CN108254351B - Raman spectrum detection method for checking articles - Google Patents
Raman spectrum detection method for checking articles Download PDFInfo
- Publication number
- CN108254351B CN108254351B CN201611257547.5A CN201611257547A CN108254351B CN 108254351 B CN108254351 B CN 108254351B CN 201611257547 A CN201611257547 A CN 201611257547A CN 108254351 B CN108254351 B CN 108254351B
- Authority
- CN
- China
- Prior art keywords
- raman spectrum
- standard
- detected
- raman
- information
- 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
Links
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
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/65—Raman scattering
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/10—Scanning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention discloses a Raman spectrum detection method for checking objects, which comprises the following steps: and a Raman spectrum acquisition step: collecting Raman spectrum of the object to be detected; and comparing and judging: comparing the acquired Raman spectrum of the object to be detected with the Raman spectrum of the standard object stored in the standard database to judge whether the object to be detected is matched with the standard object. The comparing and judging steps comprise: and classifying the Raman spectrum of the object to be detected by adopting a support vector machine so as to realize the comparison of the Raman spectrum of the object to be detected and the Raman spectrum of the standard object stored in the standard database.
Description
Technical Field
The invention relates to the field of Raman spectrum detection, in particular to a Raman spectrum detection method for checking articles, and particularly relates to a Raman spectrum detection method for quickly checking special articles entering and exiting.
Background
Raman spectroscopy is a molecular vibration spectrum that reflects the fingerprint characteristics of molecules and can be used for detection of substances. Raman spectrum detection detects and identifies substances by detecting raman spectra generated by raman scattering effects of the analyte on excitation light. The Raman spectrum detection method is widely applied to the fields of liquid security inspection, jewelry detection, explosive detection, drug detection, pesticide residue detection and the like.
The existing field detection for the special articles in and out mainly adopts the real-time monitoring and supervision functions of a field high-definition camera device, field inspection personnel unpack the box to detect, whether the cargo information is consistent with the declaration form information or not is compared, and if the cargo information is consistent, the cargo is considered to be released.
However, the above-described method has mainly the following drawbacks: (1) the work load of the on-site inspection personnel is large. The checking of whether the approval ticket is consistent with the label information requires checking the label information one by a field checking staff, and manually comparing whether various information (mostly non-Chinese language information such as English) such as complex goods names, specifications, characters and quantity are consistent. (2) checking that the vulnerability is obvious. The checking mode of checking the tag information cannot verify whether the cargo tag information is consistent with the actual content, and cannot effectively identify the situation of using low-risk or non-special articles to pack, store or transport high-risk special articles.
Disclosure of Invention
The present invention has been made in order to solve at least one of the above-mentioned drawbacks.
The invention aims to provide a Raman spectrum detection method for checking articles, in particular to a Raman spectrum detection method for quickly checking special articles entering and exiting, which can enhance the accuracy of on-site checking of the articles, realize simultaneous matching checking of multiple information through one-time detection and accelerate the on-site checking efficiency of the articles.
In order to achieve the above object, the technical solution of the present invention is achieved by:
according to one aspect of the present invention, there is provided a raman spectrum detection method for inspection of an article, comprising the steps of:
and a Raman spectrum acquisition step: collecting Raman spectrum of the object to be detected; and
and (3) comparing and judging: comparing the collected Raman spectrum of the object to be detected with the Raman spectrum of the standard object stored in the standard database to judge whether the object to be detected is matched with the standard object,
the method is characterized in that the comparing and judging steps comprise: and classifying the Raman spectrum of the object to be detected by adopting a support vector machine so as to realize the comparison of the Raman spectrum of the object to be detected and the Raman spectrum of the standard object stored in the standard database.
According to some embodiments, the step of classifying raman spectra of the article to be inspected using a support vector machine comprises:
a step of establishing a classifier: selecting a training sample, measuring the training sample to obtain a Raman spectrum of the training sample, performing sparse transformation on the Raman spectrum of the training sample, and then establishing a classifier by adopting a support vector machine algorithm; and
classification: and classifying the Raman spectrum of the object to be detected by adopting the established classifier.
According to some embodiments, the sparse transform comprises the steps of:
acquiring a main component M of Raman spectrum information y of a training sample;
training samples are paired using the formula y=mcSparse representation is carried out on Raman spectrum information y of the spectrum to obtain a series of sparse representation vectors c i The method comprises the steps of carrying out a first treatment on the surface of the And
the target vector c sparsely representing the raman spectrum information y is calculated using the following target function:
wherein the number of non-zero elements in the vector c is fixed to be k, N is the length of the vector c, k is less than or equal to N, and i is the sequence number of the vector c.
According to some embodiments, the comparing and determining step further comprises:
carrying out similarity measurement on the Raman spectrum of the object to be detected and the Raman spectrum of the standard object stored in the standard database;
when the result of the similarity measurement is larger than a preset threshold value, judging that the object to be detected is matched with the standard object;
and when the result of the similarity measurement is equal to or lower than a preset threshold value, classifying the Raman spectrum of the object to be detected by adopting a support vector machine so as to realize the comparison of the Raman spectrum of the object to be detected and the Raman spectrum of the standard object stored in the standard database.
According to some embodiments, the performing similarity measurement between the raman spectrum of the object to be detected and the raman spectrum of the standard object stored in the standard database includes:
and calculating the correlation coefficient of the characteristic vector of the Raman spectrum of the object to be detected and the characteristic vector of the Raman spectrum of the standard object stored in the standard database, and taking the calculated correlation coefficient as a result of the similarity measurement.
According to some embodiments, the raman spectrum detection method further comprises the steps of:
establishing a standard database: the raman spectra of the surface package and the actual content of the standard article are collected sequentially to form a standard database comprising raman spectra of the surface components of the standard article and raman spectra of the components of the actual content.
According to some embodiments, the raman spectrum acquisition step comprises: scanning the surface and the interior of the object to be detected to acquire the Raman spectrum of the surface component of the object to be detected and the Raman spectrum of the component of the actual content;
the step of performing similarity measurement on the Raman spectrum of the object to be detected and the Raman spectrum of the standard object stored in the standard database comprises the following steps: respectively carrying out similarity measurement on the Raman spectrum of the surface component of the object to be detected and the Raman spectrum of the surface component of the standard object stored in the standard database, and the Raman spectrum of the component of the actual content of the object to be detected and the Raman spectrum of the component of the actual content of the standard object stored in the standard database; and is also provided with
The result of the similarity measure being greater than a preset threshold value only includes a case where the results of both similarity measures are greater than a preset threshold value.
According to some embodiments, the step of establishing a criteria database further comprises: and collecting the name, source company, specification size and picture information of the standard object, and storing the information into a standard database.
According to another aspect of the present invention, there is also provided an electronic apparatus including:
a memory for storing executable instructions; and
a processor for executing executable instructions stored in memory to perform the method of any one of the above aspects or embodiments.
According to the detection method provided by the embodiment of the invention, the process that the on-site inspection personnel compare multiple pieces of information of the object to be inspected one by one is avoided, and the on-site inspection speed is increased; meanwhile, the recognition method of characteristic information matching is adopted, so that the accuracy of on-site inspection is improved.
Drawings
FIG. 1 illustrates a flow chart of a detection method for an outbound special article inspection according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of the support vector machine principle;
FIG. 3 shows a flow chart for building a support vector machine classifier;
FIG. 4 shows a flow chart of a detection method employing an improved support vector machine in accordance with an embodiment of the present invention; and
fig. 5 shows a block diagram of an example hardware arrangement of an electronic device for performing a detection method according to an embodiment of the invention.
Detailed Description
The technical scheme of the invention is further specifically described below through examples and with reference to the accompanying drawings. In the specification, the same or similar reference numerals denote the same or similar components. The following description of embodiments of the present invention with reference to the accompanying drawings is intended to illustrate the general inventive concept and should not be taken as limiting the invention.
Herein, for convenience of description, the steps of the method are described using expressions such as "first, second", "A, B, C" or "S1/S10, S2/S20", etc., but such expressions should not be construed as limiting the order in which the steps are performed unless specifically stated.
Hereinafter, the technical idea of the present invention will be described in detail mainly taking as an example an entry-exit special article, which refers generally to microorganisms, human tissues, human genetic resources, biological products, blood products, etc. Because of the various special articles, various goods sources and complex package in the inspection and quarantine inspection site, many special articles have the characteristics of small sample size, difficult observation by naked eyes and unknown components; in addition, the circulation of special articles has greater transportation difficulty and potential risk compared with general goods; for on-site inspection, many special objects such as biological agents including microorganisms and blood cannot realize the unsealing detection of the inspection site. Together, these reasons determine that the detection of an outbound special item is a difficult inspection quarantine inspection job. By adopting the Raman spectrum detection method for checking the articles, which is disclosed by the embodiment of the invention, the special articles entering and exiting can be checked quickly, and meanwhile, the accuracy of on-site checking can be improved.
According to the embodiment of the invention, a complete integral Raman spectrum information base of the specific articles with the entrance and the exit needs to be established, and the information base can be called a standard database. Then, in the process of checking on site, selecting standard articles existing in a standard database according to the article names indicated on the goods bill, then carrying out surface and internal scanning on the articles to be checked, then comparing and analyzing the spectrogram of the articles to be checked obtained by scanning with the standard spectrogram in the standard database, and combining a similarity measurement algorithm and an improved support vector machine algorithm to finish the rapid characteristic identification of special articles entering and exiting.
When the standard database of the specific articles for entry and exit is established, the standard spectrogram of the representative specific articles for entry and exit can be established first, or can be added cumulatively in the using process or provided by a quarantine inspection unit. In one example, for each outbound special item, the raman signature spectra of its surface package and the actual content can be collected sequentially at the site where the content of the special item is visible without the interference of other objects, and then the raman spectrum information base is built up in order to form the standard database.
According to one embodiment of the invention, the standard database holds the overall information of the outbound special article. On the other hand, when naming standard articles in the standard database, marking information such as the name, source company, specification size, picture and the like of the standard articles; on the other hand, each specific article is provided with two corresponding spectrograms, one spectrogram is used for collecting the spectrogram information of the surface components on the surface of the specific site, the other spectrogram is used for collecting the spectrogram information of the internal actual components of the same site, and the spectrogram information of the internal actual components is collected. When the to-be-detected object is scanned and compared, the spectrogram information acquired by the surface is required to be compared with the surface standard spectrum in the spectrogram library, the acquired internal spectrogram information is compared with the internal standard spectrum in the spectrogram library, and the to-be-detected object can be considered to belong to the same object as the standard object after exceeding the threshold value twice, so that the matching accuracy is improved.
Fig. 1 shows a flowchart of a detection method for inspection of an outbound special article according to an embodiment of the present invention, which is described in detail below with reference to fig. 1, and the detection method for inspection of an inbound special article according to an embodiment of the present invention mainly includes the following steps:
and a Raman spectrum acquisition step: selecting an object to be detected on the detection equipment, and scanning the object to be detected to acquire a Raman spectrum of the object to be detected;
and (3) comparing and judging the Raman spectrum: comparing the acquired Raman spectrum of the object to be detected with the standard Raman spectrum of the object marked in the goods list stored in the standard database by adopting the identification algorithm according to the embodiment of the invention so as to judge whether the acquired Raman spectrum and the standard Raman spectrum are matched; the recognition algorithm of the embodiment of the present invention will be described in detail below;
in the raman spectrum comparing and judging step, if the object to be detected is judged to be matched with the object marked in the cargo list, the data can be saved, and the examination is ended; if it is determined that the inspected article does not match the marked article in the manifest, then the following steps need to be continued;
chemical substance identification and judgment steps: detecting actual components of the object to be detected by adopting a conventional Raman spectrum identification method to identify chemical substances contained in the object to be detected, and then comparing the identified chemical substances with chemical substances of the objects marked in the goods list stored in the standard database to judge whether the two are matched;
in the step of identifying and judging the chemical substances, if the information of the object to be detected is consistent with the information of the object marked in the goods list, the data can be saved and the examination is finished under the condition that the chemical substances of the object to be detected can be identified; if the detected article is inconsistent with the information of the marked article in the cargo list, the subsequent processing, such as manual unpacking inspection and the like, is needed; in the case that the chemical substance of the object to be inspected cannot be identified, the object to be inspected needs to be inspected again and then the collected data is saved for subsequent processing.
In the step of comparing and determining raman spectra in the method, the identification algorithm according to the embodiment of the present invention needs to be adopted, and the collected raman spectra of the object to be detected and the standard raman spectra of the objects marked in the cargo list stored in the standard database are compared to determine whether the raman spectra match, and the identification algorithm according to the embodiment of the present invention will be described in detail below with reference to the accompanying drawings.
According to one embodiment of the invention, a Raman spectrum technology is combined with a similarity measurement algorithm and a support vector machine improvement algorithm, and firstly, similarity measurement is carried out on an object to be detected and a standard object by calculating a correlation coefficient; secondly, when the correlation coefficient of the object to be detected and the standard object is lower than a threshold value, an improved support vector machine algorithm is provided pertinently, the high-dimensional vector is subjected to dimension reduction processing, and the object to be detected with the correlation coefficient lower than the threshold value is classified, so that the influence of noise generated by Raman spectrum equipment on the identification result is avoided, and the identification speed and accuracy are improved. Before describing a particular method, an algorithmic model used in the method is described.
[ Algorithm model ]
(1) Similarity measurement
The raman spectrum recognition technology is an application technology for classifying and recognizing objects to be detected, according to one embodiment of the invention, after spectrum pretreatment and feature extraction are completed, key information capable of reflecting the composition of substances is obtained, spectrum information contained in a spectrum signal is extracted, and the objects to be detected are classified and recognized according to spectrum information difference, wherein the spectrum information difference is calculated by adopting a similarity measurement method.
It should be noted that, the spectrum preprocessing generally includes denoising, baseline correction, normalization processing, and the like, and in this document, the original spectrum acquired needs to be preprocessed, which is not described in detail below for brevity.
The correlation coefficient (Correlation coefficient) is a measure of the degree of linear correlation between the study variables and is a measure of the correlation between vectors. Is provided with a feature vector X (X) 1 ，x 2 ，...，x n )，Y(y 1 ，y 2 ，...，y n ) The correlation coefficient of the two is defined as follows:
where n represents the length of the feature vector X, Y, respectively representing the mean of the vector X, Y, i representing the i-th data of the vector.
In this embodiment, the correlation coefficient is selected as the criterion of similarity measurement, so that the loss of the euclidean distance to the information and the amplification of the mahalanobis distance to the small deviation can be avoided, and the similarity between the two feature vectors can be better determined.
(2) Support Vector Machine (SVM)
In raman spectrum measurement, the raman spectrum is biased due to sample uniformity differences, instrument noise, fluorescent background and the like; errors can also occur during spectral processing, denoising, baseline correction, etc. In the identification process, the accuracy rate of the feature identification of the substances by only adopting the correlation coefficient is not high, so that in the feature identification process of the Raman spectrum of the specific articles entering and exiting in the inspection and quarantine field, a support vector machine is introduced to classify the articles to be inspected which are equal to or lower than or slightly lower than a threshold value.
The support vector machine is a bi-classification model whose basic model is defined as the most spatially-separated linear classifier. The principle of which is shown in figure 2.
Let training sample set X be X i I=1, 2,..n, samples fall into two categories, w 1 And w 2 And is linearly separable. The general form of the linear discriminant function l is: g (x) =ω·x+b, but such hyperplane is not unique. Thus, the classification recognition problem of the support vector machine translates into a classification hyperplane problem that finds the largest interval.
Selecting the appropriate ω, b such that the sample class w 1 And w 2 Satisfy g (x) =1 and g (x) =respectively-1, i.e. of the formula:
the geometrical meaning of the support vector machine in the linear classifier is in the hyperplane l 1 And l 2 The distance Gap between the two electrodes takes the maximum value. In the process, for the distance hyperplane l 0 The most recent samples are subjected to a normalization process, such Gap/2 = 2/|||omega|, the optimization function of the support vector machine with linear sortable support is as follows:
s.t.y i (ω·x i +b)≥1，i＝1，2，...，n
the following formula is available according to Lagrange's multiplier method and Karush-Kuhn-Tucker condition (KKT condition), where λ is Lagrange's multiplier.
λ i ≥0，i＝1，2，...，N
λ i [y i (ω·x+b)-1]＝1，i＝1，2，...，N
(3) Improved support vector machine
Because the Raman spectrum is preprocessed to obtain the high-dimensional vector for representing the substance to be identified, the training and learning time process is longer in the process of directly adopting the high-dimensional vector for identification and classification.
In order to change the problem, in the process of carrying out identification by adopting a support vector machine, a certain sparse transformation is firstly carried out on a Raman spectrum, and the specific operation comprises the following two steps: 1. acquiring a main component M of Raman spectrum information y of a training sample; 2. sparse representation is carried out on Raman spectrum information y of a training sample by using a formula y=Mc, and a series of sparse representation vectors c are obtained i The number of non-zero elements in the vector c is fixed to be k, and k is less than or equal to N; and a target vector c for sparsely representing the raman spectrum information y is calculated using the following target function:
where N is the length of vector c and i is the number of vector c.
By using the objective function, find the resultAnd taking a minimum target vector c to represent the spectrum information y by using the target vector c after the dimension reduction, thereby realizing sparse representation of the Raman spectrum.
In the process of classifying substances by selecting unmodified support vectors, because of high dimensionality of spectrum information and non-zero property of most information, RBF kernel functions, polynomial kernel functions and the like are required to be mapped to higher dimensions for classification processing; the improved support vector machine realizes sparse representation of the spectrum, enhances the distinguishing property of spectrum information, and can be used for spectrum classification by adopting a linear kernel support vector machine, so that the training and testing speed is higher, the required storage space is less, and the time is reduced in the training and learning process.
The classifier building process mainly comprises three steps, taking two classifications as examples, and a specific building flow is shown in fig. 3.
And selecting a proper amount of samples as training samples, measuring the samples to obtain Raman spectra, and preprocessing the Raman spectra through baseline correction, denoising, normalization and the like to obtain test sample data. The method comprises the steps of dividing a sample to be measured into a positive sample and a negative sample, wherein the positive sample is the spectrum information of a certain article to be measured, and the negative sample is the spectrum information of a non-article to be measured. And carrying out sparse representation on the spectrum, and carrying out model establishment by adopting a support vector machine to obtain the SVM classifier.
[ method ]
Based on the above-described calculation model, a raman spectrum detection method for inspection of an article according to an embodiment of the present invention is further described below in conjunction with fig. 4, and may include the steps of:
and a Raman spectrum acquisition step: collecting Raman spectrum of the object to be detected; and
and (3) comparing and judging: comparing the acquired Raman spectrum of the object to be detected with the Raman spectrum of the standard object stored in the standard database to judge whether the object to be detected is matched with the standard object.
According to one embodiment, the comparing and determining steps specifically include: calculating a correlation coefficient of a characteristic vector of the Raman spectrum of the object to be detected and a characteristic vector of the Raman spectrum of the standard object stored in the standard database, and judging that the object to be detected is matched with the standard object when the calculated correlation coefficient is larger than a preset threshold value.
Further, the comparing and determining step further includes: and when the calculated correlation coefficient is equal to or lower than a preset threshold value, classifying the Raman spectrum of the object to be detected by adopting a support vector machine so as to judge whether the object to be detected is matched with the standard object.
Specifically, the step of classifying raman spectra of the object to be detected by using a support vector machine includes:
a step of establishing a classifier: selecting a training sample, measuring the training sample to obtain a Raman spectrum of the training sample, performing sparse transformation on the Raman spectrum of the training sample, and then establishing a classifier by adopting a support vector machine algorithm; and
classification: and classifying the Raman spectrum of the object to be detected by adopting the established classifier.
In the sparse transformation process of the raman spectrum of the training sample, the objective function defined by the formula (6) can be adopted to perform sparse change on the raman spectrum of the training sample so as to obtain sparse representation of the raman spectrum, so that training learning time is shortened, and training and testing speeds are higher.
[ experiment verification ]
Next, experimental verification is performed on a raman spectrum detection method for inspection of an article according to an embodiment of the present invention.
(1) Experimental conditions and methods
The experiment randomly extracts 380 inbound special articles in the daily checking process of the Beijing city inbound inspection and quarantine bureau to carry out Raman spectrogram acquisition and support vector machine algorithm verification experiment. Table 1 shows a category statistics table.
TABLE 1 statistical table of 380 types of entry and exit special article species
The instrument used in the experiment is a RT6000 handheld Raman spectrometer of the same Fang Wei visual technology Co., ltd, and the excitation wavelength is 785nm; resolution is 6-9 cm -1 The method comprises the steps of carrying out a first treatment on the surface of the Wave number range of 200-3200 cm -1 。
The Raman spectrograms acquired by the experiment are subjected to noise reduction, baseline correction and normalization pretreatment operation, and all spectrogram analysis and comparison verification are obtained through Matlab (R2010 a) and C++ programming methods.
The accuracy verification of the improved support vector machine comprises two aspects, namely, a substance with the same identification result as a true value is judged to be correct (pass), and a substance with a different identification result from the true value is judged to be wrong (fail). And verifying the accuracy of the model by using the sample to be tested, and judging the accuracy of the algorithm model. The specific flow is shown in fig. 4.
In the experiment, preprocessing such as baseline correction, denoising, normalization and the like is carried out on the acquired Raman spectrum, similarity comparison is carried out on the acquired Raman spectrum and a sample in a standard spectrum library, and if the correlation coefficient is greater than a threshold value, the acquired Raman spectrum is judged to be matched; and classifying the Raman spectra which are smaller than or equal to the threshold value by adopting an improved support vector machine, and if the classification result shows that the Raman spectra are matched, namely, the Raman spectra are matched, otherwise, the Raman spectra are not matched.
(2) Experimental results
All test samples were analyzed using a modified support vector machine, wherein 6 classes of materials, including blood products, viruses, antibodies, and the like, were combined. The class of substance standard databases has a range of discrimination between 0.7460-0.7822 and an average discrimination between classes of 0.7623. The matching rate of the substance compliance check using the similarity measure, the support vector machine and the modified support vector machine algorithm was counted as shown in table 2.
Table 2 list of test results
In the compliance verification process, the correct matching rates using the three methods are shown in table 2. In the inspection process of sequentially selecting the thresholds of 0.85, 0.86, 0.87, 0.88, 0.89 and 0.90, comparison can prove that: when a lower threshold is selected, a better checking result can be obtained by adopting a correlation coefficient, and the matching rate gradually decreases along with the further increase of the threshold; in the process of checking by adopting a support vector machine and an improved support vector machine algorithm, the matching rate of the object to be checked is higher than that of the object to be checked by directly adopting the similarity measurement. Therefore, the support vector machine classification judgment process is added to the to-be-inspected object with low similarity, the possibility of missed inspection can be reduced, and meanwhile, the accuracy of the classifier is improved due to the improvement of the sparse expression of the support vector machine algorithm, so that the improved support vector machine algorithm obtains an optimal result in the inspection process.
According to another embodiment of the present invention, there is further provided a raman spectrum detection method for inspection of articles, which may directly use an improved support vector machine algorithm to perform the above-described comparing and determining steps without performing a similarity measure.
Specifically, the method may comprise the steps of:
and a Raman spectrum acquisition step: collecting Raman spectrum of the object to be detected; and
and (3) comparing and judging: and classifying the Raman spectrum of the object to be detected by directly adopting a support vector machine so as to judge whether the object to be detected is matched with the standard object.
Specifically, the step of classifying raman spectra of the object to be detected by using a support vector machine includes:
a step of establishing a classifier: selecting a training sample, measuring the training sample to obtain a Raman spectrum of the training sample, performing sparse transformation on the Raman spectrum of the training sample, and then establishing a classifier by adopting a support vector machine algorithm; and
classification: and classifying the Raman spectrum of the object to be detected by adopting the established classifier.
Similarly, in the sparse transformation process of the raman spectrum of the training sample, the objective function defined by the above formula (6) may be used to perform sparse transformation on the raman spectrum of the training sample, so as to obtain sparse representation of the raman spectrum, so as to reduce training learning time.
That is, in this embodiment, the raman spectra of all the articles to be measured are classified and identified using the modified support vector machine algorithm without distinguishing whether the correlation coefficient thereof is greater than or equal to the threshold value. It should be understood that this embodiment may include other steps similar to the above-described embodiment, such as including various pretreatment steps, etc., in addition to this distinction.
According to yet another embodiment of the invention, there is also provided an electronic device, fig. 5 being a block diagram illustrating an example hardware arrangement 500 of the electronic device. The hardware arrangement 500 includes a processor 506 (e.g., a microprocessor (μp), a Digital Signal Processor (DSP), etc.). Processor 506 may be a single processing unit or multiple processing units for performing the different actions of the method steps described herein. The arrangement 500 may further comprise an input unit 502 for receiving signals from other entities, and an output unit 504 for providing signals to other entities. The input unit 502 and the output unit 504 may be arranged as a single entity or as separate entities.
Further, the arrangement 500 may include at least one readable storage medium 508 in the form of non-volatile or volatile memory, such as electrically erasable programmable read-only memory (EEPROM), flash memory, and/or a hard disk drive. The readable storage medium 508 comprises a computer program 510, the computer program 510 comprising code/computer readable instructions which, when executed by the processor 506 in the arrangement 500, enable the hardware arrangement 500 and/or a device comprising the hardware arrangement 500 to perform a procedure such as described above in connection with fig. 1, 3-4 and any variations thereof.
The computer program 510 may be configured as computer program code having an architecture of, for example, computer program modules 510A-510C. Thus, in an example embodiment when the hardware arrangement 500 is used in a device, for example, code in a computer program of the arrangement 500 comprises: module 510A for …. The code in the computer program further comprises: module 510B for …. The code in the computer program further comprises: module 510C for ….
The computer program modules may substantially execute the various actions in the flows shown in figures 1, 3-4 to simulate a device. In other words, when different computer program modules are executed in the processor 506, they may correspond to the different units described above in the device.
Although the code means in the embodiment disclosed above in connection with fig. 5 are implemented as computer program modules which, when executed in the processor 506, cause the hardware arrangement 500 to perform the actions described above in connection with fig. 1-4, in alternative embodiments at least one of the code means may be implemented at least partly as hardware circuitry.
The processor may be a single CPU (central processing unit), but may also comprise two or more processing units. For example, a processor may include a general purpose microprocessor, an instruction set processor, and/or an associated chipset and/or special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)). The processor may also include on-board memory for caching purposes. The computer program may be carried by a computer program product connected to the processor. The computer program product may include a computer readable medium having a computer program stored thereon. For example, the computer program product may be a flash memory, a Random Access Memory (RAM), a Read Only Memory (ROM), an EEPROM, and the above-described computer program modules may be distributed in alternative embodiments in the form of memory within the UE into different computer program products.
It should be understood by those skilled in the art that, in the embodiment of the present invention, the technical concept of the present invention is described in detail by taking inspection of an entry special article as an example, but the present invention is not limited to inspection of an entry special article, and may be applied to fields such as rapid detection of medicines, screening of biological samples, and the like.
Although the present invention has been described with reference to the accompanying drawings, the examples disclosed in the drawings are intended to illustrate preferred embodiments of the invention and are not to be construed as limiting the invention.
Although a few embodiments of the present general inventive concept have been shown and described, it would be appreciated by those skilled in the art that changes may be made in these embodiments without departing from the principles and spirit of the general inventive concept, the scope of which is defined in the claims and their equivalents.
Claims (6)
1. A raman spectrum detection method for inspection of an outbound special article, comprising the steps of:
and a Raman spectrum acquisition step: scanning the surface and the interior of a specific site of an object to be detected; and
and (3) comparing and judging: comparing the collected Raman spectrum of the object to be detected with the Raman spectrum of the standard object stored in the standard database to judge whether the object to be detected is matched with the standard object,
the comparing and judging steps comprise:
respectively carrying out similarity measurement on the Raman spectrum of the surface component of the object to be detected and the Raman spectrum of the surface component of the standard object stored in the standard database, and the Raman spectrum of the component of the actual content of the object to be detected and the Raman spectrum of the component of the actual content of the standard object stored in the standard database;
when the result of the similarity measurement is larger than a preset threshold value, judging that the object to be detected is matched with the standard object;
when the result of the similarity measurement is equal to or lower than a preset threshold value, classifying the Raman spectrum of the object to be detected by adopting a support vector machine so as to realize the comparison of the Raman spectrum of the surface package and the actual content of the object to be detected and the Raman spectrum of the surface package and the actual content of the standard object stored in the standard database;
the step of classifying the Raman spectrum of the object to be detected by adopting the support vector machine comprises the following steps:
a step of establishing a classifier: selecting a training sample, measuring the training sample to obtain a Raman spectrum of the training sample, performing sparse transformation on the Raman spectrum of the training sample, and then establishing a classifier by adopting a support vector machine algorithm; and
classification: classifying Raman spectra of the object to be detected by adopting the established classifier;
the sparse transform comprises the steps of:
acquiring a main component M of Raman spectrum information y of a training sample;
sparse representation is carried out on Raman spectrum information y of a training sample by using a formula y=Mc, and a series of sparse representation vectors c are obtained i The method comprises the steps of carrying out a first treatment on the surface of the And
the target vector c sparsely representing the raman spectrum information y is calculated using the following target function:
wherein the number of non-zero elements in the vector c is fixed to be k, N is the length of the vector c, k is less than or equal to N, and i is the sequence number of the vector c.
2. The raman spectrum detection method according to claim 1, wherein the performing similarity measurement between the raman spectrum of the object to be detected and the raman spectrum of the standard object stored in the standard database comprises:
and calculating the correlation coefficient of the characteristic vector of the Raman spectrum of the object to be detected and the characteristic vector of the Raman spectrum of the standard object stored in the standard database, and taking the calculated correlation coefficient as a result of the similarity measurement.
3. The raman spectrum detection method according to claim 1, characterized by further comprising the steps of:
establishing a standard database: the raman spectra of the surface package and the actual content of the standard article are collected sequentially to form a standard database comprising raman spectra of the surface components of the standard article and raman spectra of the components of the actual content.
4. A raman spectrum detection method according to claim 3 wherein when the result of the similarity measure is greater than a preset threshold value, determining that the object to be detected matches the standard object comprises:
the result of the similarity measure being greater than a preset threshold value only includes a case where the results of both similarity measures are greater than a preset threshold value.
5. The raman spectrum detection method according to claim 1, wherein said step of establishing a standard database further comprises: and collecting the name, source company, specification size and picture information of the standard object, and storing the information into a standard database.
6. An electronic device, comprising:
a memory for storing executable instructions; and
a processor for executing executable instructions stored in a memory to perform the method of any one of claims 1-5.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611257547.5A CN108254351B (en) | 2016-12-29 | 2016-12-29 | Raman spectrum detection method for checking articles |
| PCT/CN2017/111624 WO2018121122A1 (en) | 2016-12-29 | 2017-11-17 | Raman spectroscopy detection method for checking goods, and electronic device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611257547.5A CN108254351B (en) | 2016-12-29 | 2016-12-29 | Raman spectrum detection method for checking articles |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108254351A CN108254351A (en) | 2018-07-06 |
| CN108254351B true CN108254351B (en) | 2023-08-01 |
Family
ID=62706829
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611257547.5A Active CN108254351B (en) | 2016-12-29 | 2016-12-29 | Raman spectrum detection method for checking articles |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN108254351B (en) |
| WO (1) | WO2018121122A1 (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11428638B2 (en) * | 2018-01-12 | 2022-08-30 | The Regents Of The University Of California | Spectroscopic biological material characterization |
| CN110489586A (en) * | 2019-01-07 | 2019-11-22 | 公安部第一研究所 | A kind of matched method of sample detection database hierarchy |
| CN113811880A (en) * | 2019-06-07 | 2021-12-17 | 巴斯夫涂料有限公司 | Device and method for forming at least one ground truth database for an object recognition system |
| CN111141719A (en) * | 2019-12-27 | 2020-05-12 | 梧州市食品药品检验所 | Rapid nondestructive identification method for anti-AIDS medicine |
| CN111426648B (en) * | 2020-03-19 | 2023-04-07 | 甘肃省交通规划勘察设计院股份有限公司 | Method and system for determining similarity of infrared spectrogram |
| CN111861171B (en) * | 2020-07-08 | 2021-07-27 | 京源中科科技股份有限公司 | Special article monitoring system, method and device based on Internet of things |
| CN113324967A (en) * | 2020-09-18 | 2021-08-31 | 山东农业大学 | Method for rapidly identifying DFD beef |
| CN112161965B (en) * | 2020-09-22 | 2022-08-05 | 复旦大学附属妇产科医院 | Method, device, computer equipment and storage medium for detecting traditional Chinese medicine property |
| CN112232673A (en) * | 2020-10-16 | 2021-01-15 | 拜耳医药保健有限公司 | Early warning method, device, equipment and medium for processing genetic resources |
| CN112686768A (en) * | 2020-12-30 | 2021-04-20 | 山东省食品药品检验研究院 | Quick identification system of medicine preparation enterprise material |
| CN112834481B (en) * | 2020-12-31 | 2023-09-05 | 宁波海关技术中心 | Raman spectrum enhancement measurement system and measurement method |
| CN112834451B (en) * | 2021-01-12 | 2023-04-18 | 深圳网联光仪科技有限公司 | Sample identification method and device based on infrared spectrum and storage medium |
| CN113092407B (en) * | 2021-04-20 | 2024-01-30 | 深圳华创智信科技有限公司 | Near infrared spectrum-based method for identifying varieties of original places of five-normal rice flower fragrance rice |
| CN113466206A (en) * | 2021-06-23 | 2021-10-01 | 上海仪电（集团）有限公司中央研究院 | Raman spectrum analysis system based on big data |
| CN113656559B (en) * | 2021-10-18 | 2022-01-25 | 印象(山东)大数据有限公司 | Data analysis method and device based on metering platform and electronic equipment |
| CN114970596B (en) * | 2022-01-19 | 2024-08-23 | 福建省产品质量检验研究院(福建省缺陷产品召回技术中心) | Raman spectrum analysis-based fish meal adulteration rapid detection method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103903010A (en) * | 2014-03-28 | 2014-07-02 | 哈尔滨工程大学 | Method for classifying hyperspectral images based on sparse characteristics and same neighborhood properties |
| JP2015052551A (en) * | 2013-09-09 | 2015-03-19 | 株式会社かんでんエンジニアリング | Component-in-oil measuring method and oil-filled electric instrument |
| CN104764734A (en) * | 2015-04-30 | 2015-07-08 | 江苏易谱恒科技有限公司 | Identification method for fruit juice flavors and fragrances based on Raman spectrum and SVM algorithm |
| CN104865279A (en) * | 2014-02-26 | 2015-08-26 | 清华大学 | Customs supervision item rapid-inspection equipment and method |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1967846A1 (en) * | 2007-03-05 | 2008-09-10 | National University of Ireland Galway | En ensemble method and apparatus for classifying materials and quantifying the composition of mixtures |
| CN101403696A (en) * | 2008-10-21 | 2009-04-08 | 浙江大学 | Method for measuring gasoline olefin content based on Raman spectrum |
| CN103217411B (en) * | 2013-04-15 | 2015-03-25 | 浙江大学 | Raman-spectrum-based method for detecting content of oleic acid, linoleic acid and saturated fatty acid in edible vegetable oil |
| CN104215623B (en) * | 2013-05-31 | 2018-09-25 | 欧普图斯(苏州)光学纳米科技有限公司 | Laser Raman spectroscopy intelligence discrimination method and system towards conglomerate detection |
| CN105806824B (en) * | 2016-03-22 | 2019-03-19 | 武汉轻工大学 | Edible oil type method for quick identification based on multi-source optical spectrum Fusion Features |
-
2016
- 2016-12-29 CN CN201611257547.5A patent/CN108254351B/en active Active
-
2017
- 2017-11-17 WO PCT/CN2017/111624 patent/WO2018121122A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015052551A (en) * | 2013-09-09 | 2015-03-19 | 株式会社かんでんエンジニアリング | Component-in-oil measuring method and oil-filled electric instrument |
| CN104865279A (en) * | 2014-02-26 | 2015-08-26 | 清华大学 | Customs supervision item rapid-inspection equipment and method |
| CN103903010A (en) * | 2014-03-28 | 2014-07-02 | 哈尔滨工程大学 | Method for classifying hyperspectral images based on sparse characteristics and same neighborhood properties |
| CN104764734A (en) * | 2015-04-30 | 2015-07-08 | 江苏易谱恒科技有限公司 | Identification method for fruit juice flavors and fragrances based on Raman spectrum and SVM algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 结构化压缩感知研究进展;刘芳;《自动化学报》;20131231;第1980-1995页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2018121122A1 (en) | 2018-07-05 |
| CN108254351A (en) | 2018-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108254351B (en) | Raman spectrum detection method for checking articles | |
| CN108362662B (en) | Near infrared spectrum similarity calculation method and device and substance qualitative analysis system | |
| Muro et al. | Forensic body fluid identification and differentiation by Raman spectroscopy | |
| US7679740B2 (en) | Method and apparatus for multimodal detection | |
| McLaughlin et al. | Discrimination of human and animal blood traces via Raman spectroscopy | |
| US11493447B2 (en) | Method for removing background from spectrogram, method of identifying substances through Raman spectrogram, and electronic apparatus | |
| US7990532B2 (en) | Method and apparatus for multimodal detection | |
| EP3088871B1 (en) | Raman spectrum detection method | |
| CN108240978B (en) | Self-learning qualitative analysis method based on Raman spectrum | |
| CN107818298A (en) | General Raman spectral characteristics extracting method for machine learning material recognition | |
| WO2015096788A1 (en) | Raman spectroscopy method used for detecting sample in accommodating body | |
| WO2017019988A1 (en) | Gender and race identification from body fluid traces using spectroscopic analysis | |
| US20120134582A1 (en) | System and Method for Multimodal Detection of Unknown Substances Including Explosives | |
| CN108398416A (en) | A kind of mix ingredients assay method based on laser Raman spectroscopy | |
| CN104376325A (en) | Method for building near-infrared qualitative analysis model | |
| CN104777143A (en) | Method for similarity identification of expired drugs based on Raman spectroscopy | |
| CN108169201B (en) | Raman spectrum detection method for deducting package interference | |
| CN112712108A (en) | Raman spectrum multivariate data analysis method | |
| WO2018103541A1 (en) | Raman spectrum detection method and electronic apparatus for removing solvent perturbation | |
| Rajesh et al. | Studies on identification of medicinal plant variety based on NIR spectroscopy using plant leaves | |
| Burlacu et al. | Convolutional Neural Network detecting synthetic cannabinoids | |
| Zhao et al. | Performance Improvement of Handheld Raman Spectrometer for Mixture Components Identification Using Fuzzy Membership and Sparse Non-Negative Least Squares | |
| WO2018121151A1 (en) | Method for identifying raman spectrogram, and electronic device | |
| Suthagar et al. | Determination of apple, lemon, and banana ripening stages using electronic nose and image processing | |
| WO2006019991A2 (en) | Method and apparatus for multimodal detection |
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 |