CN207894825U - Drawing pigment identifying system based on multispectral imaging - Google Patents
Drawing pigment identifying system based on multispectral imaging Download PDFInfo
- Publication number
- CN207894825U CN207894825U CN201820207087.3U CN201820207087U CN207894825U CN 207894825 U CN207894825 U CN 207894825U CN 201820207087 U CN201820207087 U CN 201820207087U CN 207894825 U CN207894825 U CN 207894825U
- Authority
- CN
- China
- Prior art keywords
- camera
- pigment
- image
- spectral
- light
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The drawing pigment identifying system based on multispectral imaging that the utility model is related to a kind of, the system includes lighting source, bandpass filter group and camera, the bandpass filter group is made of the optical filter of multiple corresponding different spectral coverages, and the camera is connected with image processing unit；Lighting source is used to send out illuminating ray to drawing to be measured, and the light that drawing to be measured reflects enters camera after penetrating optical filter.The utility model solve it is existing based on chemical analysis or based on insertion type present in the drawing pigments recognition methods such as light and substance reciprocation, have the technical problems such as damage, efficiency be low.The utility model identifying system and method have many advantages, such as that spatial resolution is high, precision is high, non-contact, not damaged.
Description
Technical field
The drawing pigment identifying system based on multispectral imaging that the utility model is related to a kind of.
Background technology
The recognition methods of traditional drawing pigment is broadly divided into three classes at present：The first kind is the method using chemical analysis, this
Kind method is with high accuracy, but needs to be sampled from drawing ontology, and analysis result is only effective to sample, can not spread to
Sampling location is with exterior domain；Second class be include that X-ray diffraction, Laser Roman spectroscopic analysis of composition are based on light or ray and are interacted with substance
The analysis method of effect, such methods have preferable analysis result, but since analytic process is often related in molecular atoms
Excitation of the level to pigment, can lead to the variation of pigment molecule structure, to have damaging analysis method as a kind of, by
In the particularity of the drawing class historical relic art work, such method is difficult to be widely popularized in practical applications.Meanwhile this method
Also it is only capable of analyzing tiny area, it is difficult to meet the requirement of drawing whole picture surface analysis.Third class is anti-based on pigment spectrum
The analysis method of rate is penetrated, the instrument of generally use is fiber spectrometer.Such method is studied the light of region pigment by acquisition
Reflectivity is composed to identify its type, but this method is only capable of obtaining the pigment spectral information of tiny area every time, it is difficult to be generalized to more
On a large scale, it is difficult to effectively play a role in practical applications.
Utility model content
The utility model aim is to provide a kind of drawing pigment identifying system based on multispectral imaging, solves existing
Based on chemical analysis or based on insertion type present in the drawing pigments recognition methods such as light and substance reciprocation, have damage, effect
The low technical problem of rate.
The technical solution of the utility model is：A kind of drawing pigment identifying system based on multispectral imaging, it is special
Different place is：Including lighting source, bandpass filter group and camera, the bandpass filter group is by multiple corresponding different spectral coverages
Optical filter composition, the camera is connected with image processing unit；Lighting source is used to send out illumination light to drawing to be measured
Line, the light that drawing to be measured reflects enter camera after penetrating optical filter.
Further, the union covering of the spectral-transmission favtor curve of each optical filter can in above-mentioned bandpass filter group
It is light-exposed near infrared band, the spaced distribution of spectral transmission peak value of each optical filter.
Further, above-mentioned camera is that monochrome CMOS industry lines sweep camera, and the imaging sensor of camera has a line 8192
The spectral sensitivity curve coverage area of a pixel, pixel is 400-900nm.
Further, above-mentioned camera is connected by CameraLink interfaces with image processing unit.
Further, above-mentioned lighting source is cold light light source.
The utility model also provides a kind of drawing pigment recognition methods based on multispectral imaging, is characterized in that,
Include the following steps：
1) image obtains；
1.1) lighting source treat each other survey drawing surface send out illuminating ray；
1.2) before an optical filter in bandpass filter group being placed on camera lens, the light of drawing reflection to be measured
Line enters camera lens after penetrating optical filter；
1.3) it is transmitted to image processing unit after camera acquisition image；
1.4) camera is placed on after replacing previous optical filter with another optical filter in bandpass filter group
Before camera lens, camera acquires image and is transmitted to image processing unit again；
1.5) step 1.4) is repeated until completing the figure that whole optical filters in bandpass filter group correspond to spectral coverage
As acquisition, the multispectral image of drawing to be measured is obtained；
2) image procossing；
2.2) spectral reflectance recovery：
The relational expression p=CLr+e for solving image pixel value and pigment spectral reflectivity in multispectral image, obtains waiting surveying and drawing
The spectral reflectivity curve of target pixel location in the multispectral image of picture；
Wherein, p is the multispectral image pixel value vector of M*1, and C is the camera spectrum sensitivity matrix of M*N, and L is N*N
Light source light spectrum radiates diagonal matrix, and r is the spectral reflectivity vector of N*1, and e is the additive noise vector of M*1, and M is image channel
Number, N is Spectral dimension；
3) data analysis；
The spectral reflectivity curve for the target pixel location that step 2) obtains is compared to obtain mesh with colorant data library
Mark the pigment type and feature of location of pixels.
Further, the relational expression of image pixel value and pigment spectral reflectivity in multispectral image is solved in step 2.2)
Method include the following steps：
2.2.1 the spectral reflectivity property database of drawing pigment sample database and drawing pigment sample database) is established, and with
The spectral reflectivity property database is as learning sample；
2.2.2) ignore additive noise vector e, and by image pixel value in multispectral image and pigment spectral reflectivity
Relational expression p=CLr+e is reduced to p=Hr；Wherein, H is to represent camera and the M*N matrixes of light source light spectrum characteristic；
2.2.3 it) asks and calculates W=RP+；Wherein, W is the estimation to matrix H, and R is the spectral reflectance rate matrix of learning sample, P
It is multispectral image pixel matrix, P+It is the pseudo inverse matrix of P；
2.2.4) by step 2.2.3) obtained matrix W acts on multispectral image pixel value vector p, it is acquired by r*=Wp
The spectral reflectivity r* of target pixel location.
Further, step 2) further includes the step 2.1) image preprocessing executed before step 2.2)：To target picture
Pixel value at plain position (x, y) carries out Shading corrections.
Further, the method for progress Shading corrections is in step 2.1)：It acquires at target pixel location (x, y)
Correct preceding pixel value pT(x, y) then calculates pixel value p ' after correctionT(x,y)：
Wherein, SLHorizontal, the p for correctionw(x, y) is pixel value of the reference white plate at the position (x, y).
Further, further include that the step 2.3) coloured image executed side by side with step 2.2) is rebuild in step 2)：
2.3.1 it) calculates and corresponds to pigment spectral reflectivity rλColor tristimulus specifications X, Y, Z：
Wherein, Wx,λ、Wy,λ、Wz,λIt is the color matching function for corresponding respectively to color tristimulus specifications X, Y, Z；
2.3.2) color tristimulus specifications is converted to the tristimulus specifications in the spaces sRGB：
Wherein,
2.3.3 the transmission function of RGB color component) is calculated：
The coloured image that drawing to be measured is completed according to the transmission function being calculated is rebuild.
The beneficial effects of the utility model are：
1, high spatial resolution：The highest spatial imaging resolution ranging from 1200DPI that the utility model may be implemented, i.e.,
The detailed information that about 20 micrometer lengths can be sampled can meet the high-space resolution of the various complete breadth surface informations of drawing
Rate is imaged, the details of the drawing surface information that high-definition picture is included enrich the remote superman's eye of degree in one's power, be effective district
The pigment of split screen different zones or feature of painting provides powerful guarantee；
2, Non-contact nondestructive is hindered：The utility model is acquired for a kind of optical information and analysis method, multispectral image
Data acquisition is similar to shooting photo, will not be contacted between instrument and equipment and drawing, and the cold light source of independent research exists
While ensureing illumination validity when multi-optical spectrum image collecting, drawing will not be made because of high brightness or the golf calorific value of light source
At damage, therefore, this method belongs to a kind of non-contact, undamaged analysis method；
3, high accuracy：The physical principle of the pigment identification of the utility model is the spectral reflectivity curve based on pigment,
Since different material in molecular atoms level there is different constituent and structure, different pigment to have different spectral reflectances
Rate curve, and spectral reflectivity curve has uniqueness, there are one-one relationships with pigment composition.Establish pigment sample and its
On the basis of spectral reflectivity, then it is compared with pigment library of spectra by obtaining pigment spectral reflectivity, to real
The purpose of existing pigment identification.Key among these is accurately to obtain pigment spectral reflectivity characteristic, accessed by this method
Spectral reflectivity accuracy is up to 99% or more；
4, full width face：Pigment identification in the utility model is based on image pixel, for arbitrarily selected in image
One pixel, can identify the type of the pixel position pigment.Selection interested pixel point can obtain pigment
Type has the characteristics that quick, efficient, intuitive.
Description of the drawings
Fig. 1 is the system composition schematic diagram of drawing pigment identifying system of the utility model based on multispectral imaging.
Fig. 2 is the spectral-transmission characteristics curve of bandpass filter group in the utility model.
Fig. 3 is the flow diagram of drawing pigment recognition methods of the utility model based on multispectral imaging.
Wherein, reference numeral is：1- lighting sources, 2- drawing to be measured, 3- bandpass filter groups, 4- optical filters, 5-
Camera, 6- image processing units.
Specific implementation mode
The utility model is a kind of traditional drawing pigment identifying system and recognition methods based on multi-optical spectrum imaging technology, is somebody's turn to do
System is specifically included the image collection processing system based on camera and image processing unit and is constituted with bandpass filter group
Beam splitting system, equipped with pigment spectral reflectance recovery algorithm, drawing pigment sample database and its light on image processing unit
Compose reflectivity Characteristics database.
Referring to Fig. 1, the component part of the utility model drawing pigment identifying system preferred embodiment include lighting source 1,
Bandpass filter group 3 and camera 5, bandpass filter group 3 are made of the optical filter 4 of multiple corresponding different spectral coverages, camera 5 with
Image processing unit 6 is connected.Drawing 2 to be measured is lain against on original text platform, and camera 5 is suspended to be measured by certain Design of Mechanical Structure
On drawing 2, camera depends on desired imaging resolution at a distance from drawing to be measured, and the optical axis of camera is perpendicular to drawing to be measured
The plane at place.
Lighting source 1 is used to send out illuminating ray to drawing 2 to be measured, and the light of 2 reflection of drawing to be measured penetrates optical lightscreening
Enter camera 5 after piece 4.
Camera 5 can select monochromatic CMOS industry line to sweep camera, and imaging sensor has 8192 pixels of a line, pixel
Number is enough to ensure that the acquisition of high spatial resolution images.The spectral sensitivity curve coverage area of pixel is 400nm-900nm, including
Entire visible light region and part near infrared region.Camera has high dynamic range and High Data Rate simultaneously, it is ensured that imaging system
System can obtain the spectrum picture of drawing with higher speed.High sensitive is but also the camera has the ability to distinguish extremely subtle
Spectral reflectivity curve difference.Camera is connected to the same configuration with CameraLink interfaces with CameraLink interfaces and exists
On image pick-up card on host computer, the course of work of camera is controlled by image pick-up card.
In addition to the industrial camera of superior performance, the utility model has also been devised matching by the optical elements structure such as camera lens
At optical system, can reach the image spatial resolution for meeting drawing surface detail analysis, and image have it is high several
What precision.
The size of optical filter in bandpass filter group 3 is more than the diameter of camera lens, can be directly placed in
Before camera lens and it is completely covered by entire camera lens.Referring to Fig. 2, the spectrum of each optical filter 4 is saturating in bandpass filter group 3
The union covering visible light of rate curve is penetrated near infrared band, the spectral transmission peak value of each optical filter have it is rational between
Every distribution, and adjacent spectrum transmittance graph has minimum coincidence.
The image acquisition circuit that camera is swept based on monochromatic CMOS industry line is controlled in mechanical scanning transmission and stepper motor movement
Under the cooperation of system, optical lens, lighting source etc., the image content of drawing to be measured is converted into digital picture.Multispectral image
Acquisition realized by switching one group of optical filter, the image in a channel is acquired using optical filter every time,
Complete multispectral image is obtained by way of gradually switchable optics optical filter.In the drawing pigment spectral reflectance established in advance
On the basis of rate diagram database, using spectral reflectance recovery algorithm by the mostly light of each pixel position in multispectral image
Spectrogram picture is converted to the spectral reflectivity curve of pigment at this, so by with drawing pigment spectral reflectivity curve database
It compares to realize the identification of pigment.
As shown in figure 3, drawing pigment recognition methods of the utility model based on multispectral imaging is broadly divided into three steps：
Image acquisition, image procossing and data analysis.Image obtains the high-resolution multi-spectral image for acquiring drawing to be measured；At image
Reason may include two steps, be first pretreatment, be secondly two operations carried out parallel, respectively spectral reflectance recovery
It is rebuild with coloured image.Data analysis phase be operator by observation to coloured image after reconstruction and by with calculating
The interested pixel of machine interactive selection is to know selected location of pixels pigment type and feature.Point of pigment type and feature
Analysis is that realization is compared with colorant data library by its spectral reflectivity curve.
It elaborates with reference to specific implementation step the utility model drawing pigment recognition methods.
1) image obtains；
1.1) lighting source treat each other survey drawing surface send out illuminating ray；
1.2) before an optical filter in bandpass filter group being placed on camera lens, the light of drawing reflection to be measured
Line enters camera lens after penetrating optical filter；
1.3) it is transmitted to image processing unit after camera acquisition image；
1.4) camera is placed on after replacing previous optical filter with another optical filter in bandpass filter group
Before camera lens, camera acquires image and is transmitted to image processing unit again；
1.5) step 1.4) is repeated until completing the figure that whole optical filters in bandpass filter group correspond to spectral coverage
As acquisition, the multispectral image of drawing to be measured is obtained；
2) image procossing；
2.2) spectral reflectance recovery：
The relational expression p=CLr+e for solving image pixel value and pigment spectral reflectivity in multispectral image, obtains waiting surveying and drawing
The spectral reflectivity curve of target pixel location in the multispectral image of picture；
Wherein, p is the multispectral image pixel value vector of M*1, and C is the camera spectrum sensitivity matrix of M*N, and L is N*N
Light source light spectrum radiates diagonal matrix, and r is the spectral reflectivity vector of N*1, and e is the additive noise vector of M*1, and M is image channel
Number, N is Spectral dimension；
3) data analysis；
The spectral reflectivity curve for the target pixel location that step 2) obtains is compared to obtain mesh with colorant data library
Mark the pigment type and feature of location of pixels.
For step 2) spectral reflectance recovery, it is assumed that multispectral that there is M channel, image pixel value in multispectral image
It can be described by following formula with the relationship of pigment spectral reflectivity
P=∫ C (λ) L (λ) r (λ)+e
Wherein, the multispectral image pixel value that p is M*1 is vectorial (i.e. imaging sensor response vector), and C (λ) is the phase of M*1
Machine spectrum sensitivity vector, L (λ) are light source light spectrum radiation vectors, and r (λ) is pigment spectral reflectivity curve, and e is the additivity of M*1
Noise vector.
Consider for terseness mathematically, above formula can be expressed with easier matrix and vector：
P=CLr+e
Wherein, C is the camera spectrum sensitivity matrix of M*N, and L is the light source light spectrum radiation diagonal matrix of N*N, and r is N*1
Spectral reflectivity vector.
There are linear relationships between the response and pigment spectral reflectivity of the formula image sensor, thus, by light
Spectrum reflectivity can be expressed to the transmission function that imaging sensor responds by matrix.Sensor response, camera spectral response curve,
Vector relations between light source light spectral power distributions and the spectral reflectivity of research object can pass through direct or indirect approach
To solve.Direct method is usually required about the prioris such as camera sensitivity curve and light source light spectral power distributions.Although this
The accuracy of one method is indubitable, but Feasible degree in practice is very low.The spectral response curve of camera and the light of light source
Spectral power distributions are often unknowable, even if can be obtained from manufacturer, since the performance of light source etc. became with time and use environment
Change, the accuracy of data is also difficult to keep for a long time.Another kind is to utilize learning sample without the method for solving of priori.
Learning sample can be used to estimate the spectral characteristic of estimation camera, light source, camera lens etc., the system change without these parameters
Problem.As long as learning sample selection is appropriate, then the accuracy estimated can be ensured.The utility model is by independently building drawing
Conventional pigment sample database and its spectral reflectivity property database were formed as learning sample to solve multispectral image
Transition matrix in journey.
The relational expression p=CLr+e of image pixel value and pigment spectral reflectivity can be reduced to p=in multispectral image
Hr, wherein H is to represent camera and the M*N matrixes of light source light spectrum characteristic, and error e, which is ignored, to be disregarded.
In the case of no priori, relational expression p=Hr can be solved by pseudoinverse.Pseudo- inversion model is considered as
It is the modification estimated wiener using regression analysis.Wherein, indicate that the spectral reflectance rate matrix of learning sample, P indicate mostly light with R
For spectrogram as pixel matrix, W is the estimation to matrix H.So that the W that ‖ R-WP ‖ are minimized is represented by：
W=RP+=RPt(PPt)-1
Wherein, P+Represent the pseudo inverse matrix of P.By the way that W to be acted on to the pixel value vector p of multispectral image position,
Spectral reflectivity r* can be expressed as：
R*=Wp
Based on the solution procedure of learning sample utilized above, image in multispectral image is solved in the present embodiment step 2.2)
Pixel value and the method for the relational expression of pigment spectral reflectivity include the following steps：
2.2.1 the spectral reflectivity property database of drawing pigment sample database and drawing pigment sample database) is established, and with
The spectral reflectivity property database is as learning sample；
2.2.2) ignore additive noise vector e, and by image pixel value in multispectral image and pigment spectral reflectivity
Relational expression p=CLr+e is reduced to p=Hr；Wherein, H is to represent camera and the M*N matrixes of light source light spectrum characteristic；
2.2.3 it) asks and calculates W=RP+；Wherein, W is the estimation to matrix H, and R is the spectral reflectance rate matrix of learning sample, P
It is multispectral image pixel matrix, P+It is the pseudo inverse matrix of P；
2.2.4) by step 2.2.3) obtained matrix W acts on multispectral image pixel value vector p, it is acquired by r*=Wp
The spectral reflectivity r* of target pixel location.
Due to the limitation of the uneven optical device manufacturing technology level of optical profile, image sensing is reached in imaging process
Usually there is non-uniform phenomenon in the light on device surface.Ideally, it is assumed that be more than one piece of white standard of viewing field of camera to area
Sample carries out Image Acquisition, then response having the same is answered from first pixel of sensor to the last one pixel, and in reality
In border, the response for being typically at the pixel in sensor centre position is responded higher than the pixel of both sides, and the response of both sides pixel is all
As the increase with intermediate pixel distance tapers off trend, ultimately cause bright among generated image and both sides are gradually dimmed
Phenomenon.The phenomenon of distribution of light sources unevenness is referred to as Shading in this image.In order to eliminate distribution of light sources present in image not
The step 2) of the problem of uniformity, the present embodiment further include the step 2.1) image preprocessing executed before step 2.2)：It is right
Pixel value at target pixel location (x, y) carries out Shading corrections.
The relationship of pixel value and brightness of illumination in image at location of pixels (x, y) can be described by following formula：
P (x, y)=IL(x,y)*c(x,y)
p(x,y):Image pixel value
IL(x,y):Light-source brightness value
(x,y)：Location of pixels
OD(x,y):Optical density (OD)
By discussed above it is found that under conditions of the conditions such as light source immobilize, each pixel position in imaging sensor
There are a difference of coefficients between the pixel value and desired pixel value at place, can realize brightness of image not by obtaining this coefficient
The elimination of consistency.The relationship between the image pixel value before image pixel value and correction after correction can be retouched by following formula
It states：
p′T(x, y)=σS*pT(x,y)
Wherein, SLHorizontal, the p for correctionw(x, y) is pixel value of the reference white plate at the position (x, y), pT(x, y) is target
Correction preceding pixel value at location of pixels (x, y), p 'T(x, y) is pixel value after the correction at target pixel location (x, y), σSFor
Shading correction coefficient.
Based on the above analysis, the method that Shading corrections are carried out in step 2.1) is specifically：Acquire target pixel location
Correction preceding pixel value p at (x, y)T(x, y) then calculates pixel value p ' after correctionT(x,y)：
Wherein, SLHorizontal, the p for correctionw(x, y) is pixel value of the reference white plate at the position (x, y).
If not considering light conditions and observer, substance causes the essential attribute of color to be spectral reflectivity.Obtain material
The spectral reflectivity of material obtains the basic physical attribute of the material color, this eliminates the need for traditional RGB camera and is obtaining color
Metamerism phenomenon when color.It further include the step 2.3) coloured image executed side by side with step 2.2) in the present embodiment step 2)
It rebuilds：
2.3.1 it) calculates and corresponds to pigment spectral reflectivity rλColor tristimulus specifications X, Y, Z：
Wherein, Wx,λ、Wy,λ、Wz,λIt is the color matching function for corresponding respectively to color tristimulus specifications X, Y, Z；
2.3.2) color tristimulus specifications is converted to the tristimulus specifications in the spaces sRGB：
Wherein,
When the Y-component in white reference XYZ tristimulus specifications is set as 1, and the value that will exceed [0,1] range in rgb value
It is cut.
2.3.3 the transmission function of RGB color component) is calculated：
The coloured image that drawing to be measured is completed according to the transmission function being calculated is rebuild.
Spectral reflectance recovery algorithm designed by the utility model has low computation complexity, memory cost and time
The features such as cost is small.The methods of traditional Wiener filtering often requires that the light of the spectral response curve of camera, camera lens in imaging system
The spectral power distribution of spectrum transmission curve even light source is needed known or is individually solved out, could therefrom estimate light extraction
The spectral characteristic of spectrum filter piece or other light-splitting devices.And the algorithm in the utility model without individually estimate camera, camera lens,
The spectral characteristic of the imaging system components such as light source, and the spectral characteristic of entire imaging system is handled as a whole,
And finally for the estimation of pigment spectral reflectivity curve then by being rung by calculated imaging system spectrum as a whole
The combination of characteristic and pigment spectral characteristic database is answered to realize.This mode had both avoided more stringent and severe in conventional method
The known a priori condition at quarter, and effectively prevent, by the deviation accumulation in substep solution procedure, high accuracy capable of being obtained
Pigment spectral reflectivity curve, to improve reliability and the accuracy of the identification of system pigment.
Claims (5)
1. a kind of drawing pigment identifying system based on multispectral imaging, it is characterised in that：Including lighting source, bandpass filter
Group and camera, the bandpass filter group are made of the optical filter of multiple corresponding different spectral coverages, at the camera and image
Unit is managed to be connected；Lighting source is used to send out illuminating ray to drawing to be measured, and the light of drawing reflection to be measured penetrates optical lightscreening
Enter camera after piece.
2. the drawing pigment identifying system according to claim 1 based on multispectral imaging, it is characterised in that：The band logical
The union covering visible light of the spectral-transmission favtor curve of each optical filter is near infrared band, each optics in filter set
The spaced distribution of spectral transmission peak value of optical filter.
3. the drawing pigment identifying system according to claim 1 or 2 based on multispectral imaging, it is characterised in that：It is described
Camera is that monochrome CMOS industry lines sweep camera, and the imaging sensor of camera has 8192 pixels of a line, the spectrum sensitive of pixel
Curve coverage area is 400-900nm.
4. the drawing pigment identifying system according to claim 3 based on multispectral imaging, it is characterised in that：The camera
It is connected with image processing unit by CameraLink interfaces.
5. the drawing pigment identifying system according to claim 1 or 2 based on multispectral imaging, it is characterised in that：It is described
Lighting source is cold light light source.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201820207087.3U CN207894825U (en) | 2018-02-06 | 2018-02-06 | Drawing pigment identifying system based on multispectral imaging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201820207087.3U CN207894825U (en) | 2018-02-06 | 2018-02-06 | Drawing pigment identifying system based on multispectral imaging |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN207894825U true CN207894825U (en) | 2018-09-21 |
Family
ID=63543309
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201820207087.3U Active CN207894825U (en) | 2018-02-06 | 2018-02-06 | Drawing pigment identifying system based on multispectral imaging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN207894825U (en) |
-
2018
- 2018-02-06 CN CN201820207087.3U patent/CN207894825U/en active Active
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106896069B (en) | A kind of spectrum reconstruction method based on color digital camera single width RGB image | |
| CN105913093B (en) | A kind of template matching method for Text region processing | |
| Foster et al. | Frequency of metamerism in natural scenes | |
| Kim et al. | 3D imaging spectroscopy for measuring hyperspectral patterns on solid objects | |
| US7894662B2 (en) | Method for using image depth information in identifying illumination fields | |
| Liang et al. | Remote spectral imaging with simultaneous extraction of 3D topography for historical wall paintings | |
| DE102015210440A1 (en) | Image inspection device, image inspection method, image inspection program, computer readable recording medium and recording device | |
| Pillay et al. | Hyperspectral imaging of art: Acquisition and calibration workflows | |
| CN102721471B (en) | Method for conversion of spectrum into chromaticity | |
| Nayak et al. | Ground truth study on fractal dimension of color images of similar texture | |
| CN108267426B (en) | Drawing pigment identification system and method based on multispectral imaging | |
| Nocerino et al. | Mapping VIS and UVL imagery on 3D geometry for non-invasive, non-contact analysis of a vase | |
| Wang et al. | A novel low rank smooth flat-field correction algorithm for hyperspectral microscopy imaging | |
| CN207894825U (en) | Drawing pigment identifying system based on multispectral imaging | |
| Li Vigni et al. | Coupling 2D‐wavelet decomposition and multivariate image analysis (2D WT‐MIA) | |
| Kuzio et al. | Comparing practical spectral imaging methods for cultural heritage studio photography | |
| Luzi et al. | Automatic evaluation of vertex structural defects on the anode surface of a low-light-level image intensifier based on proposed individual image processing strategies | |
| Bringier et al. | Photometric reconstruction of a dynamic textured surface from just one color image acquisition | |
| Manakov | Evaluation of computational radiometric and spectral sensor calibration techniques | |
| Simon et al. | Integration of high resolution spatial and spectral data acquisition systems to provide complementary datasets for cultural heritage applications | |
| Tominaga et al. | Estimation of surface properties for art paintings using a six-band scanner | |
| Tateishi et al. | Study on the spectral quality preservation derived from multisensor image fusion techniques between JERS-1 SAR and Landsat TM data | |
| Plata et al. | Trichromatic red-green-blue camera used for the recovery of albedo and reflectance of rough-textured surfaces under different illumination conditions | |
| Soszyńska et al. | Feasibility study of hyperspectral line-scanning camera imagery for remote sensing purposes | |
| Ciortan et al. | Fluorescence transformation imaging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant |