0.08058 0.04961 0.02443 0.2776 0.06904
D5YP4-1 0.06137 0.04362 0.02141 0.01637 0.17511 0.03459
D5YP4-2 0.02648 0.02423 0.01115 0.02183 0.09718 0.04703
D5YP4-3 0.20292 0.14723 0.07778 0.04303 0.24488 0.08241
D5YP5-1 0.12585 0.11091 0.04334 0.0667 0.11672 0.06435
D5YP5-2 0.05202 0.04499 0.01392 0.02668 0.02703 0.03691
D5YP5-3 0.0627 0.0495 0.02202 0.0209 0.16673 0.02881
D5YP6-1 0.05943 0.04892 0.02641 0.01796 0.06576 0.02534
D5YP6-2 0.08646 0.07716 0.02567 0.07512 0.19755 0.09069
D5YP6-3 0.1498 0.1089 0.0587 0.03587 0.39988 0.07066
D5YP7-1 0.04234 0.02448 0.02187 0.0054 0.32279 0.04489
D5YP7-2 0.02233 0.01456 0.01063 0.00569 0.20359 0.03308
D5YP7-3 0.05943 0.04892 0.02641 0.01796 0.06576 0.02534
D5YP8-1 0.08646 0.07716 0.02567 0.07512 0.19755 0.09069
D5YP8-2 0.1498 0.1089 0.0587 0.03587 0.39988 0.07066
D5YP8-3 0.04234 0.02448 0.02187 0.0054 0.32279 0.04489
D5YP9-1 0.02233 0.01456 0.01063 0.00569 0.20359 0.03308
D5YP9-2 0.02638 0.01781 0.01764 0.01051 0.1764 0.04667
D5YP9-3 0.04781 0.0369 0.023 0.02459 0.42332 0.06618
D5YP10-1 0.02744 0.01904 0.01882 0.01153 0.1601 0.03803
D5YP10-2 0.02656 0.01776 0.01678 0.00713 0.17441 0.03989
D5YP10-3 0.03258 0.02194 0.02225 0.00513 0.19853 0.02527
D6YP1-1 0.05348 0.04235 0.01289 0.00366 0.08298 0.01314
D6YP1-2 0.10796 0.07701 0.02608 0.00651 0.20752 0.04086
D6YP1-3 0.06804 0.05061 0.01653 0.00418 0.16127 0.0297
D6YP2-1 0.07973 0.0575 0.02295 0.00193 0.1816 0.01905
D6YP2-2 0.02395 0.02332 0.01026 0.00751 0.086 0.03351
D6YP2-3 0.1078 0.08041 0.02045 0.00931 0.25514 0.07536
D6YP3-1 0.16883 0.12768 0.04446 0.01012 0.18353 0.05078
D6YP3-2 0.03811 0.0336 0.01102 0.00838 0.06324 0.03079
D6YP3-3 0.13154 0.09113 0.03044 0.00547 0.35945 0.05266
D6YP4-1 0.10439 0.07674 0.02551 0.00228 0.22212 0.02632
D6YP4-2 0.02966 0.02931 0.01139 0.00866 0.08187 0.03972
D6YP4-3 0.11724 0.08606 0.02434 0.00825 0.13049 0.03288
D6YP5-1 0.05708 0.04879 0.01381 0.01517 0.01199 0.03031
D6YP5-2 0.03981 0.03213 0.00595 0.00563 0.024 0.01281
D6YP5-3 0.00736 0.00744 0.00228 0.00274 0.14909 0.02556
D6YP6-1 0.03718 0.03072 0.00835 0.00292 0.0294 0.0094
D6YP6-2 0.04269 0.02563 0.0129 0.00165 0.1597 0.02212
D6YP6-3 0.08828 0.05934 0.01547 0.00215 0.38835 0.02412
D6YP7-1 0.05209 0.03461 0.01104 0.00215 0.28096 0.03988
D6YP7-2 0.05158 0.03207 0.01013 0.00264 0.30859 0.04272
D6YP7-3 0.03718 0.03072 0.00835 0.00292 0.0294 0.0094
D6YP8-1 0.04269 0.02563 0.0129 0.00165 0.1597 0.02212
D6YP8-2 0.08828 0.05934 0.01547 0.00215 0.38835 0.02412
D6YP8-3 0.05209 0.03461 0.01104 0.00215 0.28096 0.03988
D6YP9-1 0.05158 0.03207 0.01013 0.00264 0.30859 0.04272
D6YP9-2 0.04418 0.02546 0.00922 0.00147 0.20001 0.03113
D6YP9-3 0.03638 0.02299 0.00666 0.00039 0.31614 0.02562
D6YP10-1 0.02969 0.04283 0.00505 0.04396 0.14356 0.06491
D6YP10-2 0.02871 0.02411 0.00782 0.00572 0.15739 0.04436
D6YP10-3 0.01984 0.01458 0.00628 0.00093 0.17616 0.02171
D7YP1-1 0.02692 0.02187 0.00822 0.00754 0.02555 0.02032
D7YP1-2 0.02645 0.02363 0.0097 0.0074 0.05327 0.02282
D7YP1-3 0.07286 0.05166 0.0131 0.00878 0.08689 0.02909
D7YP2-1 0.03814 0.02962 0.01204 0.00759 0.09519 0.03854
D7YP2-2 0.20016 0.17207 0.04393 0.04629 0.04212 0.06953
D7YP2-3 0.03352 0.02614 0.00862 0.008 0.1238 0.0439
D7YP3-1 0.04508 0.03253 0.01041 0.00595 0.14172 0.0209
D7YP3-2 0.15413 0.09291 0.03021 0.01185 0.34282 0.06181
D7YP3-3 0.05645 0.04456 0.01448 0.0099 0.0804 0.04408
D7YP4-1 0.07666 0.05692 0.01524 0.05481 0.21794 0.08633
D7YP4-2 0.14207 0.10034 0.0303 0.01027 0.15727 0.04022
D7YP4-3 0.1328 0.12076 0.04006 0.04428 0.17335 0.07534
D7YP5-1 0.08788 0.06016 0.01497 0.00542 0.09872 0.02238
D7YP5-2 0.01664 0.01167 0.00278 0.00168 0.16591 0.02152
D7YP5-3 0.13792 0.1058 0.03979 0.01358 0.08637 0.03336
D7YP6-1 0.06704 0.04478 0.01362 0.00367 0.19577 0.03014
D7YP6-2 0.10989 0.07425 0.02263 0.00739 0.29172 0.04205
D7YP6-3 0.07005 0.04272 0.01391 0.00214 0.30298 0.04212
D7YP7-1 0.18129 0.22104 0.21365 0.36247 0.54057 0.38038
D7YP7-2 0.13792 0.1058 0.03979 0.01358 0.08637 0.03336
D7YP7-3 0.06704 0.04478 0.01362 0.00367 0.19577 0.03014
D7YP8-1 0.10989 0.07425 0.02263 0.00739 0.29172 0.04205
D7YP8-2 0.07005 0.04272 0.01391 0.00214 0.30298 0.04212
D7YP8-3 0.18129 0.22104 0.21365 0.36247 0.54057 0.38038
D7YP9-1 0.03862 0.02573 0.01052 0.00089 0.18869 0.02811
D7YP9-2 0.02662 0.01576 0.00472 0.00131 0.15501 0.0223
D7YP9-3 0.00826 0.00773 0.0048 0.00282 0.04177 0.01449
D7YP10-1 0.05086 0.02783 0.01095 0.0013 0.34475 0.03973
D7YP10-2 0.03712 0.01818 0.00634 0.00014 0.29957 0.0246
D7YP10-3 0.01442 0.01096 0.00466 0.00461 0.13774 0.0326
From the data of 210 samples taken, all samples at a wavenumber of 1080cm can be seen in conjunction with Table 2-1Left and right, wave number 1650cm-1And wave number 1535cm-1Strong absorption peaks appear at the left and right sides, and the wave number is 1315cm-1、1392cm-1、1452cm-1、2854cm-1、2873cm-1、2935cm-1、2960cm-1、3012cm-1、3289cm-1The absorption peaks appear at the left and right. The vast majority of samples were at a wavenumber of 1650cm-1And wave number 1535cm-1The left and right absorption peaks have similar heights, and most samples have a wave number of 1650cm-1Has a peak intensity of greater than 1535cm at wave number-1The peak intensity of (c). Most samples were at wavenumber 1080cm-1The absorption peak intensity was greatest at the left and right, and all samples had a wave number of 3289cm-1Has a moderate intensity absorption peak with peak intensity less than 1080cm-1Wave number 1650cm-1And wave number 1535cm-1Absorption peak and more than 4000cm wave number-1-1300cm-1Other peaks of functional groups.
The blood trace infrared spectrum data acquired by the attenuated total reflection infrared spectroscopy have relatively stable and available conditions; on the other hand, the wave number was 1536cm-1The peak intensity value at (A) with respect to the wave number was 1650cm-1The peak intensity values at (a) correspond to y =0.7744x-0.002942 (R)2=0.9281), which means that the relative relationship between the two peaks of the amide I and the amide II is less influenced by factors such as sample nonuniformity and experimental operation difference, and is beneficial to detailed research on the important peak positions.
4. Establishing a corresponding relation model of infrared spectrum data of sample traces and the remaining time by using a neural network and verifying the precision,
establishing a BP neural network by using the function of a neural network toolbox in MATLAB data analysis software and taking the remaining time of a trace sample and infrared spectrum data as variables, wherein the training type is TRAINLM, namely an L-M optimization algorithm; selecting MSE as the error type; the number of layers is 2 (1 hidden layer and 1 output layer), the transfer function of the hidden layer and the output layer is TANSIL with default system, and because the L-M optimization algorithm needs to occupy a large amount of memory resources, the requirement of overlarge neuron number on a computer is high, three neural networks of network _ all _ train _ n1, network _ all _ train _ n3 and network _ all _ train _ n5 are constructed by respectively setting the neuron number to 1, 3 and 5.
The three neural networks with the built structures are respectively trained by utilizing the training set, time is set to be 180 in training parameters (the training time of network _ all _ training _ n5 is 300), the max _ fail value is set to be 20, the rest parameters are default parameters, variable AAall is used as input data inputs, TrainDay is used as target data Targets, the neural networks are trained, after the neural networks are trained according to preset parameters, APall is used as the input data inputs to predict the corresponding days, and the predicted days are respectively stored as variables network _ all _ training _ n1_ outputs, network _ all _ in _ n3_ outputs and network _ all _ training _ n5_ outputs.
And (4) verifying by using the verification set, inputting infrared spectrum data in the verification set into the three groups of models obtained by training, and feeding back corresponding remaining days in the simulated verification set by the three groups of models.
Comparing the corresponding remaining days and the real remaining days in the verification set of the three groups of model simulations by using a data processing tool, and calculating a judgment coefficient R of the remaining days20.7071, 0.8435, 0.5677 respectively, and therefore a model built using the neural network _ all _ rainlm _ n3 is available.
Determination coefficient R in the present invention2Also called a decision coefficient, which has the main function of evaluating the interpretation degree of the regression model on the change of the dependent variable and is an index for judging the accuracy degree of the model, generally speaking, the decision coefficient R2Greater than 0.8 indicates good fitting of the model, and the calculation formula is as follows:
R2= regression squared sum/total squared sum, where regression squared sum = total squared sum-residual squared sum.
5. Inputting the infrared spectrum data of the material detection trace into a model meeting the precision requirement to obtain the leave-behind time of the simulated material detection trace
And (3) recording infrared spectrum data of the material detection trace into the model established by the neural network _ all _ train _ n3, and feeding back the simulated material detection trace remaining time D =5.5 by the model.
6. Evaluation of remaining time of material trace
The model feedback simulation shows that the leave-behind days of the material detection trace D =5.5 and the ratio of the leave-behind days of the material detection trace to the maximum value 7 of the leave-behind time of the sample trace is less than or equal to 0.9, so that the leave-behind days of the material detection trace is evaluated to be 5.5.

Claims (5)

1. A method for measuring trace remaining time based on an attenuated total reflection infrared spectrum is characterized by comprising the following steps:
1) collecting infrared spectrum data of a material detection trace by using an attenuated total reflection infrared spectroscopy, wherein the material detection trace comes from a scene, and the material detection trace is a blood mark, a handwriting, a sweat spot or a finger print;
2) simulating the material detecting trace to make sample trace, classifying the sample,
3) collecting a sample data set, wherein the data set comprises infrared spectrum data of sample traces with different remaining time and corresponding remaining time,
4) establishing a corresponding relation model of infrared spectrum data of sample traces and the remaining time by using a neural network and verifying the precision,
5) inputting the infrared spectrum data of the material detecting trace into a model meeting the precision requirement to obtain the leave-behind time of the simulated material detecting trace,
6) evaluating the material detecting trace remaining time, comparing the simulated material detecting trace remaining time with the maximum value of the sample trace remaining time, if the ratio is less than or equal to 0.9, evaluating the simulated material detecting trace remaining time as the measured material detecting trace remaining time, if the ratio is greater than 0.9, expanding the sample data set, prolonging the sample trace remaining time, repeating the steps 3) to 5), comparing the simulated material detecting trace remaining time after expanding the sample data set with the maximum value of the sample trace remaining time until the ratio is less than or equal to 0.9, evaluating the simulated material detecting trace remaining time as the measured material detecting trace remaining time, and completing the measurement.
2. The method for measuring the trace remaining time based on the attenuated total reflection infrared spectroscopy as claimed in claim 1, wherein in the step 2), the sample trace is made by using the same material as the material detection trace carrier, and the sample trace is processed according to the following steps of 7: 3 proportion, and randomly dividing the trace into a training sample trace and a verification sample trace.
3. The method for measuring the trace remaining time based on the attenuated total reflection infrared spectroscopy of claim 1, wherein in the step 3), the remaining environment of the trace of the material to be tested is simulated, the attenuated total reflection infrared spectroscopy is used, the wave number range and the resolution are required to be the same as those in the step 1), the infrared spectroscopy data of the sample trace with different remaining times are collected at uniform time intervals, the infrared spectroscopy data and the corresponding remaining times are summarized into a sample data set, the maximum value of the remaining time of the sample trace is recorded, the corresponding relation between the infrared spectroscopy data and the remaining times in the sample data set is reserved, and the sample data set is divided into a training set and a verification set according to training and verification.
4. The method for measuring the trace remaining time based on the attenuated total reflection infrared spectroscopy as claimed in claim 3, wherein in the step 4), the neural network is trained by using the infrared spectrum data and the remaining time in the training set as variables, the model of the corresponding relationship between the infrared spectrum data and the remaining time of the trace of the sample is established, and the validation set is used to verify the accuracy of the model, such as the determination coefficient R2>0.8, the model is available, otherwise, the process from training the neural network to verifying the model precision by using the verification set is repeated until a judgment coefficient R2>0.8。
5. The method for measuring the trace remaining time based on the attenuated total reflection infrared spectroscopy as claimed in claim 1, wherein in the step 5), the infrared spectrum data of the material detection trace is inputted into the available model, and the simulated remaining time of the material detection trace is obtained.
CN202010452130.4A 2020-05-26 2020-05-26 Method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy Active CN111366554B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010452130.4A CN111366554B (en) 2020-05-26 2020-05-26 Method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010452130.4A CN111366554B (en) 2020-05-26 2020-05-26 Method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy

Publications (2)

Publication Number Publication Date
CN111366554A CN111366554A (en) 2020-07-03
CN111366554B true CN111366554B (en) 2020-08-11

Family

ID=71207539

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010452130.4A Active CN111366554B (en) 2020-05-26 2020-05-26 Method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy

Country Status (1)

Country Link
CN (1) CN111366554B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112098362B (en) * 2020-09-15 2024-02-27 南京工程学院 Cancellous bone in-vitro time assessment method and system based on near infrared spectrum characteristics
CN113293217B (en) * 2021-05-27 2022-02-18 山西医科大学 Method for identifying blood trace remaining time and mRNA (messenger ribonucleic acid) composition thereof

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090036781A1 (en) * 2007-07-31 2009-02-05 Kei Utsugi External condition control device based on measurement of brain functions
US20190285550A1 (en) * 2015-07-30 2019-09-19 The Research Foundation For The State University Of New York Spectroscopic methods for body fluid age determination

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090036781A1 (en) * 2007-07-31 2009-02-05 Kei Utsugi External condition control device based on measurement of brain functions
US20190285550A1 (en) * 2015-07-30 2019-09-19 The Research Foundation For The State University Of New York Spectroscopic methods for body fluid age determination

Also Published As

Publication number Publication date
CN111366554A (en) 2020-07-03

Similar Documents

Publication Publication Date Title
CN111366554B (en) Method for measuring trace remaining time based on attenuated total reflection infrared spectroscopy
CN104198432B (en) The method of application near-infrared spectrum technique Undamaged determination fire-resistant coating for steel structure brand
CN109799195B (en) High-precision quantitative analysis method for laser-induced breakdown spectroscopy
CN106124508A (en) A kind of verification retrieval system and authentication method
CN109064553B (en) Solid wood board section morphology inversion method based on near infrared spectrum analysis
CN111553877B (en) Damage identification and service life evaluation method based on terahertz ceramic matrix composite blade
CN102841070B (en) Method for identifying types of crude oil by using two-dimensional correlation infrared synchronization spectrum
CN106796197A (en) The method that sample is characterized by mass spectrum imaging
CN113008805B (en) Radix angelicae decoction piece quality prediction method based on hyperspectral imaging depth analysis
CN109669023A (en) A kind of soil attribute prediction technique based on Multi-sensor Fusion
CN110068544A (en) Material identification network model training method and tera-hertz spectra substance identification
Thickett et al. Using non-invasive non-destructive techniques to monitor cultural heritage objects
CN110082310A (en) A kind of near infrared band EO-1 hyperion diagnostic method of rubber tree LTN content
CN110208238A (en) It is a kind of based on SVM models coupling image to the accurate positioning method of cancerous lung tissue
CN105092513A (en) Method for determining content of drug components in mixture drug
CN110441254A (en) A kind of near-infrared frequency comb spectrometer of plastics for identification
CN108548794A (en) A kind of biological products method for transferring near infrared model
CN110070004B (en) Near-earth hyperspectral data expansion method applied to deep learning
CN110567894A (en) method and device for rapidly detecting pH value of aqueous solution
CN112716447A (en) Oral cancer classification system based on deep learning of Raman detection spectral data
CN108489928B (en) Method for detecting textile fiber components by short-wave infrared extinction spectrum
CN111220575A (en) Cell detection method based on terahertz near-field spectrum
Martin et al. Impact of the sampling procedure on the specific surface area of snow measurements with the IceCube
JPH02271211A (en) Evaluating method of painting sharpness
Galagan et al. The use of backpropagation artificial neural networks in thermal tomography

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