CN114878506A - Ivory identification method based on near infrared spectrum - Google Patents
Ivory identification method based on near infrared spectrum Download PDFInfo
- Publication number
- CN114878506A CN114878506A CN202210408852.9A CN202210408852A CN114878506A CN 114878506 A CN114878506 A CN 114878506A CN 202210408852 A CN202210408852 A CN 202210408852A CN 114878506 A CN114878506 A CN 114878506A
- Authority
- CN
- China
- Prior art keywords
- ivory
- spectrum
- sample
- near infrared
- identification
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000002329 infrared spectrum Methods 0.000 title claims abstract description 38
- 238000001228 spectrum Methods 0.000 claims abstract description 101
- 238000012795 verification Methods 0.000 claims abstract description 19
- 238000012937 correction Methods 0.000 claims abstract description 15
- 238000004458 analytical method Methods 0.000 claims abstract description 14
- 241001503485 Mammuthus Species 0.000 claims description 24
- 230000002159 abnormal effect Effects 0.000 claims description 17
- 230000003595 spectral effect Effects 0.000 claims description 17
- 238000010200 validation analysis Methods 0.000 claims description 7
- 241000283095 Elephas maximus Species 0.000 claims description 5
- 241000283092 Loxodonta Species 0.000 claims description 5
- 238000002835 absorbance Methods 0.000 claims description 5
- 238000004497 NIR spectroscopy Methods 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 4
- 241000406668 Loxodonta cyclotis Species 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000004451 qualitative analysis Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000012217 deletion Methods 0.000 claims description 2
- 230000037430 deletion Effects 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 8
- 230000000877 morphologic effect Effects 0.000 abstract description 7
- 238000012216 screening Methods 0.000 abstract description 6
- 238000010170 biological method Methods 0.000 abstract description 2
- 238000013480 data collection Methods 0.000 abstract 1
- 230000001737 promoting effect Effects 0.000 abstract 1
- 210000000988 bone and bone Anatomy 0.000 description 7
- 238000002360 preparation method Methods 0.000 description 7
- 230000003247 decreasing effect Effects 0.000 description 6
- 239000000463 material Substances 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 241000282326 Felis catus Species 0.000 description 4
- 241000894007 species Species 0.000 description 4
- 241001465754 Metazoa Species 0.000 description 3
- 238000010521 absorption reaction Methods 0.000 description 3
- 239000003086 colorant Substances 0.000 description 3
- 238000009795 derivation Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 239000000779 smoke Substances 0.000 description 3
- 241000196324 Embryophyta Species 0.000 description 2
- 239000011324 bead Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008094 contradictory effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 235000013569 fruit product Nutrition 0.000 description 2
- 238000012845 near infrared spectroscopy analysis Methods 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 241000502561 Acacia irrorata Species 0.000 description 1
- 235000014653 Carica parviflora Nutrition 0.000 description 1
- 241000243321 Cnidaria Species 0.000 description 1
- 241000283080 Proboscidea <mammal> Species 0.000 description 1
- 241000512272 Tridacna Species 0.000 description 1
- 210000003056 antler Anatomy 0.000 description 1
- 238000000149 argon plasma sintering Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000002308 calcification Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013524 data verification Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010422 painting Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- G01N21/3563—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
- G01N21/25—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
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (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 Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides an ivory identification method based on near infrared spectrum, which comprises the steps of firstly collecting a sample, selecting a spectrum collection part of the sample, and carrying out near infrared spectrum scanning and spectrum data collection on the sample; selecting a correction set and verification spectrum data of modeling, determining a characteristic spectrum, selecting an optimal main factor value and an optimal F value, and establishing a near infrared spectrum identification model of ivory; and analyzing the near infrared spectrum data of the sample to be detected by using the model, and giving a sample identification result. The invention establishes the ivory authenticity identification model by utilizing the near infrared analysis method for the first time, and compared with the traditional morphological identification, the method has the characteristics of objectivity, digitalization and quantifiability; compared with a molecular biological method, the method disclosed by the invention does not need complex pretreatment, is a method which is high in detection speed, simple and convenient to operate and free of damage to samples, is suitable for being used as a quick examination and preliminary screening method for field law enforcement of a supervision department, and has important significance for promoting the protection and utilization of species resources.
Description
Technical Field
The invention relates to a spectral nondestructive identification method of ivory. A near infrared spectrum detection technology is used for detecting and identifying object teeth and imitations thereof, and belongs to the technical field of biology.
Background
Ivory products are widely accepted and appreciated by people because of their unique and high-value colors, smooth texture, natural and perfect combination with artificial chisels. However, in order to obtain ivory, a large number of elephants are killed, and the bloody true phase is hidden behind the exquisite and absolute dental sculpture handicraft. The species listed in International trade convention for endangered wild animal and plant species (CITES) is likely to have strictly limited trade.
The identification of the ivory is mainly based on the schlieren angle structure, and the existence of the schlieren angle structure and the size of the outer edge schlieren angle are the main characteristics for identifying whether the ivory is an ivory or a modern ivory or mammoth ivory. However, if the product is a craft product, the product has no outer edge schlieren structure, or some ivory has no characteristic lines, and the identification is only carried out from the morphological characteristics, so that the product cannot be judged. Although there are cases of successful identification of ivory from the viewpoint of molecular biology in the extraction of DNA from ivory, even modern ivory has a very high degree of calcification and petrifaction and an extremely low DNA content despite the fact that the ivory is buried underground, there is a high difficulty in extracting DNA. In addition, there is also a method for identifying physical characteristics of ivory, such as mohs hardness, density, refractive index, etc., but the effect is generally close to that of an analog, and it is difficult to be used as a basis for independent judgment. Near Infrared Spectroscopy (NIRS) is a nondestructive identification method, and detection and identification are carried out by utilizing different spectral characteristics of different substances in a Near Infrared spectrum region. The present invention utilizes near infrared spectrum technology to establish the method of identifying and distinguishing ivory and ivory imitation and some other natural matter.
Disclosure of Invention
The invention provides an ivory identification method based on near infrared spectrum, which comprises the following steps:
s1, collecting a sample;
s2, selecting a spectrum acquisition part of a sample and scanning a near infrared spectrum;
s3, establishing a near infrared spectrum identification model of the sample;
s4, judging a near infrared spectrum identification result: and analyzing the near infrared spectrum data of the sample to be detected by using a near infrared spectrum analysis model to give a sample identification result.
Further, the sample spectrum collecting part selected in step S2 should be selected to have a typical color part of the sample for scanning.
Further, the scanning parameters of the near-infrared spectrometer in the step S2 are set as: the spectrum scanning range is 1000 nm-1800 nm, the resolution is 11nm, and the scanning times are 30 times. Each sample was scanned 2 times, and the average of 2 scans was taken as the original spectrum.
Further, in the step S3, a calibration set for modeling and spectral data verification are selected, a characteristic spectrum segment is determined, an optimal main factor value and an optimal F value are selected, and a near infrared spectrum identification model of ivory is established.
Further, in the near infrared spectroscopic analysis model in the step S3, the F value is set to 0.57, and the main factor is set to 2.
Further, the judgment of the near infrared spectrum identification result in the step S4: one sample is measured with a plurality of spectra, and the ivory is judged as the ivory as long as the result of one spectrum shows the ivory.
Further, the establishment of the near infrared spectrum analysis model comprises the following steps:
(1) collecting a sample;
(2) collecting a near infrared spectrum: acquiring basic spectrum by using a spectrometer with a spectrum scanning range of 1000-1800 nm, a resolution of 11nm and a scanning frequency of 30 times; scanning in a near-infrared diffuse reflection mode, collecting a basic spectrum, scanning each sample for 2 times, and taking the obtained average value as an original spectrum of the sample; scanning each sample by the same method in sequence to obtain an average value; identifying and deleting the acquired abnormal spectrum sample;
(3) dividing a correction set and a verification set according to the properties of a known sample, wherein the correction base is an ivory spectrum, and the verification set is a non-ivory spectrum;
(4) spectrum pretreatment: preprocessing the calibration set and the validation collection spectra;
(5) extracting characteristic wave bands from the preprocessed diffuse reflection spectrum, and selecting a sensitive spectrum band for ivory identification;
(6) establishing and verifying ivory models: calculating by using the extracted characteristic wavelength through the Mahalanobis distance, establishing an identification model by adopting an SIMCA qualitative analysis method, and deducing to obtain an optimal main factor value and an optimal F value;
(7) and (3) identifying the accuracy: detecting the accuracy of the model by using the verification set and the samples of known classes;
(8) and (3) predicting unknown samples: and (6) carrying out pattern recognition on the unknown sample through the established ivory identification model.
Furthermore, the near infrared spectrum analysis model is established in the step (2) to identify and delete the acquired abnormal spectrum sample, including the direct deletion of the spectrum with obvious spectral characteristics and material differences of the same properties and the spectrum with absorbance exceeding 1.5.
Further, the ivory spectrum in the step (3) of establishing the near infrared spectroscopic analysis model includes spectra of an african elephant tooth, an asian elephant tooth and a mammoth elephant tooth.
Further, the near infrared spectrum analysis model establishing step (4) is used for preprocessing the correction set and the verified collection spectrum, and comprises standard normal variable transformation, detrending correction, Savitzky-Golay smoothing, Savitzky-Golay derivative and mean centralization.
Furthermore, the sensitive spectral bands of the ivory identification in the step (5) of establishing the near infrared spectrum analysis model are 1160-1200 nm, 1430-1500 nm, 1680-1710 nm and 1720-1750 nm respectively.
More specifically, the invention provides a near infrared spectrum analysis technology-based ivory identification method with accurate high-efficiency lossless prediction result, which comprises the following steps:
(1) collecting a sample: collecting at least 10 pieces of African elephant ivory, Asian elephant ivory, mammoth ivory original tooth or products, and collecting samples of imitation products including plastic ivory products, ivory fruit products and the like and natural materials such as ox horn, antler, tridacna, red coral and the like;
(2) collecting a near infrared spectrum: and (3) carrying out basic spectrum acquisition by using a spectrometer with a spectrum scanning range of 1000-1800 nm, a resolution of 11nm and a scanning frequency of 30 times. Scanning and collecting spectra in a near-infrared diffuse reflection mode, scanning each sample for 2 times, and taking the obtained average value as the original spectrum of the sample; scanning each sample in sequence in the same near infrared diffuse reflection mode to obtain an average value; identifying and deleting the acquired abnormal spectrum sample, such as directly deleting the spectrum with obvious spectral characteristics and material differences with the same properties, the spectrum with absorbance exceeding 1.5 and the like;
(3) based on the properties of the known samples, the calibration set (ivory spectra, including spectra of african elephant ivory, asian elephant ivory and mammoth ivory) and the validation set (non-ivory spectra) were partitioned;
(4) spectrum pretreatment: the calibration set and the validated collection spectra are preprocessed, including standard normal transformation (SNV), Detrending (DT), Savitzky-Golay smoothing, Savitzky-Golay derivative and mean centering, among others. Noise and background absorption is achieved through pretreatment, and interference information such as light scattering is reduced and eliminated;
(5) extracting characteristic wave bands from the preprocessed diffuse reflection spectrum, wherein 4 sensitive spectrum bands for ivory identification are 1160-1200 nm, 1430-1500 nm, 1680-1710 nm and 1720-1750 nm respectively;
(6) establishing and verifying ivory models: calculating by applying the screened characteristic wavelength through the Mahalanobis distance, establishing an identification model by adopting a SIMCA qualitative analysis method, and deducing to obtain an optimal main factor value and an optimal F value;
(7) and (3) identifying the accuracy: detecting the correctness of the model by using the verification set and a sample of a known class (a sample known as ivory or not);
(8) and (3) predicting unknown samples: and (6) carrying out pattern recognition on the unknown sample through the established ivory identification model.
The beneficial effects of the invention include: the method establishes the ivory authenticity identification model by utilizing the near infrared analysis method for the first time, and compared with the traditional morphological identification, the method has the characteristics of objectivity, digitalization and quantifiability; compared with the prior art such as a molecular biological method and the like, the method disclosed by the invention does not need complex sample pretreatment, is high in detection speed, simple and convenient to operate, is lossless to the sample, and is suitable for being used as a rapid examination primary screening method for field law enforcement of supervision departments, such as the rapid primary screening of customs in international airports and international mail examination fields.
Drawings
Fig. 1A to 1H are spectrograms and model statistical charts. FIGS. 1A to 1B: raw (fig. 1A) and derivative (fig. 1B) spectra of 2 beige ivory and 2 dark gray black ivory; fig. 1C to 1D: original (fig. 1C) and derivative (fig. 1D) spectrograms of 2 blue-black mammoth ivory integuments and 2 beige mammoth ivory teeth; FIGS. 1E to 1F: original (fig. 1E) and derivative (fig. 1F) spectra of ivory products of different thicknesses; FIGS. 1G to 1H: abnormal spectra found when correcting set spectra (including modern ivory and mammoth ivory) were qualitatively analyzed during SIMCA modeling.
Figure 2 is a photograph of a typical sample of ivory and a scan pattern. 2A: a typical beige ivory picture shaft head; 2B: a typical gray black ivory painting shaft head; 2C: a typical blue-black mammoth ivory cortex; 2D: a typical beige mammoth; 2E: an ivory blade; 2F: scanning the thin sheet sample on a table; 2G: the sheet-type sample was scanned while being held in hand.
Fig. 3A to 3F are ivory near infrared spectrum identification modeling analysis diagrams. FIG. 3A: the corrected sample F value was 2.5;
FIG. 3B: verifying that the sample F value is 2.5; FIG. 3C: the corrected sample F value was 1.14; FIG. 3D: verifying that the sample F value is 1.14; FIG. 3E: the corrected sample F value was 0.57; FIG. 3F: the validation sample F value was 0.57.
Detailed description of the preferred embodiments
1. Influence of sample properties on test results:
1.1 sample
153 ivory preparations were collected, including 99 modern ivory but indistinguishable asian or african ivory preparations, 26 african ivory preparations, 18 asian ivory preparations, and 10 mammoth ivory preparations. Most ivory is in the typical ivory white or beige color, and the surface of a sample is dark, such as black gray and the like, due to 5 ivory products which are long-term or improper to store; 8 mammoth ivory products with deep blue tooth skins. In addition, most ivory products are scroll heads, bracelets, stamps, beads, safety buttons and the like, have certain thickness, and some ivory products are in a thin sheet shape, such as 3 ivory fans, 2 ivory blades, 5 ivory piano key thin sheets and the like. Because different parts of the same sample have different colors and shapes, a plurality of spectral information are often acquired simultaneously for different parts of the same sample, and therefore the number of actually obtained spectral samples is larger than that of samples.
1.2 Spectrum scanning method
And (3) carrying out basic spectrum acquisition by using a spectrometer with a spectrum scanning range of 1000-1800 nm, a resolution of 11nm and a scanning frequency of 30 times. Scanning and collecting spectra in a near-infrared diffuse reflection mode, scanning each sample for 2 times, and calculating an average value to serve as an original spectrum of the sample; each sample was scanned in turn using the same method and the average was obtained.
1.3 influence of sample color and thickness on test results
1.3.1 spectral comparison
FIG. 1A is an original spectrum of 2 beige ivory stubs (typical figure is 2A in FIG. 2) and 2 dark gray black ivory stubs (typical figure is 2B in FIG. 2), the 4 spectra are not much different, but the absorbance of the dark sample is significantly higher than the white sample, as judged from the waveform. Fig. 1B is a derivative spectrum of the original spectrum of fig. 1A after 1-fitting 1 derivation with a window number of 3 (the number of data points involved in derivation is 3), and the 4 spectra in fig. 1B do not differ much. It is illustrated that the type 2B grayish black in fig. 2 has little influence on the spectrum. Fig. 1C is an original spectrogram of 2 mammoth ivory skins (typically, 2C in fig. 2) and 2 mammoth ivory skins (typically, 2D in fig. 2), in which the difference between the absorption peaks of the patterned mammoth ivory skins and the patterned mammoth ivory skins is a certain difference between the absorption peaks of the patterned mammoth ivory skins and the patterned mammoth ivory skins in the 1000 nm-1400 nm region, and the difference is still present in the patterns after the spectral processing (fig. 1D).
FIG. 1E is the original spectrum of 0.69mm thick fan blade of ivory, 2.86mm ivory blade (2E in FIG. 2) and 21.6mm long and 17.4mm short ivory beads, and FIG. 1F is the derivative spectrum (derivation conditions are as above). Fig. 1E and 1F both show that the thickness of different samples has little effect on the spectrum pattern, and the scanning mode, i.e. the sample is placed on the table for scanning (2F in fig. 2) or the sample is held in the hand for scanning (2G in fig. 2), has little difference in the spectrum obtained by scanning.
1.3.2 modeling analysis
Fig. 1G and 1H are abnormal spectra found when the calibration set spectra (including modern ivory and mammoth ivory) were qualitatively analyzed during the modeling with SIMCA. Of the 208 calibration set spectra, 23 abnormal spectra were found under the conditions that the F value is 2.5 as a default value and the recommended main factor value is 3, wherein 9 abnormal spectra are spectral samples collected from the black part of the mammoth ivory and 10 abnormal spectra are spectral samples of the fan blade of the ivoth ivothy.
1.4 conclusion
The color of the sample has certain influence on the scanning result, like the same substance, the absorbance of the dark and light parts has difference, even if the visual spectrum graph has no difference, the software is easy to be divided into abnormal samples after calculation and analysis, and the spectrum of the sample still has certain difference with the spectrum of other samples. Similarly, if the thickness of the sample is too thin (for example, 10 fan bone pieces in 1.3.2, which are judged to be abnormal samples, are all less than 1mm in thickness), erroneous judgment is also easily caused. When detecting a sample, if the colors of different parts of the sample are different and the thicknesses are different, the part with the typical color of the substance should be scanned, and the part with a certain thickness should be scanned at the same time, so as to avoid causing misjudgment.
2. Model building for ivory identification
2.1 samples and Spectrum samples
153 samples and 208 spectral samples for the calibration set and 77 samples and 175 spectral samples for the validation set of 21 classes of different properties, detailed in table 1.
2.2 model building
According to the descriptions (1) to (6) in the "summary of the invention", a preliminary model is established. The model building process involves main factor-F value clustering judgment, and when 208 correction set spectrum samples and 175 verification set spectrum samples are modeled, the recommended main factor value is 3, and at the moment, the sample contains 96.5% of information content. The initial default F threshold is about 2.5, as shown in fig. 3A, about 20 samples of the correction samples at this time are listed as abnormal, and the rest are normal correction samples; however, the verification sample is also shown (fig. 3B), and a considerable part of the verification sample is listed as a normal sample, which means that the non-ivory material sample corresponding to the part of the spectrum is also identified as ivory. Adjusting the F value to about 1.14, the number of correction samples that are classified as abnormal increases (FIG. 3C), but a significant number of verification samples are classified as normal (FIG. 3D). The F value was further adjusted to about 0.57 (fig. 3E and 3F), and statistics showed that the number of normal samples in the calibration sample was 155 (155/208 × 100% — 75.4%), the number of abnormal samples was 53, the number of normal samples in the verification sample was decreased to 13 (13/175 × 100% — 7.4%), and the number of abnormal samples was increased to 162. Although only 53 calibration samples were included in the abnormal samples (i.e., the ivory samples corresponding to the 53 spectra were determined to be non-ivory), 13 verification samples were considered to be normal samples (i.e., the non-ivory samples corresponding to the 13 spectra were determined to be ivory). It was further confirmed that the samples misjudged as ivory were mainly animal bones including horse bear bones and gold cat bones.
In the case where the F value remains 0.57, the main factor value is adjusted. It was found that when the main factor was adjusted from 3 to 1 or 2, the proportion of normal samples in the correction set increased, while at the same time the proportion of normal samples in the verification set also increased, e.g. when decreased to 2, the proportion of normal samples in the correction set increased from 75.4% to 88.9% (185/208 x 100%), and the proportion of normal samples in the verification set increased from 7.4% to 12.0% (21/175 x 100%); while adjusting the main factor from 3 to 4 or 5, the proportion of normal samples in the calibration set decreased, while the number of normal samples in the verification set also decreased, e.g., increased to 4, the proportion of normal samples in the calibration set decreased from 75.4% to 60.1% (125/208 x 100%), and the proportion of normal samples in the verification set decreased from 7.4% to 4.6% (8/175 x 100%). Therefore, under the condition that the F value is kept unchanged, the main factor is reduced from 3 to 1 or 2, so that false negative is reduced, but the false positive rate is increased; if the main factor is increased from 3 to 4 or 5, the false negative rate will increase and the false positive rate will decrease.
Whether the F value or the main factor value is adjusted, the aim is to classify as many samples as possible in the correction set into normal samples, thereby reducing the false negative of the detection result; meanwhile, as few samples as possible in the verification set are classified into normal samples, and the false positive of the detection result is reduced. And adjusting the F value and the main factor value to keep the false positive rate and the false negative rate at a lower level, and generally comparing false negatives and more tolerating false positives according to the detection purpose, such as preliminary screening for fighting smuggling of endangered animals and plants. Therefore, the F value is finally set to 0.57, the main factor is set to 2, the settings are saved, and the model building is completed.
3. Model validation of ivory identification
3.1 accuracy of recognition
The accuracy of the sample identification in table 1 was verified using the model established at 2.2. The morphological identification and molecular biological identification methods refer to the methods of customs service general administration of methods for identifying ivory and ivory products (trial implementation), department of Sichu (2019) No. 75.
The three different methods are determined as follows. And (3) judging the near infrared spectrum identification result: the sample is judged to be ivory, namely, a plurality of spectra are measured on one sample, and the sample is judged to be ivory as long as the result of one spectrum is displayed as ivory; otherwise, it is judged as "non-ivory". Judging a morphological identification result: the ivory is judged to be an ivory, namely, the ivory has an obvious Schneider structure, and the hot needle has no hole mark and no smoke; non-ivory, namely, no schlieren structure, hole marks are poked by a hot needle, and white smoke is emitted; the 'unidentifiable' is that no Schneider structure exists, no hole mark exists in the hot needle stamp, and no smoke exists. And (3) judging the molecular biological identification result: the test is judged as "ivory", namely the molecular test result is an African elephant, an Asian elephant or a mammoth; "non-ivory", i.e., species identified as other than elephant; "unidentified", i.e., the species was not identified.
The spectral identification accuracy is 100% of the number of samples identified accurately per total number of samples. The identification result is based on the result (comprehensive judgment result) of combining morphology with molecular biology, namely one of the two methods is identified as 'ivory', namely 'ivory'; one method is identified as 'non-ivory', namely 'non-ivory'; if both methods are "unidentifiable", then "unidentifiable"; when two methods are contradictory, namely one method results in ivory, and the other method results in non-ivory, the verification is required, but the contradictory results are not generally generated.
The spectra of the samples in table 1 were detected using the established model, and the obtained identification results showed (table 1) that all samples identified as ivory morphologically and/or molecularly biologically were ivory, and the identification results of this model were also ivory without false negatives, so that the identification accuracy of ivory reached 100%. However, when identifying the bones of the bear and the golden cat, the false positive can be completely avoided in actual work because the false positive rate is 100% and 66.7%, respectively, but even if the bones of the bear and the golden cat are identified by naked eye sensory analysis, the bones of the bear and the golden cat can be very easily distinguished from the ivory. However, for samples made of other materials, including ivory imitations which are most easily identified, such as plastic imitations of ivory and ivory fruit products, the identification accuracy can reach 100%.
Because in the primary screening procedure for ivory preparations, once negative (i.e., non-ivory) in the primary screening, identification may not be subsequently performed, while false positives may continue to be identified in subsequent tests. Therefore, the ivory identification method based on the near infrared spectrum can effectively avoid misjudging the sample as negative (false negative), does not put through real ivory products, and is more beneficial to striking behaviors such as ivory buying and selling and the like.
3.2 prediction of unknown samples
The 20 suspected ivory preparations captured by customs were identified. A near infrared spectrum identification method, namely, the model established in the step 2 is utilized; the morphological and molecular biological identification methods are referred to the methods in Customs 'general administration of ancient molar and ivory products identification methods (trial) in the' Shujia 'Ke' No. 2019, 75. The three different methods are determined in the same manner as 3.1.
And (3) according to the result obtained by the near infrared detection, 18 samples in 20 samples are consistent with comprehensive judgment, wherein 11 samples are ivory, and 7 samples are non-ivory. The samples (No. 13 and No. 14 samples in the table 2) with the results of the 2 near infrared spectrums which do not accord with the comprehensive judgment are judged as ivory, the results of the comprehensive judgment cannot be identified, but the true ivory has no typical texture and is easy to have the condition that DNA cannot be extracted, and the samples are macroscopically identified as the suspected ivory samples, so that the two samples have the same probability of being the true ivory.
TABLE 1 samples used for modeling and results of identification
Note: if the sample has a large volume and different sections or the color, thickness and the like of each part are different, 2-4 original spectra are collected at different parts of the same sample according to the circumstances;
a : the results of African ivory are obtained by molecular identification of 26 samples, wherein 23 samples have the specific Schneider structure of ivory, so the morphology can be judged as ivory, and the rest 3 samples have no any grain, so the morphology cannot be judged;
b : the results of Asian ivory are obtained by molecular identification of 18 samples, wherein 14 samples have the specific Schneider structure of the ivory, so the morphology can be judged as the ivory, and the rest 4 samples have no grain, so the morphology cannot be judged;
c : 99 samples all have the special texture of the ivory, so the morphology can be judged as the ivory, but the molecular biology can not extract DNA, so the specific ivory can not be identified;
d : the mammoth ivory is a sample which is sent for inspection in the same batch, the samples have the specific Schleman structure of the ivory, the Schleman angle formed by the Schleman structure is less than 90 degrees, and the samples are provided with blueskin; but 4 molecular biological identification results are mammoth mammoths, and the rest species cannot be identified; in view of the morphological characteristics described above, samples that were not detected by molecular biology were also classified as mammoth ivory.
TABLE 2 identification results of suspected ivory products
Claims (10)
1. An ivory identification method based on near infrared spectrum is characterized by comprising the following steps:
s1, collecting a sample;
s2, selecting a spectrum acquisition part of a sample and scanning a near infrared spectrum;
s3, establishing a near infrared spectrum identification model of the sample;
s4, judging a near infrared spectrum identification result: and analyzing the near infrared spectrum data of the sample to be detected by using a near infrared spectrum identification model to give a sample identification result.
2. The method of claim 1, wherein the sample spectrum collection portion is selected in step S2 to have a color typical of the sample.
3. The method according to claim 1, wherein the parameters of the near infrared spectrum scan in step S2 are: the spectrum scanning range is 1000 nm-1800 nm, the resolution is 11nm, the scanning times are 30 times, and each sample is scanned for 2 times.
4. The method as claimed in claim 1, wherein the near infrared spectroscopy analysis model in the step of S3 has an F value set to 0.57 and a main factor set to 2.
5. The method according to claim 1, wherein the determination of the near infrared spectrum identification result in the step S4: one sample is measured with a plurality of spectra, and the ivory is judged as the ivory as long as the result of one spectrum shows the ivory.
6. The method according to claim 1, wherein the establishing of the near infrared spectroscopy analysis model comprises the following steps:
(1) collecting a sample;
(2) collecting a near infrared spectrum: acquiring basic spectrum by using a spectrometer with a spectrum scanning range of 1000-1800 nm, a resolution of 11nm and a scanning frequency of 30 times; scanning in a near-infrared diffuse reflection mode, collecting a basic spectrum, scanning each sample for 2 times, and taking the obtained average value as an original spectrum of the sample; scanning each sample by the same method in sequence to obtain an average value; identifying and deleting the acquired abnormal spectrum sample;
(3) dividing a correction set and a verification set according to the properties of a known sample, wherein the correction base is an ivory spectrum, and the verification set is a non-ivory spectrum;
(4) spectrum pretreatment: preprocessing the calibration set and the validation collection spectra;
(5) extracting characteristic wave bands from the preprocessed diffuse reflection spectrum, and selecting a sensitive spectrum band for ivory identification;
(6) establishing and verifying ivory models: calculating by using the extracted characteristic wavelength through the Mahalanobis distance, establishing an identification model by adopting an SIMCA qualitative analysis method, and deducing to obtain an optimal main factor value and an optimal F value;
(7) and (3) identifying the accuracy: the validation set and samples of known classes are used to detect the accuracy of the model.
7. The method of claim 2, wherein the step (2) of identifying and deleting the collected abnormal spectral samples comprises direct deletion of spectra with spectral features and substances of the same nature that differ significantly and spectra with absorbance exceeding 1.5.
8. The method of claim 2, wherein said ivory spectra in step (3) comprise spectra of an african elephant, an asian elephant, and a mammoth elephant ivory.
9. The method of claim 2, wherein the calibration set and the validated collection spectra are preprocessed in step (4) and include standard normal variable transformation, detrending correction, Savitzky-Golay smoothing, Savitzky-Golay derivatives, and mean centering.
10. The method according to claim 2, wherein the sensitive bands for ivory identification in step (5) are 1160-1200 nm, 1430-1500 nm, 1680-1710 nm, 1720-1750 nm, respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210408852.9A CN114878506A (en) | 2022-04-19 | 2022-04-19 | Ivory identification method based on near infrared spectrum |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210408852.9A CN114878506A (en) | 2022-04-19 | 2022-04-19 | Ivory identification method based on near infrared spectrum |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114878506A true CN114878506A (en) | 2022-08-09 |
Family
ID=82671646
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210408852.9A Pending CN114878506A (en) | 2022-04-19 | 2022-04-19 | Ivory identification method based on near infrared spectrum |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114878506A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117740712A (en) * | 2024-02-19 | 2024-03-22 | 中国海关科学技术研究中心 | Ivory for customs ports and method and system for rapid preliminary screening and identification of products thereof |
-
2022
- 2022-04-19 CN CN202210408852.9A patent/CN114878506A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117740712A (en) * | 2024-02-19 | 2024-03-22 | 中国海关科学技术研究中心 | Ivory for customs ports and method and system for rapid preliminary screening and identification of products thereof |
| CN117740712B (en) * | 2024-02-19 | 2024-04-19 | 中国海关科学技术研究中心 | Ivory for customs ports and method and system for rapid preliminary screening and identification of products thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Qiao et al. | Pork quality and marbling level assessment using a hyperspectral imaging system | |
| Fuentes et al. | Automated grapevine cultivar classification based on machine learning using leaf morpho-colorimetry, fractal dimension and near-infrared spectroscopy parameters | |
| Moscetti et al. | Detection of mold-damaged chestnuts by near-infrared spectroscopy | |
| Huang et al. | Non-destructively sensing pork’s freshness indicator using near infrared multispectral imaging technique | |
| CN107884340B (en) | Spectroscopic characterization of seafood | |
| Downey et al. | Discrimination between fresh and frozen-then-thawed beef m. longissimus dorsi by combined visible-near infrared reflectance spectroscopy: A feasibility study | |
| CN109540831B (en) | Wolfberry variety identification method based on hyperspectral imaging technology | |
| Abdullah et al. | Automated inspection system for colour and shape grading of starfruit (Averrhoa carambola L.) using machine vision sensor | |
| García-Rey et al. | Prediction of texture and colour of dry-cured ham by visible and near infrared spectroscopy using a fiber optic probe | |
| Li et al. | Wood defect classification based on two-dimensional histogram constituted by LBP and local binary differential excitation pattern | |
| CN101692052A (en) | Hyperspectrum image technique-based method and hyperspectrum image technique-based device for identifying true and false famous tea | |
| CN114878506A (en) | Ivory identification method based on near infrared spectrum | |
| WO2007000165A1 (en) | Online recording of wavelength absorption spectra in meat | |
| Prades et al. | Methodology for cork plank characterization (Quercus suber L.) by near-infrared spectroscopy and image analysis | |
| CN108801971A (en) | Mould based on high light spectrum image-forming technology infects the detection method of Chinese chestnut | |
| Yin et al. | Non-destructive detection of foreign contaminants in toast bread with near infrared spectroscopy and computer vision techniques | |
| Sun et al. | Predicting pork color scores using computer vision and support vector machine technology | |
| JP2002328088A (en) | Discriminating method of abnormal meat by near-infrared spectroscopy | |
| CN103076300B (en) | Method for judging and analyzing traditional Chinese medicine resource fingerprint information by specific mode identification model | |
| Shimoyama et al. | Non-destructive analysis of the two subspecies of African elephants, mammoth, hippopotamus, and sperm whale ivories by visible and short-wave near infrared spectroscopy and chemometrics | |
| CN111595805A (en) | Possibility-clustering Chinese cabbage pesticide residue qualitative analysis method | |
| Prades et al. | Evaluation of parameters that determine cork plank quality (Quercus suber L.) by near infrared spectroscopy | |
| CN108956528A (en) | The near-infrared Undamaged determination method of mahogany furniture | |
| CN113866122A (en) | Method for rapidly identifying chicken varieties and application thereof | |
| Bond | Characterization of wood features using color, shape, and density parameters |
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 |