CN114279976A - Mural soluble salt content detection method based on reflection spectrum - Google Patents
Mural soluble salt content detection method based on reflection spectrum Download PDFInfo
- Publication number
- CN114279976A CN114279976A CN202111614138.7A CN202111614138A CN114279976A CN 114279976 A CN114279976 A CN 114279976A CN 202111614138 A CN202111614138 A CN 202111614138A CN 114279976 A CN114279976 A CN 114279976A
- Authority
- CN
- China
- Prior art keywords
- mural
- spectrum
- standard
- soluble salt
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a mural soluble salt content detection method based on a reflection spectrum, which comprises the following steps: making a plurality of standard mural sample blocks containing soluble salts with different mass concentrations; at preset time intervals, sampling each standard mural sample block for at least 1 time by using a spectral radiometer to obtain an original reflection spectrum of the standard mural sample block; carrying out reciprocal logarithmic transformation on the original reflection spectrum to obtain a reciprocal logarithmic spectrum; extracting a characteristic wave band; taking the mass concentration of soluble salt of a standard mural sample block as a dependent variable, taking the parameter value of a reciprocal logarithmic spectrum at a characteristic wave band as an independent variable, and establishing a mural sample soluble salt content prediction model; and (3) acquiring a punctiform spectrum curve of the designated position of the mural to be detected by using a spectrum radiometer, and predicting the mass concentration of soluble salt by using a soluble salt content prediction model. The method has the beneficial effects of carrying out nondestructive, efficient and real-time detection on the content of soluble salt in any area of the mural.
Description
Technical Field
The invention relates to the technical field of mural detection. More specifically, the invention relates to a mural soluble salt content detection method based on a reflection spectrum.
Background
The mural is a color painting attached to ancient buildings, and is also a carrier reflecting human social life, religious belief and historical culture. However, the wall painting is degraded due to natural erosion and is interfered by people, so the preservation status is always worried. The interaction between the mural and soluble salt in the environment is obvious, the salt can be continuously dissolved, crystallized and expanded along with the temperature, once the content is accumulated to a certain concentration, the salt can move to the surface layer of the mural through capillary water to generate the phenomena of enrichment and crystallization, and therefore the mural is induced to generate various irreversible mural diseases such as the shortbread herpes and the like. Once the salt damage of the mural is formed, the artistic value of the mural is irreversibly and seriously weakened. Therefore, the nondestructive detection method has very important and urgent practical significance for carrying out efficient, real-time and high-precision nondestructive detection on the content of soluble salt in the mural.
At present, the detection of soluble salt in mural painting mainly comprises the following two methods, the first method is to use an ion chromatographic analyzer to analyze the type and content of salt in mural painting ground layer; secondly, rapidly detecting and analyzing the mural salt damage by utilizing capillary electrophoresis; the two methods need to carry out on-site sampling on the position to be detected (the ground layer disease) in the full-picture mural, so that the salt content detection is difficult to carry out on any specified area of the full-picture mural, the integrity of the mural can be damaged, meanwhile, the detection salt ion concentration range is limited by a peak value, the precision and the accuracy can be reduced along with the sampling quality, and the salt content dynamic monitoring cannot be carried out.
Disclosure of Invention
An object of the present invention is to solve at least the above problems and to provide at least the advantages described later.
The invention also aims to provide a mural soluble salt content detection method based on the reflection spectrum, which realizes nondestructive, efficient and real-time detection of soluble salt content in any area of the mural.
To achieve these objects and other advantages in accordance with the purpose of the invention, there is provided a mural soluble salt content detection method based on reflection spectrum, comprising the steps of: making a plurality of standard mural sample blocks containing soluble salts with different mass concentrations;
sampling each standard mural sample block for at least 1 time by using a spectral radiometer at preset time intervals, wherein a pretreatment spectrum of the standard mural sample block in the wavelength range of 350-;
carrying out reciprocal logarithmic transformation on the preprocessed spectrum to obtain a reciprocal logarithmic spectrum;
extracting a characteristic wave band according to the reciprocal logarithmic spectrum;
taking the mass concentration of soluble salt of a standard mural sample block as a dependent variable, taking the parameter value of a reciprocal logarithmic spectrum at a characteristic wave band as an independent variable, performing linear regression to obtain an optimal fitting line, and establishing a mural sample soluble salt content prediction model;
and (3) acquiring a point-like spectrum curve of the designated position of the mural to be detected by using a spectrum radiometer, preprocessing, carrying out reciprocal logarithmic transformation, acquiring characteristic spectrum parameters at a characteristic wave band, bringing the characteristic spectrum parameters into a mural sample soluble salt content prediction model, and predicting the mass concentration of soluble salt.
Preferably, the method for making the standard mural sample block comprises the following steps:
manufacturing a coarse mud layer;
and manufacturing a fine mud layer on the surface of the coarse mud layer, and naturally drying in the shade to obtain a plurality of standard mural sample blocks, wherein soluble salt with corresponding mass concentration is uniformly mixed in the fine mud layer of each standard mural sample block.
Preferably, the mass concentration range of the soluble salt formed by the plurality of standard mural sample blocks is 0-1%, wherein the mass concentration difference of the two adjacent standard mural sample blocks with the mass concentration is not more than 0.2% according to the mass concentration of the soluble salt in the standard mural sample blocks from small to large.
Preferably, the mass concentrations of soluble salts in the standard mural sample blocks are arranged from small to large, and the mass concentration difference of any two adjacent standard mural sample blocks with the mass concentrations is equal.
Preferably, the soluble salt is one of sulfate, nitrate and chloride.
Preferably, the soluble salt is sodium sulfate.
Preferably, each sampling is specifically:
dividing each standard mural sample block into 3 x 3 blocks, and randomly selecting a central point of one block for data acquisition in each row and each column, wherein the method for acquiring data from each central point comprises the following steps:
s1, collecting an initial reflection spectrum curve within the wavelength range of 350-2500nm by aligning a probe of the spectrum radiometer with a standard mural sample block and a central point;
s2, horizontally rotating the probe of the spectrum radiometer by 90 degrees for next acquisition;
and S3, repeating the step S2 until the probe of the spectrum radiation instrument horizontally rotates 270 degrees.
Preferably, the pre-processing after each sampling comprises the following steps:
sa, taking the arithmetic mean value of the breakpoint correction spectrum curves of the same central point as the spectrum curve of the central point;
and Sb, taking the arithmetic mean value of the average spectrum curves corresponding to the same standard mural sample block.
Preferably, the preprocessing after each sampling further includes performing breakpoint correction on the initial reflection spectrum curve before performing the step Sa to obtain a breakpoint correction spectrum curve;
after the step Sb is carried out, Savitzky-Golay smoothing processing is carried out on the spectrum curve acquired by each time of the standard mural sample block, and a reflection spectrum of the standard mural sample block in the wavelength range of 350-2500nm is obtained.
Preferably, the method of extracting the characteristic band from the reciprocal logarithmic spectrum includes the steps of:
selecting a wave band range with a spectrum showing obvious peaks and valleys and a spectrum curve showing stable monotonous change;
for the determined waveband range, after the pretreatment spectrum and the salt concentration are subjected to a p-0.05 Pearson correlation significance detection, the characteristic waveband of the reciprocal logarithmic spectrum is determined.
The invention at least comprises the following beneficial effects:
utilize spectral technique, acquire mural different regions and salt damage degree department punctiform reflectance spectrum, strengthen through spectral feature, establish salt ion concentration and spectral feature parameter's fitting regression equation, construct the prediction inversion system of mural soluble salt content, realized carrying out harmless, high efficiency, real-time detection to arbitrary regional soluble salt content in the mural, and have better detection precision, in the mural salt damage degree appraisal and the mural permanent save field, great economic value and practical meaning have, it is specific:
firstly, a hyperspectral technology is adopted to collect dot spectra of the murals, and the wavebands of the murals are wide in response range and sensitive to the spectrum superposed on surface substances, so that the requirement on refined murals salt detection can be met, and the problems that the traditional murals soluble salt detection mode is single and secondary damage can be caused to the murals are solved;
secondly, selecting mass concentration of soluble salt of the mural as a salt content inversion index in constructing standard mural sample blocks, wherein the mass concentration of the soluble salt of the mural is more practical, and in the manufacturing process, firstly manufacturing a coarse mud layer, continuously manufacturing a fine mud layer on the surface of the coarse mud layer, then putting the coarse mud layer into a mold, and placing the mold in a room for natural drying in the shade, wherein soluble salt with corresponding mass concentration is uniformly mixed in the fine mud layer of each standard mural sample block, so that the manufactured sample is closer to the ground layer of the actual mural;
thirdly, in the data acquisition process, a Sudoku method is adopted for each standard mural sample block, a central point is determined firstly, each central point is subjected to direction change measurement for four times, and in the data preprocessing process, preprocessing is carried out in a combined mode of breakpoint elimination, spectrum averaging and SG smoothing, so that noise generated by instrument elements, manual operation and the like in the acquisition process is effectively eliminated;
fifthly, the spectrum characteristics are enhanced through logarithmic transformation of reciprocal, so that the influence of illumination conditions and terrain differences on the spectrum can be reduced, the spectrum differences in a visible light range can be improved and compared, and random factor errors are reduced; specifically, the optimal characteristic band of sodium sulfate salt is determined to be 1415nm, and a mural sample sodium sulfate content prediction model is established (correlation coefficient R is 0.858, determination coefficient R2 is 0.737, and root mean square error RMSE is 0.121).
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention.
Drawings
FIG. 1 is a flow chart of a method for detecting soluble salt content of murals based on reflection spectrum according to one embodiment of the present invention;
FIG. 2 is a diagram showing a standard mural sample block Sudoku point selection rule according to one embodiment of the present invention;
FIG. 3 is a general graph of a pre-processing spectral curve according to one embodiment of the present invention;
FIG. 4 is a logarithmic spectrum of the reciprocal of one embodiment of the present invention;
FIG. 5 is a graph of the reflectance of the reciprocal log spectrum in the range of 1000-1800nm according to one embodiment of the present invention;
FIG. 6 is a line of best fit of the reciprocal log spectrum at 1415nm according to one embodiment of the present invention;
FIG. 7 is a reflectance curve of the pretreatment spectrum in the range of 500-1500nm according to one embodiment of the present invention;
FIG. 8 is a line of best fit of the preprocessed spectra at 800nm according to one embodiment of the present invention;
FIG. 9 is a best fit line at 1413nm for the preprocessed spectra according to one embodiment of the present invention;
FIG. 10 is a spectrum of envelope elimination according to one embodiment of the present invention;
FIG. 11 is a reflectance curve of the envelope elimination spectrum in the range of 350-600nm according to one embodiment of the present invention;
FIG. 12 is a best fit line of the envelope elimination spectrum at 420nm according to one embodiment of the present invention;
FIG. 13 is a diagram of the prediction result of a sodium sulfate content prediction model (LR) of a mural sample according to one embodiment of the present invention;
FIG. 14 is a diagram of the prediction result of the sodium sulfate content prediction model (R-1) of the mural sample according to one embodiment of the present invention;
FIG. 15 is a diagram of the prediction result of the sodium sulfate content prediction model (R-2) of the mural sample according to one embodiment of the present invention;
fig. 16 is a diagram of a prediction result of a sodium sulfate content prediction model (CR) of a mural sample according to one embodiment of the present invention.
Detailed Description
The present invention is further described in detail below with reference to examples so that those skilled in the art can practice the invention with reference to the description.
< example 1>
As shown in figure 1, the mural soluble salt content detection method based on the reflection spectrum is characterized by comprising the following steps:
step one, manufacturing a standard mural sample block
Making a plurality of standard mural sample blocks containing soluble salts with different mass concentrations, specifically comprising: the standard mural sample block (1 block) with the mass concentration of 0%, the standard mural sample block (1 block) with the mass concentration of 0.2%, the standard mural sample block (1 block) with the mass concentration of 0.3%, the standard mural sample block (1 block) with the mass concentration of 0.4%, the standard mural sample block (1 block) with the mass concentration of 0.5%, the standard mural sample block (2 blocks) with the mass concentration of 0.6%, the standard mural sample block (1 block) with the mass concentration of 0.8%, and the standard mural sample block (1 block) with the mass concentration of 1%;
wherein each standard mural sample block is made in a rectangular wooden mold (16cm by 11cm by 1.8cm), and the making method comprises the following steps
Making a coarse mud layer, uniformly mixing sand (coarse) and coarse loess, fully grinding, removing coarse stones by using a linen, adding wheat straws and deionized water, uniformly mixing, wrapping for 24 hours by using a preservative film, standing, and uniformly spreading into the film;
fine mud layers are manufactured on the surfaces of the coarse mud layers, soluble salts with corresponding mass concentrations are uniformly mixed in the fine mud layers of each standard mural sample block, wherein the fine mud layers are specifically manufactured as follows: manually shredding the purchased ancient building hemp knives, and removing knotted hemp balls and untwisted hemp ropes. Mixing the fine silt with raw soil (fine) and clarified board soil (fine), adding deionized water and a soluble salt solution, uniformly mixing, wrapping with a preservative film for 24 hours, standing up, and uniformly paving the surface of the coarse silt layer in the film to form a fine silt layer, wherein the clarified board soil is riverbed deposited soil with the particle size of 0.001-0.005 mm;
naturally drying in the shade indoors, and culturing and monitoring the temperature and the humidity of the sample by using a soil detector in the natural drying process to ensure that the culture conditions of the standard mural sample blocks are kept consistent;
step two, measuring the spectral data of the standard mural sample block
A spectral radiometer is selected for measuring the spectral reflectivity of a standard mural sample block;
preheating the instrument for 30 minutes before using, carrying out an experiment in a dark room, firstly correcting a white board, and padding black lint under a standard mural sample block after stabilization;
respectively sampling standard mural sample blocks with mass concentrations of 0%, 0.2%, 0.4%, 0.6% (one of them), 0.8% and 1% for 4 times, wherein the sampling interval between any two adjacent sampling times of 4 times is preset time (24 hours);
respectively sampling standard mural sample blocks with mass concentrations of 0.3%, 0.5% and 0.6% (the rest blocks) for 1 time;
wherein, sampling each time is specifically as follows:
dividing the corresponding standard mural sample block into 3 × 3 blocks (as shown in fig. 2, equally dividing, 3 rows and 3 columns in total), randomly selecting a central point of one block for data acquisition in each row and each column, that is, selecting 3 central points for each standard mural sample block, wherein the rows and columns of the 3 central points are different, and further, the method for data acquisition of each central point comprises the following steps:
s1, collecting an initial reflection spectrum curve by aligning a probe of the spectrum radiometer with a central point and being vertical to the standard mural sample block;
s2, horizontally rotating the probe of the spectrum radiometer by 90 degrees for next acquisition, and acquiring another initial reflection spectrum curve;
and S3, repeating the step S2 until the probe of the spectrum radiation instrument horizontally rotates 270 degrees, namely, each central point obtains 4 corresponding initial reflection spectrum curves.
During collection, the parameters of the spectral radiometer are set as follows:
| parameter name | Specific numerical value |
| Spectral range | 350-2500nm |
| Number of channels | 2151 |
| Spectral width | 2.5nm |
| Spectral resolution | 50-1000nm@3nm；1001-2500nm@8nm |
In conclusion, each acquisition of each standard mural sample block obtains 12 initial reflection spectrum curves, and the acquisition is carried out for 27 times in total to obtain 324 initial reflection spectrum curves;
thirdly, preprocessing the spectrum data to obtain a preprocessed spectrum (R);
firstly, performing breakpoint Correction (Splice Correction) on an initial reflection spectrum curve by using View Spec Pro to obtain a breakpoint Correction spectrum curve;
after eliminating abnormal values caused by point position edges and measurement jitter in each breakpoint correction spectrum curve, performing double-averaging on the initial reflection spectrum curve acquired each time by each standard mural sample block to obtain an actually-measured reflection spectrum curve of the standard mural sample block with corresponding concentration acquired this time, wherein the double-averaging specifically comprises the following steps:
sa, taking the arithmetic mean value of the breakpoint correction spectrum curves of the same central point as the spectrum curve of the central point;
sb, taking the arithmetic mean value of the average spectrum curves corresponding to the same standard mural sample block;
carrying out SG Smoothing (Savitzky Golay Smoothing) treatment on the actually measured reflection spectrum curve, wherein the size of an SG Smoothing window is 21, and obtaining a reflection spectrum (preprocessing spectrum) of a standard mural sample block in the wavelength range of 350-2500 nm;
in summary, the preprocessing yielded acceptable 27 preprocessed spectra for the standard mural sample block (as shown in fig. 3);
step three, enhancing spectral characteristics
Performing LR conversion on the preprocessed spectrum (taking the logarithm of the reflectance value after reciprocal operation) to obtain a reciprocal logarithmic spectrum (LR) (as shown in fig. 4, in which, for convenience of observation, only data obtained by sampling 4 th time of a standard mural sample block with mass concentration of 0%, 0.2%, 0.4%, 0.6% (one of them), 0.8%, 1% is shown in fig. 3, and the same is true for fig. 3, 5, 7, 10, 11;
step four, extracting characteristic wave bands
Based on the reciprocal logarithmic spectrum, the characteristic band range is determined following two principles:
firstly, the spectrum changes violently and presents distinct peaks and valleys, and as can be seen from fig. 4, the two wavelength ranges at which the distinct peaks and valleys appear are respectively about 1415nm and 1941 nm;
secondly, selecting a wave band with a spectrum curve showing stable monotonous change in a section of wave band range, wherein the absorption valley reflectivity of the sample spectrum at 1941nm is greatly changed along with the salt concentration because the water conventional absorption wave band at about 1950nm does not show stable monotonous change;
in conclusion, the characteristic wave band range is determined to be 1410nm-1420 nm;
for the characteristic wave band range, after p-0.05 Pearson correlation significance detection is carried out on the logarithmic spectrum of the reciprocal and the mass concentration of sodium sulfate in SPSS, the characteristic wave band of the logarithmic spectrum of the reciprocal is determined (as shown in FIG. 5, the characteristic wave band is 1415nm, and the observation shows that the 1415nm wave band has an obvious reflection peak and the fluctuation is small);
step five, establishing a prediction model of the sodium sulfate content of the mural sample
Taking the concentration of soluble salt in a standard mural sample block as a dependent variable, taking the parameter value of a reciprocal logarithmic spectrum at a characteristic wave band as an independent variable, performing linear regression to obtain a best fit line (as shown in figure 6), and establishing a mural sample sodium sulfate content prediction model (LR);
mural sample sodium sulfate content prediction model (LR): -3.3906R1415nm+2.9633
Wherein Y is the content (mass concentration) of sodium sulfate salt in the sample, and R is1415nmIs the logarithmic value of the reciprocal of the spectral reflectance at 1415 nm;
step six: acquiring a punctiform spectrum curve of the designated position of the mural to be detected by using a spectrum radiometer;
and carrying out reciprocal logarithmic transformation after pretreatment to obtain characteristic spectrum parameters at a characteristic wave band, and then bringing the parameters into a mural sample sodium sulfate content prediction model to predict the mass concentration of the sodium sulfate salt.
< comparative example 1>
The mural soluble salt content detection method based on the reflection spectrum is characterized by comprising the following steps:
step one, manufacturing a standard mural sample block, which is the same as the step one in the embodiment 1;
secondly, preprocessing the spectral data, and obtaining a preprocessed spectral curve in the same way as the second step in the embodiment 2;
thirdly, preprocessing the spectrum without enhancing the spectral characteristics;
step four, extracting characteristic wave bands
Based on the preprocessed spectra (fig. 3), the characteristic bands are determined following two principles:
firstly, the spectrum changes violently and presents an inflection point at a distinct peak and valley, and as can be seen from fig. 7, absorption valleys are clearly present at wavebands of 367nm, 1413nm, 1941nm and 2211 nm;
secondly, selecting a wave band in which a spectrum curve is in stable monotonicity change in a section of wave band range, wherein the edge wave band at 367nm has larger noise influence, and the reason that the change of the reflectivity of an absorption valley at 1941nm along with the salt concentration is larger is that the wave band is a conventional absorption wave band of water at about 1950nm, the peak valley at 2211nm is frequently replaced, the signal-to-noise ratio is overlarge, namely, the 3 wave bands do not show stable monotonicity change;
in conclusion, the characteristic wave band range is determined to be 1410nm-1420 nm;
further, the spectral curve is observed to show that the spectral reflectivity is in an ascending trend within a wave band of 350nm to 1400nm, wherein the slope is increased fastest at 800nm of 600 + and the reflectivity is increased well at 800nm, no obvious rate difference exists between the stable acceleration and different salt concentrations, then the slope increasing speed is gradually decreased, and the characteristic wave band range is determined to be 795nm to 805nm in order to compare the fitting effect of different wave bands under the same model;
for the characteristic wave band range, after the preprocessing spectrum and the salt concentration are subjected to p-0.05 Pearson correlation significant detection in the SPSS, the characteristic wave band of the reciprocal logarithmic spectrum is determined (as shown in FIG. 7, the characteristic wave bands are 800nm and 1413nm respectively, and it can be known from observation that an obvious absorption valley appears at the 1413nm wave band, the reflectivity potential increase at 800nm is obvious, the reflectivity is in a positive correlation monotonic relation along with the increase of the salt content, and the intervals between the graduation levels are relatively uniform);
step five, establishing a prediction model of the sodium sulfate content of the mural sample
Taking the concentration of soluble salt in a standard mural sample block as a dependent variable, taking parameter values of a pretreatment spectrum at a characteristic wave band as independent variables, performing linear regression to obtain an optimal fit line (shown in figures 8 and 9), and establishing a mural sample sodium sulfate content prediction model, wherein the mural sample sodium sulfate content prediction model (R-1) is established for the characteristic wave band of 800nm, and the mural sample sodium sulfate content prediction model (R-2) is established for the characteristic wave band of 1413 nm;
prediction model of sodium sulfate content in mural sample (R-1): 5.4263R800nm-1.8904
Wherein Y is the content of sodium sulfate salt in the sample, R800nmIs the spectral reflectance at 800 nm;
prediction model of sodium sulfate content of mural sample (R-2): 6.7574R1413nm-2.7787
Wherein Y is the content of sodium sulfate salt in the sample, R1413nmIs the spectral reflectance at 1413 nm.
< comparative example 2>
The mural soluble salt content detection method based on the reflection spectrum is characterized by comprising the following steps:
step one, manufacturing a standard mural sample block, which is the same as the step one in the embodiment 1;
step two, pretreatment of spectral data, the same as step two of example 2;
step three, enhancing spectral characteristics
Performing envelope elimination (normalizing the spectrum reflectivity) on the preprocessed spectrum to obtain an envelope elimination spectrum (CR), as shown in fig. 10;
step four, extracting characteristic wave bands
Based on the envelope elimination spectrum, the characteristic band range is determined according to the following two principles:
firstly, the spectrum changes violently and presents an inflection point at an obvious peak-valley position, and as can be seen from fig. 10, the wave band ranges at which the obvious peak-valley position appears obviously present absorption valleys at the wave bands of 420nm, 1413nm, 1941nm and 2211nm, respectively;
secondly, selecting a wave band with a spectrum curve showing stable monotonous change in a section of wave band range, wherein the stable monotonous change is not shown at 1413nm, 1941nm and 2211 nm;
in conclusion, the characteristic wave band range is determined to be 415nm-425 nm;
for the characteristic wavelength range, after p-0.05 pearson correlation significant detection is performed on the preprocessed spectrum and the salt concentration in the SPSS, the characteristic wavelength range of the reciprocal logarithmic spectrum is determined (as shown in fig. 11, the characteristic wavelength ranges are respectively 420nm, and it can be seen that a characteristic absorption valley with strong correlation appears at the 420nm wavelength range, and the characteristic absorption valley is similar to a steep threshold);
step five, establishing a prediction model of the sodium sulfate content of the mural sample
Taking the concentration of soluble salt in a standard mural sample block as a dependent variable, taking the parameter value of an envelope removal spectrum at a characteristic wave band as an independent variable, performing linear regression to obtain a best fit line (as shown in figure 12), and establishing a mural sample sodium sulfate content prediction model (CR);
fresco sample sodium sulfate content prediction model (CR): -4.521R420nm+1.3737
Wherein Y is the content of sodium sulfate salt in the sample, R420nmThe spectral reflectance dip depth after envelope removal at 420 nm.
1. Standard mural sample block fabrication
In the construction of standard mural sample blocks, the mass concentration of soluble mural salt is selected as a salt content inversion index, so that the method is more practical, in the manufacturing process, a coarse mud layer is manufactured firstly, a fine mud layer is manufactured on the surface of the coarse mud layer continuously, then the coarse mud layer is placed in a mold and placed indoors for natural drying in the shade, wherein the soluble salt with the corresponding mass concentration is uniformly mixed in the fine mud layer of each standard mural sample block, and the sample manufacturing quality is improved.
2. Relating to spectral data acquisition and pre-processing
The method is characterized in that a Sudoku method is adopted for each standard mural sample block, a central point is determined firstly, direction change measurement is performed for four times for each central point, preprocessing is performed in a combined mode of breakpoint elimination, spectrum averaging and SG smoothing in a data preprocessing process, noise generated due to instrument elements, manual operation and the like in an acquisition process is effectively eliminated through an obtained spectrum curve, the spectrum curve changes regularly along with salt concentration, and peak-valley band characteristics are enhanced.
3. Enhancement of spectral features
3.1, analysis of the pretreated spectral curve (FIG. 2) reveals that:
the reflectivity of the mural sample and the salt content are in a positive correlation trend overall, the spectral reflectivity of the mural sample is in a gradually increasing trend along with the increase of the salt concentration of the sample block, the reflectivity of 1% and 0.8% of the salt content is obviously higher than that of the rest groups, a generation-breaking interval is formed transversely, and the primary analysis is caused by white frost salt spots formed on the surface of the mural sample.
Absorption valleys are obviously generated at wavebands of 367nm, 1413nm, 1941nm and 2211nm, a reflection peak is obviously generated at 2136nm, and the maximum value of the reflectivity is 0.5821, wherein the reason that the reflectivity of the absorption valleys of the sample spectrum at 1941nm greatly changes along with the salt concentration is that the waveband is the conventional absorption waveband of water at 1950nm or so.
Within the wave band of 350nm-1400nm, the spectral reflectivity is in an ascending trend, wherein the slope is increased fastest at the 800nm of 600-plus, the reflectivity is increased well at the 800nm, the increase rate is stable, no obvious rate difference exists between different salt concentrations, and then the increase rate of the slope is gradually decreased. A regular steep threshold appears around 1400nm, and then fluctuation of the reflectivity is large in the range from 2500 nm.
Comparison of the pretreatment spectrum (R), the envelope elimination spectrum (CR), and the reciprocal logarithmic spectrum (LR) of 3 spectral curves revealed that:
after the metal salt ion electron transfer molecular group in the mural is subjected to spectral enhancement in visible light and near infrared, the reflection characteristic is amplified, the detail characteristic is highlighted and hidden, the peaks and valleys of the spectrum are more prominent, the number of times of the peaks and valleys is more, and the influence of the salt content with different concentrations on the spectral reflectivity of the surface of a mural sample is more finely quantized; the application of LR transformation can reduce the influence on the spectrum due to illumination conditions and terrain differences, improve and compare the spectrum differences in the visible light range, and reduce random factor errors; the envelope removal is realized by normalizing the spectral reflectivity to form more absorption valleys, so that the correlation between the spectral characteristics and the salt content with different concentrations is enhanced, and more visual expression is realized.
4. Precision evaluation of prediction model
4.1, preparing a verification group mural sample block, specifically comprising a verification mural sample block (1 block) with the mass concentration of 0.1%, a verification mural sample block (1 block) with the mass concentration of 0.2%, a verification mural sample block (1 block) with the mass concentration of 0.3%, a verification mural sample block (1 block) with the mass concentration of 0.5%, a verification mural sample block (1 block) with the mass concentration of 0.7%, and a verification mural sample block (1 block) with the mass concentration of 0.9%, wherein the preparation method is the same as in example 1;
sampling the verification sample blocks for at least 1 time respectively, wherein the sampling method is the same as that of embodiment 1;
preprocessing the sampled data to obtain a preprocessing spectrum of the verification mural sample block in the wavelength range of 350-;
carrying out reciprocal logarithmic transformation on the preprocessed spectrum to obtain a reciprocal logarithmic spectrum, and carrying out envelope elimination on the preprocessed spectrum to obtain an envelope elimination spectrum;
for the pretreatment spectrum of the verification group, determining the characteristic wave bands to be 800nm and 1413 nm;
for the log spectrum of the reciprocal of the validation set, the characteristic band was determined to be 1415 nm;
determining the characteristic wave band to be 420nm for the envelope elimination spectrum of the verification group;
for the verification group, linear regression is carried out by taking the content concentration of soluble salt of the verification mural sample block as a dependent variable and taking the parameter value of the corresponding spectrum at the characteristic wave band as an independent variable to obtain an optimal fit line (as shown in figure 12), and an inversion verification equation of the content of sodium sulfate in the mural sample is established;
selecting Goodness-of-fit statistic (Goodness-of-fit statistical) to perform precision analysis on the established model, and respectively calculating and comparing decision coefficients (R) of the modeling set and the verification set2) Root Mean Square Error (RMSE) and relative analytical error (RPD), wherein，R2The method is used for judging the fitting effect of the model, the RMSE is used for judging the prediction capability of the model, the RPD can reduce the influence of the difference of the attribute values of the prediction samples in different researches to a certain extent, and the reliability of the model is measured, and the result is shown in the following table 2:
TABLE 2 precision analysis data
The prediction results of the four mural sample sodium sulfate content prediction models facing the modeling group and the verification group are shown in fig. 13-16, and the capabilities of the spectrum enhancement models at different characteristics in the aspect of estimating the sodium sulfate content of the mural sample are ranked from strong to weak by combining fig. 13-16 and table 2: and sequentially obtaining an envelope elimination spectrum > pretreatment spectrum (1413nm) > pretreatment spectrum (800nm) > envelope elimination spectrum, wherein the fitting effect of different wave bands under the same model is compared at the wave bands of 800nm and 1413nm, and the difference of the modeling results of different models at the same wave band is compared at the wave band of 1414nm or so.
While embodiments of the invention have been described above, it is not limited to the applications set forth in the description and the embodiments, which are fully applicable in various fields of endeavor to which the invention pertains, and further modifications may readily be made by those skilled in the art, it being understood that the invention is not limited to the details shown and described herein without departing from the general concept defined by the appended claims and their equivalents.
Claims (10)
1. The mural soluble salt content detection method based on the reflection spectrum is characterized by comprising the following steps:
making a plurality of standard mural sample blocks containing soluble salts with different mass concentrations;
sampling each standard mural sample block for at least 1 time by using a spectral radiometer at preset time intervals, wherein a pretreatment spectrum of the standard mural sample block in the wavelength range of 350-;
carrying out reciprocal logarithmic transformation on the preprocessed spectrum to obtain a reciprocal logarithmic spectrum;
extracting a characteristic wave band according to the reciprocal logarithmic spectrum;
taking the mass concentration of soluble salt of a standard mural sample block as a dependent variable, taking the parameter value of a reciprocal logarithmic spectrum at a characteristic wave band as an independent variable, performing linear regression to obtain an optimal fitting line, and establishing a mural sample soluble salt content prediction model;
and (3) acquiring a point-like spectrum curve of the designated position of the mural to be detected by using a spectrum radiometer, preprocessing, carrying out reciprocal logarithmic transformation, acquiring characteristic spectrum parameters at a characteristic wave band, bringing the characteristic spectrum parameters into a mural sample soluble salt content prediction model, and predicting the mass concentration of soluble salt.
2. The method for detecting the soluble salt content of the mural based on the reflection spectrum according to claim 1, wherein the method for manufacturing the standard mural sample block comprises the following steps:
manufacturing a coarse mud layer;
and manufacturing a fine mud layer on the surface of the coarse mud layer, and naturally drying in the shade to obtain a plurality of standard mural sample blocks, wherein soluble salt with corresponding mass concentration is uniformly mixed in the fine mud layer of each standard mural sample block.
3. The method for detecting the content of soluble salt in mural painting based on reflection spectrum according to claim 1, wherein the mass concentration range of the soluble salt formed by a plurality of standard mural painting sample blocks is 0-1%, wherein the mass concentration difference of the two adjacent standard mural painting sample blocks with the mass concentration is not more than 0.2% according to the mass concentration of the soluble salt in the standard mural painting sample blocks from small to large.
4. The method for detecting the content of soluble salts in murals based on the reflection spectrum as claimed in claim 3, wherein the standard mural sample blocks are arranged from small to large in mass concentration of soluble salts, and the mass concentration difference between any two adjacent standard mural sample blocks is equal.
5. The method for detecting the content of soluble salt of mural painting based on the reflection spectrum as claimed in claim 3, wherein the soluble salt is one of sulfate, nitrate and chloride.
6. The method for detecting the content of soluble salt of mural painting based on reflection spectrum according to claim 5, wherein the soluble salt is sodium sulfate.
7. The mural soluble salt content detection method based on reflection spectroscopy according to claim 1, wherein each sampling specifically comprises:
dividing each standard mural sample block into 3 x 3 blocks, and randomly selecting a central point of one block for data acquisition in each row and each column, wherein the method for acquiring data from each central point comprises the following steps:
s1, collecting an initial reflection spectrum curve within the wavelength range of 350-2500nm by aligning a probe of the spectrum radiometer with a standard mural sample block and a central point;
s2, horizontally rotating the probe of the spectrum radiometer by 90 degrees for next acquisition;
and S3, repeating the step S2 until the probe of the spectrum radiation instrument horizontally rotates 270 degrees.
8. The method for detecting the soluble salt content of the mural painting based on the reflection spectrum as claimed in claim 7, wherein the pretreatment after each sampling comprises the following steps:
sa, taking the arithmetic mean value of the breakpoint correction spectrum curves of the same central point as the spectrum curve of the central point;
and Sb, taking the arithmetic mean value of the average spectrum curves corresponding to the same standard mural sample block.
9. The method for detecting soluble salt content of murals based on reflection spectrum according to claim 8, wherein the pretreatment after each sampling further comprises performing breakpoint correction on the initial reflection spectrum curve before performing step Sa to obtain a breakpoint correction spectrum curve;
after the step Sb is carried out, Savitzky-Golay smoothing processing is carried out on the spectrum curve acquired by each time of the standard mural sample block, and a reflection spectrum of the standard mural sample block in the wavelength range of 350-2500nm is obtained.
10. The method for detecting the soluble salt content of the mural painting based on the reflection spectrum as claimed in claim 1, wherein the method for extracting the characteristic wave band according to the reciprocal logarithmic spectrum comprises the steps of:
selecting a wave band range with a spectrum showing obvious peaks and valleys and a spectrum curve showing stable monotonous change;
for the determined waveband range, after the pretreatment spectrum and the salt concentration are subjected to a p-0.05 Pearson correlation significance detection, the characteristic waveband of the reciprocal logarithmic spectrum is determined.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111614138.7A CN114279976B (en) | 2021-12-27 | 2021-12-27 | Method for detecting content of soluble salt in wall painting based on reflection spectrum |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111614138.7A CN114279976B (en) | 2021-12-27 | 2021-12-27 | Method for detecting content of soluble salt in wall painting based on reflection spectrum |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114279976A true CN114279976A (en) | 2022-04-05 |
| CN114279976B CN114279976B (en) | 2023-09-19 |
Family
ID=80876066
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111614138.7A Active CN114279976B (en) | 2021-12-27 | 2021-12-27 | Method for detecting content of soluble salt in wall painting based on reflection spectrum |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114279976B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070116628A1 (en) * | 2005-06-01 | 2007-05-24 | Chwen-Yang Shew | Charged carbon nanotubes for use as sensors |
| US20140320064A1 (en) * | 2013-04-25 | 2014-10-30 | Tseng-Lu Chien | USB Charger Related Products has Built-In Fluid and Display Unit |
| CN104897592A (en) * | 2015-06-11 | 2015-09-09 | 石河子大学 | Monitoring method of salt ion content in saline soil based on hyperspectral technology |
| CN105241810A (en) * | 2015-10-10 | 2016-01-13 | 敦煌研究院 | Method and device for preparing fresco restoration material and measuring cohesiveness |
| CN105510328A (en) * | 2014-09-25 | 2016-04-20 | 鞍钢股份有限公司 | Coal petrographic analysis test and graph processing method |
| CN105651663A (en) * | 2014-11-10 | 2016-06-08 | 敦煌研究院 | Method for determining water vapor permeability of mural and stone cultural relic protection material |
| CN206208820U (en) * | 2016-11-26 | 2017-05-31 | 浙江大学 | A kind of surface of solids salt content determination sensor |
| CN108052962A (en) * | 2017-11-29 | 2018-05-18 | 西安建筑科技大学 | A kind of Spectral matching algorithm based on improved edit-distance |
| CN108982407A (en) * | 2018-07-06 | 2018-12-11 | 浙江大学 | A method of probing into the soil optimum moisture content of detection soil nitrogen using near infrared spectrum |
| CN109738380A (en) * | 2019-01-25 | 2019-05-10 | 西北农林科技大学 | A kind of high-spectrum remote-sensing judgment method of soil salinization degree |
| CN112213287A (en) * | 2020-12-07 | 2021-01-12 | 速度时空信息科技股份有限公司 | Coastal beach salinity inversion method based on remote sensing satellite image |
| AU2021100533A4 (en) * | 2021-01-28 | 2021-04-22 | Qingdao Agricultural University | Soil salinity at Yellow River Delta Inversion Method based on Landsat 8 |
| CN112782096A (en) * | 2020-12-16 | 2021-05-11 | 南京信息工程大学 | Soil organic carbon density estimation method based on reflection spectrum data |
-
2021
- 2021-12-27 CN CN202111614138.7A patent/CN114279976B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070116628A1 (en) * | 2005-06-01 | 2007-05-24 | Chwen-Yang Shew | Charged carbon nanotubes for use as sensors |
| US20140320064A1 (en) * | 2013-04-25 | 2014-10-30 | Tseng-Lu Chien | USB Charger Related Products has Built-In Fluid and Display Unit |
| CN105510328A (en) * | 2014-09-25 | 2016-04-20 | 鞍钢股份有限公司 | Coal petrographic analysis test and graph processing method |
| CN105651663A (en) * | 2014-11-10 | 2016-06-08 | 敦煌研究院 | Method for determining water vapor permeability of mural and stone cultural relic protection material |
| CN104897592A (en) * | 2015-06-11 | 2015-09-09 | 石河子大学 | Monitoring method of salt ion content in saline soil based on hyperspectral technology |
| CN105241810A (en) * | 2015-10-10 | 2016-01-13 | 敦煌研究院 | Method and device for preparing fresco restoration material and measuring cohesiveness |
| CN206208820U (en) * | 2016-11-26 | 2017-05-31 | 浙江大学 | A kind of surface of solids salt content determination sensor |
| CN108052962A (en) * | 2017-11-29 | 2018-05-18 | 西安建筑科技大学 | A kind of Spectral matching algorithm based on improved edit-distance |
| CN108982407A (en) * | 2018-07-06 | 2018-12-11 | 浙江大学 | A method of probing into the soil optimum moisture content of detection soil nitrogen using near infrared spectrum |
| CN109738380A (en) * | 2019-01-25 | 2019-05-10 | 西北农林科技大学 | A kind of high-spectrum remote-sensing judgment method of soil salinization degree |
| CN112213287A (en) * | 2020-12-07 | 2021-01-12 | 速度时空信息科技股份有限公司 | Coastal beach salinity inversion method based on remote sensing satellite image |
| CN112782096A (en) * | 2020-12-16 | 2021-05-11 | 南京信息工程大学 | Soil organic carbon density estimation method based on reflection spectrum data |
| AU2021100533A4 (en) * | 2021-01-28 | 2021-04-22 | Qingdao Agricultural University | Soil salinity at Yellow River Delta Inversion Method based on Landsat 8 |
Non-Patent Citations (6)
| Title |
|---|
| R OLMI: "Diagnostics and monitoring of frescoes using evanescent-field dielectrometry", 《MEASUREMENT SCIENCE AND TECHNOLOG 》, vol. 17, no. 8, pages 2281 - 2288, XP020103663, DOI: 10.1088/0957-0233/17/8/032 * |
| 吴霞: "基于多光谱遥感的盐渍化评价指数对宁夏银北灌区土壤盐度预测的适用性分析", 《国土资源遥感》, vol. 33, no. 2, pages 124 - 133 * |
| 樊再轩: "莫高窟第98窟酥碱壁画保护修复试验研究", 《敦煌研究.》, no. 6, pages 4 - 7 * |
| 赵莉: "对流失海外的克孜尔石窟壁画的数字化复原探索", 《 浙江大学学报(理学版)》, vol. 47, no. 6, pages 651 - 659 * |
| 陈东强 等: "准格尔盆地人工林地土壤全盐的高光谱反演", 《干旱区研究》, vol. 30, no. 3, pages 444 - 448 * |
| 魏瑞: "辽宁奉国寺建筑壁画可溶盐调查与分析", 《文物鉴定与鉴赏》, no. 13, pages 92 - 93 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114279976B (en) | 2023-09-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108593569B (en) | EO-1 hyperion water quality parameter quantitative inversion method based on spectrum morphological feature | |
| CN114018833B (en) | Method for estimating heavy metal content of soil based on hyperspectral remote sensing technology | |
| Sandak et al. | Characterization and monitoring of surface weathering on exposed timber structures with a multi-sensor approach | |
| CN110531054B (en) | Soil organic carbon prediction uncertainty estimation method based on Bootstrap sampling | |
| Tuzet et al. | Influence of light-absorbing particles on snow spectral irradiance profiles | |
| CN111965140B (en) | Wavelength point recombination method based on characteristic peak | |
| CN110987846B (en) | Nitrate concentration prediction method based on iPLS-PA algorithm | |
| CN104931453A (en) | Method for predicting water content of spread green leaves of green tea based on near infrared spectrum technology | |
| CN102937575B (en) | Watermelon sugar degree rapid modeling method based on secondary spectrum recombination | |
| CN104374711B (en) | A kind of trees blade face dust method for determination of amount and system | |
| CN110596028A (en) | High-spectrum inversion method for content of deposited rare earth La element | |
| CN115063034A (en) | Glass manufacturing on-line monitoring analysis management system based on artificial intelligence | |
| CN111879725A (en) | Spectral data correction method based on weight coefficient | |
| CN107632010B (en) | Method for quantifying steel sample by combining laser-induced breakdown spectroscopy | |
| CN110779875B (en) | Method for detecting moisture content of winter wheat ear based on hyperspectral technology | |
| CN107907490A (en) | Soil erosion EO-1 hyperion inversion method based on outdoor rainfall and indoor soil | |
| CN108827909B (en) | Rapid soil classification method based on visible near infrared spectrum and multi-target fusion | |
| CN113686811A (en) | Spectral data processing method based on double sensors | |
| CN114279976A (en) | Mural soluble salt content detection method based on reflection spectrum | |
| CN117491293A (en) | High-steep bank slope carbonate rock corrosion rapid evaluation method based on hyperspectrum | |
| CN110567385A (en) | Hyperspectral technology-based construction thickness detection method for building reflective insulation coating | |
| Liu et al. | Study on hyperspectral estimation model of total nitrogen content in soil of shaanxi province | |
| CN115266648A (en) | Optimization simulation method for intrinsic optical parameters of second-class water body | |
| Zheng et al. | Evolution of paddy soil fertility in a millennium chronosequence based on imaging spectroscopy | |
| CN112229817A (en) | Method for establishing soda saline-alkali soil heavy metal quantitative inversion model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |