Method, device and equipment for identifying mixed pigment spectral partition
By performing reflectance correction and noise reduction on the hyperspectral images of painted cultural relics, calculating the first derivative curve, extracting feature sub-intervals, and combining spectral similarity algorithms, accurate identification of mixed pigments was achieved, solving the problem of insufficient identification accuracy in existing technologies and supporting the protection and restoration of cultural relics.
Patent Information
- Application Number
- CN202310700339.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-14
- Publication Date
- 2025-12-19
- Estimated Expiration
- 2043-06-14
AI Technical Summary
Existing technologies struggle to effectively identify the components of mixed pigments, especially in painted cultural relics. Hyperspectral technology has difficulty processing the nonlinear spectra of mixed pigments, resulting in insufficient identification accuracy.
A mixed pigment spectral segmentation identification method is adopted. By acquiring hyperspectral images and dark current data, reflectance correction is performed, noise reduction is performed, the first derivative curve is calculated, feature sub-intervals are extracted, and similarity is calculated using the spectral angle cosine combined with the normalized Euclidean distance algorithm to generate the identification results of mixed pigments.
It improves the accuracy and efficiency of mixed pigment identification, effectively identifying the components of mixed pigments in painted cultural relics, and supporting the restoration and protection of cultural relics.
Smart Images
Figure CN116612108B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of remote sensing, and more particularly, to a mixed pigment spectrum partition interval identification method, device and equipment. BACKGROUND
[0002] The painted cultural relics are colorful and lifelike, and have a high artistic appreciation and archaeological research value, and are the precious gems of the brilliant culture and art of ancient times. The painted cultural relics occupy an important position in the whole cultural relics, and are an important component of the cultural relics system. However, due to the long time of existence, the painted cultural relics are inevitably affected by natural environment (humidity, high temperature, light, dust) or improper human protection measures, and different degrees of natural aging and diseases occur, which seriously affects the normal display and use of the cultural relics, reduces the historical and artistic research value, and even seriously threatens the survival of the cultural relics. The ingenious collocation of various pigments on the surface of the cultural relics constitutes meaningful color symbols, characters or patterns, is the most important carrier of heritage information, is the most direct evidence for interpreting the era characteristics and regional characteristics, and is the most critical data support for repairing and protecting the cultural relics. Therefore, the quantitative identification and analysis of the pigments of the painted cultural relics are imminent.
[0003] At present, various modern technologies have been tentatively applied to the analysis of cultural relics pigments, such as polarized light microscopy, Raman spectroscopy, X-ray diffraction, fluorescence X-ray, scanning electron microscopy-energy dispersive X-ray, electron microscopy, near-infrared spectrometer, etc. Although these technologies can basically determine the elements and structure of the pigments, they still have some limitations in practical application, such as the need for sampling, the detection of point or local area, and the difficulty in quantitative analysis of the specific content of each pigment on the whole surface of the cultural relics. The hyperspectral technology is gradually applied to the identification of cultural relics pigments because it can obtain image information and spectral information of the surface of the cultural relics without contacting the target. However, the existing identification algorithms are difficult to analyze the spectrum of mixed pigments, mainly because the mixed pigments are mostly nonlinear, and when light is irradiated on the surface of the mixed pigments, the particles will produce strong lattice vibration and electronic transition, and the mechanism is more complex, which further increases the difficulty of identification. SUMMARY
[0004] According to the embodiments of the present application, a mixed pigment spectrum partition interval identification scheme is provided. The scheme can identify the mixed pigments and improve the accuracy of the identification results.
[0005] In a first aspect of the present application, a mixed pigment spectrum partition interval identification method is provided. The method comprises:
[0006] Collecting hyperspectral images of a target object to be identified, standard reflectance plate and dark current data, and performing reflectance correction;
[0007] The reflectivity-corrected image is subjected to data noise reduction processing, and the unknown pigment spectrum and the standard pigment spectrum subjected to the noise reduction processing are converted into corresponding K / S curves;
[0008] The unknown pigment spectrum is subjected to first derivative calculation to obtain a derivative curve of the unknown pigment spectrum, and a characteristic subinterval of the K / S curve corresponding to the unknown pigment spectrum is extracted from the derivative curve of the unknown pigment spectrum;
[0009] If the number of the characteristic subintervals of the K / S curve corresponding to the unknown pigment spectrum is 1, the similarity of the K / S curves corresponding to the unknown pigment and the standard pigment is calculated in a full waveband range, and the standard pigment with the highest similarity is taken as the identification result; otherwise, the similarity of the K / S curves corresponding to the unknown pigment and the standard pigment is calculated in each characteristic subinterval to obtain a plurality of interval identification result sets, a K / S curve corresponding to a mixed pigment is generated, and the standard pigment with the highest similarity of the K / S curves corresponding to the mixed pigment and the standard pigment is taken as the identification result of the unknown pigment.
[0010] Further, the reflectivity correction comprises:
[0011] The standard reflectance plate correction data is calculated according to the standard reflectance plate hyperspectral image and the dark current data;
[0012] The dark current data is expanded into dark current correction data with the same row and column number as the target object hyperspectral image;
[0013] The target object correction data is recalculated, and the target object hyperspectral reflectivity image is calculated according to the target object correction data.
[0014] Further, the data noise reduction processing on the reflectivity-corrected image comprises:
[0015] The reflectivity-corrected image is subjected to MNF forward transformation to output first transformation data;
[0016] The first transformation data is subjected to MNF inverse transformation to obtain second transformation data.
[0017] Further, the characteristic subinterval of the K / S curve corresponding to the unknown pigment spectrum is extracted from the derivative curve of the unknown pigment spectrum, comprising:
[0018] The difference between two adjacent wave troughs in the derivative curve of the unknown pigment spectrum is taken as a section, and the larger value of the difference between the wave peak and the two wave troughs in each section is taken as the peak-valley difference of the current section;
[0019] Sort the peak-valley difference of each section, judge whether the multiple relationship between the maximum value of the peak-valley difference and the second maximum value of the peak-valley difference is greater than the multiple threshold value, if yes, only the section corresponding to the maximum value of the peak-valley difference is reserved as the spectral range of the characteristic sub-interval of the K / S curve; otherwise, the two sections corresponding to the maximum value of the peak-valley difference and the second maximum value of the peak-valley difference are reserved as the spectral range of the characteristic sub-interval of the K / S curve.
[0020] According to the spectral range of the characteristic sub-interval, the characteristic sub-intervals of the K / S curve of the unknown pigment are extracted.
[0021] Further, the similarity of the K / S curve of the unknown pigment and the standard pigment is calculated in each characteristic sub-interval, and a plurality of interval identification result sets are obtained, including:
[0022] In the range of each characteristic sub-interval, the similarity of the K / S curve of the unknown pigment and the standard pigment is calculated by using the spectral angle cosine combined with the normalized Euclidean distance algorithm, and the N standard pigments with the highest similarity are taken as the identification result of the characteristic sub-interval.
[0023] From the identification result of each characteristic sub-interval, a K / S curve corresponding to a standard pigment is taken out in turn, and is combined in pairs to obtain N^2 interval identification result sets.
[0024] Further, the K / S curve of the mixed pigment is generated, the standard pigment with the highest similarity of the mixed pigment and the K / S curve of the standard pigment is taken as the identification result of the unknown pigment, including:
[0025] Two standard pigment K / S curves in each interval identification result set are taken as end members to generate a plurality of simulated mixed K / S curves with different abundances.
[0026] The similarity of the simulated mixed K / S curve and the K / S curve of the standard pigment is calculated by using the spectral angle cosine combined with the normalized Euclidean distance algorithm, and the highest similarity value of each interval identification result set is obtained.
[0027] The highest similarity values of the interval identification result sets are compared, and the standard pigment in the interval identification result set corresponding to the highest similarity is taken as the component of the mixed pigment.
[0028] In the second aspect of the present application, a mixed pigment spectral partition identification device is provided. The device comprises:
[0029] A correction acquisition module is configured to acquire a hyperspectral image of a target object to be identified, a hyperspectral image of a standard reflector and dark current data, and perform reflectivity correction.
[0030] The noise reduction conversion module is used for carrying out data noise reduction processing on the reflectivity corrected image, and converts the unknown pigment spectrum and the standard pigment spectrum after the noise reduction processing into corresponding K / S curves.
[0031] The calculation extraction module is used for carrying out first derivative calculation on the unknown pigment spectrum to obtain a derivative curve of the unknown pigment spectrum, and extracts a characteristic subinterval of the K / S curve corresponding to the unknown pigment spectrum in the derivative curve of the unknown pigment spectrum.
[0032] The judgment and recognition module is used for judging whether the number of the characteristic subintervals of the K / S curve corresponding to the unknown pigment spectrum is 1, and if yes, calculating the similarity of the K / S curves corresponding to the unknown pigment and the standard pigment in the full wave band range, and taking the standard pigment with the highest similarity as the recognition result; otherwise, calculating the similarity of the K / S curves corresponding to the unknown pigment and the standard pigment in each characteristic subinterval to obtain a plurality of interval recognition result sets, generating a K / S curve corresponding to a mixed pigment, and taking the standard pigment with the highest similarity of the K / S curves corresponding to the mixed pigment and the standard pigment as the recognition result of the unknown pigment.
[0033] In a third aspect of the present application, an electronic device is provided. The electronic device comprises at least one processor; and a memory connected to the at least one processor in communication; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method of the first aspect of the present application.
[0034] In a fourth aspect of the present application, a non-transitory computer readable storage medium storing computer instructions is provided, and the computer instructions are used to enable the computer to execute the method of the first aspect of the present application.
[0035] It should be understood that the content described in the summary section is not intended to limit the key or important features of the embodiments of the present application, nor to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS
[0036] The above and other features, advantages, and aspects of embodiments of the present application will become more apparent by describing in detail the following embodiments with reference to the attached drawings. In the drawings, the same or similar reference numerals refer to the same or similar elements, and:
[0037] Figure 1 A flow chart of a mixed pigment spectrum interval recognition method according to an embodiment of the present application is shown;
[0038] Figure 2 A first derivative curve schematic diagram according to an embodiment of the present application is shown;
[0039] Figure 3 A schematic diagram of reflectance spectrum of mixed pigments is shown according to an embodiment of the present application;
[0040] Figure 4 A block diagram of a mixed pigments spectrum partition interval identification device according to an embodiment of the present application is shown;
[0041] Figure 5 A block diagram of an exemplary electronic device capable of implementing an embodiment of the present application is shown;
[0042] Wherein, 500 is an electronic device, 501 is a computing unit, 502 is a ROM, 503 is a RAM, 504 is a bus, 505 is an I / O interface, 506 is an input unit, 507 is an output unit, 508 is a storage unit, and 509 is a communication unit. DETAILED DESCRIPTION
[0043] To make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.
[0044] In addition, the term “and / or” herein merely describes an association relationship of associated objects, and indicates that there can be three relationships, for example, A and / or B can represent three cases of A alone, A and B together, and B alone. In addition, the character “ / ” herein generally represents an “or” relationship between the front and rear associated objects.
[0045] Figure 1 A flow chart of a mixed pigments spectrum partition interval identification method according to an embodiment of the present application is shown.
[0046] The method comprises:
[0047] S101, collecting hyperspectral images of a target object to be identified, hyperspectral images of a standard reflector plate, and dark current data, and performing reflectance correction.
[0048] In the embodiment, the hyperspectral images of the target object comprise hyperspectral images of unknown pigments and hyperspectral images of standard pigments.
[0049] The reflectance correction comprises:
[0050] First, standard reflector plate correction data is calculated according to the hyperspectral images of the standard reflector plate and the dark current data. The standard reflector plate correction data is:
[0051] Rstandard = R white - R dark
[0052] wherein, R white is the standard reflectance plate hyperspectral image; R dark is the dark current data; R standard is the standard reflectance plate corrected data.
[0053] Secondly, the dark current data is expanded into dark current corrected data with the same row and column number as the target hyperspectral image.
[0054] Thirdly, the target hyperspectral image is subtracted by the dark current corrected data to obtain target corrected data. The target corrected data is:
[0055] R corrected = R raw - R' dark
[0056] wherein, R raw is the target hyperspectral image; R' dark is the dark current corrected data; R corrected is the target corrected data.
[0057] Finally, the target corrected data is recalculated, and the target hyperspectral reflectance image is calculated according to the target corrected data, so as to complete the reflectance correction, that is:
[0058]
[0059] wherein, R is the target hyperspectral reflectance image.
[0060] S102, data denoising processing is performed on the reflectance corrected image, and the unknown pigment spectrum and the standard pigment spectrum after the denoising processing are converted into corresponding K / S curves.
[0061] In the embodiment, the data denoising processing on the reflectance corrected image comprises:
[0062] Firstly, MNF forward transformation is performed on the reflectance corrected image to output first transformation data; specifically, MNF forward transformation can be performed in a remote sensing data processing software, the band selection is 51-990 range, and the output band number is set to 10.
[0063] Secondly, MNF inverse transformation is performed on the first transformation data in the remote sensing data processing software to obtain second transformation data. The second transformation data comprises the unknown pigment spectrum and the standard pigment spectrum.
[0064] Finally, the unknown pigment spectrum and the standard pigment spectrum after the noise reduction processing are converted into corresponding K / S curves.
[0065] By the noise reduction processing, the interference of noise on the image data can be reduced.
[0066] S103, first derivative calculation is performed on the unknown pigment spectrum to obtain a derivative curve of the unknown pigment spectrum, and a characteristic subinterval of a K / S curve corresponding to the unknown pigment spectrum is extracted from the derivative curve of the unknown pigment spectrum.
[0067] In the embodiment, the characteristic subinterval of the K / S curve corresponding to the unknown pigment spectrum is extracted from the derivative curve of the unknown pigment spectrum, including:
[0068] First, the difference between two adjacent wave troughs in the derivative curve of the unknown pigment spectrum is taken as a section, and the larger value of the difference between the wave peak and the two wave troughs in each section is taken as the peak-valley difference of the current section. For example, the two adjacent wave troughs are wave trough 1 and wave trough 2, and the difference between the wave trough 1 and the wave peak is greater than the difference between the wave trough 2 and the wave peak, so the difference between the wave trough 1 and the wave peak is taken as the peak-valley difference of the current section from the wave trough 1 to the wave trough 2.
[0069] Secondly, the peak-valley differences of each section are sorted, and it is judged whether the multiple relationship between the maximum value of the peak-valley difference and the second maximum value of the peak-valley difference is greater than a multiple threshold value. If yes, only the section corresponding to the maximum value of the peak-valley difference is reserved as the spectral range of the characteristic subinterval of the K / S curve; otherwise, the two sections corresponding to the maximum value and the second maximum value of the peak-valley difference are reserved as the spectral range of the characteristic subinterval of the K / S curve. For example, the multiple threshold value is set to 10 times, and if the maximum value of the peak-valley difference is 10 times or more than the second maximum value of the peak-valley difference, the section corresponding to the second maximum value of the peak-valley difference is discarded.
[0070] Finally, the characteristic subinterval of the unknown pigment K / S curve is extracted according to the spectral range of the characteristic subinterval. It can be seen that when the multiple relationship between the maximum value of the peak-valley difference and the second maximum value of the peak-valley difference is greater than the multiple threshold value, the number of the characteristic subintervals of the unknown pigment K / S curve is one, and when the multiple relationship between the maximum value of the peak-valley difference and the second maximum value of the peak-valley difference is not greater than the multiple threshold value, the number of the characteristic subintervals of the unknown pigment K / S curve is two.
[0071] S104, if the number of the characteristic sub-intervals of the K / S curve corresponding to the unknown pigment spectrum is 1, similarity of the K / S curve corresponding to the unknown pigment and the standard pigment is calculated in the full waveband range, and the standard pigment with the highest similarity is taken as the identification result; otherwise, similarity of the K / S curve corresponding to the unknown pigment and the standard pigment is calculated in each characteristic sub-interval, a plurality of interval identification result sets are obtained, a K / S curve corresponding to a mixed pigment is generated, and the standard pigment with the highest similarity of the K / S curve corresponding to the mixed pigment and the standard pigment is taken as the identification result of the unknown pigment.
[0072] In the embodiment, if the number of the characteristic sub-intervals of the K / S curve corresponding to the unknown pigment spectrum is 1, similarity of the K / S curve corresponding to the unknown pigment and the standard pigment is calculated in the full waveband range, and the standard pigment with the highest similarity is taken as the identification result, which specifically includes:
[0073]
[0074]
[0075] Similarity (x,y) =(1-SAC (x,y) )×0.6+Ed (x,y) ×0.4
[0076] Wherein, x i represents a value of the K / S curve corresponding to the unknown pigment at the i-th waveband; y i represents a value of the K / S curve corresponding to the standard pigment at the i-th waveband; SAC (x,y) represents a spectral angle cosine value of the K / S curve corresponding to the unknown pigment and the standard pigment in the full waveband range; and Ed (x,y) represents an Euclidean distance value of the K / S curve corresponding to the unknown pigment and the standard pigment in the full waveband range.
[0077] In the embodiment, if the number of the characteristic sub-intervals of the K / S curve corresponding to the unknown pigment spectrum is 2, the following process is performed:
[0078] Firstly, in each characteristic sub-interval range, similarity of the K / S curve corresponding to the unknown pigment and the standard pigment is calculated by using the spectral angle cosine combined with the normalized Euclidean distance algorithm, and the calculation is as follows:
[0079]
[0080]
[0081] Similarity (x',y') =(1-SAC (x',y') )×0.6+Ed (x',y')× 0.4
[0082] wherein, x' i represents the value of the K / S curve corresponding to the unknown pigment at the i-th waveband of the characteristic sub-interval; y' i represents the value of the K / S curve corresponding to the standard pigment at the i-th waveband of the characteristic sub-interval; SAC (x',y') represents the spectral angle cosine value of the K / S curves corresponding to the unknown pigment and the standard pigment in the characteristic sub-interval; Ed (x',y') represents the Euclidean distance value of the K / S curves corresponding to the unknown pigment and the standard pigment in the characteristic sub-interval.
[0083] The N standard pigments with the highest similarity are taken as the identification result of the characteristic sub-interval. N represents the number of standard pigments with the highest similarity, for example, 3, N = 3, that is, the top three standard pigments with the highest similarity are retained as the identification result of each sub-interval.
[0084] Secondly, one K / S curve corresponding to a standard pigment is taken from the identification result of each characteristic sub-interval in turn, and is combined in pairs to obtain N^2 interval identification result sets. For example, the identification result Result1 of the characteristic sub-interval 1 is a1, a2, a3; the identification result Result2 of the characteristic sub-interval 2 is b1, b2, b3; the two standard pigments retained by the two sub-intervals are combined in pairs to generate 9 interval identification result sets, which are U1 = {a1, b1}, U2 = {a1, b2}, U3 = {a1, b3},..., U9 = {a3, b3}.
[0085] Thirdly, the two K / S curves of the standard pigments in each interval identification result set are taken as end members to generate a plurality of simulated mixed K / S curves with different abundances. For example, a 2*1000-dimensional abundance matrix A is obtained according to the Dirichlet distribution function, and the abundance is randomly distributed from 0 to 1; the two K / S curves of the standard pigments in the interval identification result set are combined with the abundance matrix A to generate 1000 simulated mixed K / S curves mixed with different abundances.
[0086] Then, the spectral angle cosine combined with the normalized Euclidean distance algorithm is used to calculate the similarity between the simulated mixed K / S curve and the K / S curve corresponding to the standard pigment, and the highest similarity value of each interval identification result set is obtained.
[0087] The similarity calculation is as follows:
[0088]
[0089]
[0090] Similarity(m, y) = (1 - SAC (m,y) ) × 0.6 + Ed(m,y) x 0.4
[0091] wherein, m i represents the value of the analog mixed K / S curve at the i-th waveband; y i represents the value of the standard pigment corresponding K / S curve at the i-th waveband; SAC (m,y) represents the spectral angle cosine value of the analog mixed K / S curve and the standard pigment corresponding K / S curve; ED (m,y) represents the Euclidean distance value of the analog mixed K / S curve and the standard pigment corresponding K / S curve.
[0092] The above obtains 1000 similarities, and a highest value among the 1000 similarities is reserved as a highest similarity value of the current interval identification result set.
[0093] Finally, the highest similarity value of each interval identification result set is compared, and the standard pigment in the interval identification result set corresponding to the highest similarity value is taken as a component of the mixed pigment.
[0094] The above process is further described below through specific embodiments.
[0095] From the pigment standard sample library, common azurite, malachite, orpiment and cinnabar in murals are selected to make pure pigment samples, and then the four pure pigments are mixed in pairs according to a mass ratio of 50:50 to make six groups of mixed pigment samples. The sample preparation steps are as follows: take alum particles as needed, soak in cold water until transparent, then add a small amount of boiling water, and stir until the gelatin is completely dissolved; put the pigment powder into a palette, slowly drop the gelatin solution into the palette with a dropper, and stir with fingers along the clockwise direction until the gelatin solution and the pigment powder are uniformly mixed; finally, dip the brush with the pigment, and draw on 4*4cm rice paper.
[0096] A VNIR400H hyperspectral imager of Themis Vision System Company, USA, is used, the single scene frame size is 1392 pixels*1000 pixels, the spectral resolution is 2.8nm, and the spectral range covering 400-1000nm wavelength contains a total of 1040 wavebands of visible light and near infrared. After imaging data collection, whiteboard correction and noise reduction pretreatment, the spectral curves of 4 groups of pure samples and 6 groups of mixed samples are extracted.
[0097] The first derivative of the unknown pigment reflectance spectrum is calculated, and the mixed spectrum of orpiment and cinnabar is taken as an example. The reflectance spectrum of the mixed pigment is as follows: Figure 3 , wherein the abscissa represents the waveband; and the ordinate represents the reflectance.
[0098] The range of the characteristic sub-interval is determined according to the "convex" of the first derivative curve of the unknown pigment, and if the number of sub-intervals is greater than 2, only the two most obvious characteristic sub-intervals of the "convex" are reserved. The first derivative curve is shown in Fig. 2. The abscissa represents the wave band, and the ordinate represents the derivative value. There are two obvious "convexes" on the first derivative curve of the mixed pigment, i.e. the 430-560 nm and 561-865 nm regions, indicating that the reflectivity of the mixed spectrum of lead yellow and cinnabar increases obviously in the corresponding wave band range, and then remains stable or slowly increases. Figure 2
[0099] Secondly, the reflectivity is converted into the absorption and scattering ratio (K / S) which is more in line with the linear mixing characteristics by the KM model. The K / S curves of the unknown pigment and the standard pigments are normalized by the maximum and minimum values in the range of the characteristic sub-interval, and the similarity between the K / S curves of the unknown pigment and the standard pigments is calculated by using the spectral angle cosine combined with the normalized Euclidean distance. The first three results with the highest similarity are reserved in each characteristic sub-interval. The characteristic sub-intervals of the six groups of mixed pigments are shown in Table 1.
[0100] Mixed pigments Subinterval 1 Subinterval 2 Azurite malachite 400-525 nm 790-985 nm Azurite topaz 440-605 nm 500-1000 nm Azurite cinnabar 400-515 nm 516-660 nm Malachite topaz 430-700 nm 745-1000 nm Malachite cinnabar 400-575 nm 576-650 nm Topaz cinnabar 430-560 nm 561-865 nm
[0101] Finally, one K / S curve of the standard pigment is taken out from the identification results in each characteristic sub-interval, and the K / S curves of the standard pigments taken out from different characteristic sub-intervals are combined into a set of sub-interval identification results. Then, the abundance matrix obtained according to the Dirichlet distribution function is combined to generate 1000 simulated mixed K / S curves with different abundances. The similarity between the simulated mixed K / S curves and the K / S curve of the unknown pigment is calculated again, and the highest value is reserved in each set of sub-interval identification results. The standard pigment in the set of sub-interval identification results corresponding to the highest value is the final identification result of the spectrum of the unknown pigment. The sub-interval identification results of the six groups of mixed pigments are shown in Table 2, in which the bold pigments are the identification results consistent with the true values of the mixed pigments.
[0102] Mixed pigments Subinterval 1 identification Subinterval 2 identification Azurite malachite Azurite, green agate, white feldspar Malachite, azurite, silver powder Azurite topaz Topaz, saffron, pearl powder Malachite, azurite, two green Azurite cinnabar Azurite, azurite powder, azurite Carmine, cinnabar, rouge Malachite topaz Topaz, saffron, saffron powder Two green, head green, precious malachite Malachite cinnabar Malachite, gauze powder, crystal powder Cinnabar, red coral, carmine Topaz cinnabar Saffron, topaz, saffron powder Rinsed cinnabar, cinnabar, cinnabar powder
[0103] According to the results, it is found that the identification results of the mixed pigment samples are good, and the overall identification rate is 83.3%. The results show that the mixed pigments can be identified by the present application, which has practical significance for the analysis of cultural relics pigments.
[0104] It can be seen from the above examples that the spectral curves of different color pigments have steep and straight rising edges at different positions, and the positions of the rising edges of the spectral curves of the same color pigments have slight differences, and therefore, the component recognition can be realized by analyzing the positions of the rising edges of the spectral curves. It is found through experiments that the "protrusions" of the derivative curves correspond to the positions of the rising edges of the spectral curves after the first derivative of the spectral curves is calculated, and it is found through analyzing the derivative curves of the mixed pigments that the "protrusions" corresponding to the rising edges of the component pigments on the derivative curves are consistent with the additivity.
[0105] After the characteristic intervals of the unknown pigments are extracted, the unknown pigments are recognized in the interval range, and it is found through comparing the recognition results of the intervals with the true values that the recognition results of each interval can accurately recognize the component pigments of the mixed pigments, proving that the characteristic intervals effectively contain the characteristics of the component pigments in the mixed pigments.
[0106] It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the action sequence described, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to optional embodiments, and the actions and modules involved are not necessarily necessary for the present application.
[0107] The above is the introduction of the method embodiment, and the scheme described in the present application will be further described through the device embodiment.
[0108] As shown in Figure 4 , the device 400 comprises:
[0109] The acquisition correction module 410 is configured to acquire the hyperspectral image of the target object to be recognized, the hyperspectral image of the standard reflector plate and the dark current data, and perform reflectivity correction;
[0110] The noise reduction conversion module 420 is configured to perform data noise reduction processing on the reflectivity corrected image, and convert the unknown pigment spectrum and the standard pigment spectrum after the noise reduction processing into corresponding K / S curves;
[0111] The calculation extraction module 430 is configured to perform first derivative calculation on the unknown pigment spectrum to obtain the derivative curve of the unknown pigment spectrum, and extract the characteristic sub-interval of the K / S curve corresponding to the unknown pigment spectrum from the derivative curve of the unknown pigment spectrum;
[0112] The judgment identification module 440 is configured to: if the number of the characteristic subintervals of the K / S curve corresponding to the unknown pigment spectrum is 1, calculate the similarity of the K / S curves corresponding to the unknown pigment and the standard pigments in the full wave band range, and the standard pigment with the highest similarity is taken as the identification result; otherwise, calculate the similarity of the K / S curves corresponding to the unknown pigment and the standard pigments in each characteristic subinterval, obtain a plurality of interval identification result sets, generate a K / S curve corresponding to a mixed pigment, and take the standard pigment with the highest similarity of the K / S curves corresponding to the mixed pigment and the standard pigments as the identification result of the unknown pigment.
[0113] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the described modules can refer to the corresponding process in the foregoing method embodiments, which will not be described here.
[0114] In the technical solution of the application, the acquisition, storage and application of user personal information comply with relevant laws and regulations and do not violate public order and good customs.
[0115] According to the embodiments of the application, the application further provides an electronic device and a readable storage medium.
[0116] Figure 5 A schematic block diagram of an electronic device 500 that can be used to implement embodiments of the application is shown. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the application described and / or claimed in this document.
[0117] The device 500 includes a computing unit 501 that can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded into a random access memory (RAM) 503 from a storage unit 508. Various programs and data required for the operation of the device 500 can also be stored in the RAM 503. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.
[0118] A plurality of components in the device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a magnetic disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.
[0119] The computing unit 501 can be various general and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 501 performs various methods and processes described above, such as the methods S101-S104. For example, in some embodiments, the methods S101-S104 can be implemented as a computer software program, which is tangibly embodied in a machine-readable medium, such as the storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded to the RAM 503 and executed by the computing unit 501, one or more steps of the methods S101-S104 described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the methods S101-S104 by any other appropriate means, such as by means of firmware.
[0120] The various implementations of the systems and techniques described above herein can be realized in a digital electronic circuit system, an integrated circuit system, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), an application specific standard product (ASSP), a system on a chip system (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.
[0121] Program code for carrying out operations of the methods of the present application can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces the functions / operations specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a stand-alone software package, and partially on a remote machine or entirely on a remote machine or server.
[0122] In the context of the present application, a machine-readable medium can be a tangible medium that can contain or store program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable storage medium can include, without limitation, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the foregoing.
[0123] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.
[0124] The systems and techniques described herein can be implemented in a computing system that includes a back end component, e.g., as a data server, or that includes a middleware component, e.g., an application server, or that includes a front end component, e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described herein, or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communication network. Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.
[0125] The computer system can include clients and servers. The clients and the servers are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, a server of a distributed system, or a server combined with a blockchain.
[0126] It should be understood that the steps shown in the various forms above can be reordered, added to, or removed. For example, the steps described in the present disclosure can be performed in parallel, in series, or in a different order, without limitation, as long as the desired results of the present disclosure are achieved.
[0127] The specific embodiments described above are not intended to limit the scope of the present application. Those skilled in the art will understand that various modifications, combinations, sub-combinations, and alternatives can be made to the specific embodiments without departing from the spirit and principles of the present application. Any further modifications, equivalents, and / or alternatives come within the scope of the present application as recited by the claims.
Claims
1. A mixed pigment spectrum partition interval identification method, characterized in that, The application relates to a method for identifying unknown pigments, and comprises the following steps: Collecting hyperspectral images of target objects to be identified, hyperspectral images of standard reflectors and dark current data, and performing reflectivity correction; Performing data noise reduction processing on the images after reflectivity correction, converting unknown pigment spectra and standard pigment spectra after noise reduction processing into corresponding K / S curves; First derivative calculation is performed on the unknown pigment spectrum to obtain a derivative curve of the unknown pigment spectrum, and a characteristic subinterval of the K / S curve corresponding to the unknown pigment spectrum is extracted from the derivative curve of the unknown pigment spectrum; If the number of the characteristic subintervals of the K / S curve corresponding to the unknown pigment spectrum is 1, the similarity of the K / S curves corresponding to the unknown pigment and the standard pigment is calculated in the full waveband range, and the standard pigment with the highest similarity is taken as the identification result; otherwise, the similarity of the K / S curves corresponding to the unknown pigment and the standard pigment is calculated in each characteristic subinterval, a plurality of interval identification result sets are obtained, a K / S curve corresponding to a mixed pigment is generated, and the standard pigment with the highest similarity of the K / S curves corresponding to the mixed pigment and the standard pigment is taken as the identification result of the unknown pigment; The characteristic subinterval of the K / S curve corresponding to the unknown pigment spectrum is extracted from the derivative curve of the unknown pigment spectrum, and comprises the following steps: The difference between the adjacent two wave troughs in the derivative curve of the unknown pigment spectrum is taken as a section, and the larger value of the difference between the wave peak and the two wave troughs in each section is taken as the peak-valley difference value of the current section; The peak-valley difference values of each section are sorted, the multiple relationship between the maximum value of the peak-valley difference value and the second maximum value of the peak-valley difference value is judged, if the multiple relationship is greater than a multiple threshold value, only the section corresponding to the maximum value of the peak-valley difference value is reserved as the spectral range of the characteristic subinterval of the K / S curve; otherwise, the two sections corresponding to the maximum value of the peak-valley difference value and the second maximum value of the peak-valley difference value are reserved as the spectral range of the characteristic subinterval of the K / S curve; The characteristic subinterval of the K / S curve of the unknown pigment is extracted according to the spectral range of the characteristic subinterval; The K / S curves of the two standard pigments in each interval identification result set are taken as end members to generate a plurality of simulated mixed K / S curves with different abundances; The similarity of the simulated mixed K / S curve and the K / S curve corresponding to the standard pigment is calculated by using the spectral angle cosine combined with the normalized Euclidean distance algorithm, and the highest similarity value of each interval identification result set is obtained; The highest similarity values of the interval identification result sets are compared, and the standard pigment in the interval identification result set corresponding to the highest similarity is taken as the component of the mixed pigment. The reflectivity correction comprises the following steps:
2. The method of claim 1, wherein, Calculating standard reflector correction data according to the hyperspectral images of the standard reflectors and the dark current data; Expanding the dark current data into dark current correction data with the same row and column number as the hyperspectral images of the target objects; Recomputing target object correction data, and calculating the hyperspectral reflectivity images of the target objects according to the target object correction data. The data noise reduction processing on the images after reflectivity correction comprises the following steps:
3. The method of claim 1, wherein, Performing MNF forward transformation on the reflectance corrected image to output first transformed data; Performing MNF inverse transformation on the first transformed data to obtain second transformed data.
4. The method of claim 1, wherein, The similarity of the unknown pigment and the standard pigment corresponding K / S curve is calculated in each characteristic sub-interval to obtain a plurality of interval identification result sets, including: In each characteristic sub-interval range, the similarity of the unknown pigment and the standard pigment corresponding K / S curve is calculated using spectral angle cosine combined with normalized Euclidean distance algorithm, and the N standard pigments with the highest similarity are taken as the identification result of the characteristic sub-interval; From the identification result of each characteristic sub-interval, a standard pigment corresponding K / S curve is sequentially taken out, and is combined in pairs to obtain N^2 interval identification result sets.
5. A mixed pigment spectrum partition interval identification device, characterized by, It includes: The acquisition correction module is used for acquiring the hyperspectral image of the target object to be identified, the hyperspectral image of the standard reflector and the dark current data, and performing reflectance correction; The noise reduction conversion module is used for performing data noise reduction processing on the reflectance corrected image, and converting the unknown pigment spectrum and the standard pigment spectrum after noise reduction processing into corresponding K / S curves; The calculation and extraction module is used for performing first derivative calculation on the unknown pigment spectrum to obtain the derivative curve of the unknown pigment spectrum, and extracting the characteristic sub-interval of the unknown pigment spectrum corresponding K / S curve in the derivative curve of the unknown pigment spectrum; The judgment and identification module is used for judging whether the number of characteristic sub-intervals of the unknown pigment spectrum corresponding K / S curve is 1, if yes, calculating the similarity of the unknown pigment and the standard pigment corresponding K / S curve in the full waveband range, and taking the standard pigment with the highest similarity as the identification result; otherwise, calculating the similarity of the unknown pigment and the standard pigment corresponding K / S curve in each characteristic sub-interval to obtain a plurality of interval identification result sets, generating a mixed pigment corresponding K / S curve, and taking the standard pigment with the highest similarity of the mixed pigment and the standard pigment corresponding K / S curve as the identification result of the unknown pigment; The characteristic sub-interval of the unknown pigment spectrum corresponding K / S curve is extracted in the derivative curve of the unknown pigment spectrum, including: The difference between the adjacent two valleys in the derivative curve of the unknown pigment spectrum is taken as a section, and the larger value of the difference between the peak and the two valleys in each section is taken as the peak-valley difference value of the current section; The peak-valley difference values of each section are sorted, and it is judged whether the multiple relationship between the maximum value of the peak-valley difference value and the second largest value of the peak-valley difference value is greater than a multiple threshold value, if yes, only the section corresponding to the maximum value of the peak-valley difference value is reserved as the spectral range of the characteristic sub-interval of the K / S curve; otherwise, the two sections corresponding to the maximum value and the second largest value of the peak-valley difference value are reserved as the spectral range of the characteristic sub-interval of the K / S curve; The characteristic sub-interval of the unknown pigment K / S curve is extracted according to the spectral range of the characteristic sub-interval; The mixed pigment corresponding K / S curve is generated, and the standard pigment with the highest similarity of the mixed pigment and the standard pigment corresponding K / S curve is taken as the identification result of the unknown pigment, including: two standard pigment K / S curves in each interval identification result set are taken as end members to generate a plurality of simulated mixed K / S curves with different abundances; a spectral angle cosine is combined with a normalized Euclidean distance algorithm to calculate the similarity of the simulated mixed K / S curve and the standard pigment corresponding K / S curve, to obtain the highest similarity value of each interval identification result set; the highest similarity value of each interval identification result set is compared, and the standard pigment in the interval identification result set corresponding to the highest similarity is taken as a component of the mixed pigment. 6.An electronic device comprising at least one processor; and a memory connected in communication with the at least one processor; characterized in that, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-4.
7. A non-transitory computer-readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to enable the computer to perform the method according to any one of claims 1-4.