Computer Aided Protection and Restoration of Dunhuang Mural
Computer Aided Protection and Restoration of Dunhuang Mural Dongming Lu College of Computer Science Zhejiang University Hangzhou, China [email protected]
Hongfang Han College of Computer Science Zhejiang University Hangzhou, China [email protected]
and based on that they did scientific methods to protect and restore of the caves with right materials, and had achieved some fruits.
Abstract –The paper proposes a computer aided way for culture heritage protection and restoration, which is based on virtual navigation and image processing technologies. It first creates the database for the digital information of the cave construction and the mural, and also provides some operations, such as insert, query etc, for the database management. In a virtual environment of the cave, users can navigate, mark the disease area, inquiry the information about the mural, monitor and analyze the environment data, and also do virtual mural restoration and simulate the mural evolution without damaging the heritage itself. The article introduces the following technologies: virtual construction of the cave, knowledge expression of Mural, interactive disease marking, object extracting, color restoration with mural knowledge, color harmonization for simulating the mural evolution, and also the functions, modules, workflow of our system.
This paper proposes a computer aided way for culture heritage protection and restoration which is base on virtual navigation and image processing technologies. It helps the protection experts to mange information, monitor environment, inquiry disease, analyze data, and also do color restoration and evolution simulation. The integration of information technologies and traditional methods of protection and restoration supplies a more efficient method for Dunhuang mural which is one of the best world cultural heritages.
2 2.1
Virtual construction of the cave
Virtual cave modeling is the precondition of the computer aided cultural relic protection and restoration system, which is consisted of measurement, photograph and 3d modeling. The digitization of cave provides virtual scene not only for our system, but also for cave navigation and computer aided archaeology, and at the same time, it’s a new organization of cave construction and mural information.
Keywords: Cultural heritage digitization, protection, restoration, Dunhuang mural.
1
Key techniques
Introduction
The Mogao Caves in Dunhuang City is a treasure of Buddha Culture in China, and even in the world, which integrated with architecture, color statue and mural. It is in an importance place in the art history of China and the world.
The 3d cave model is constructed with 3dmax according to the measurement, and textured by the highresolution image of mural. The mural digitization is an important step in the process of cave digitization, which records the beautiful mural in the computer by photographing and image processing. We design a specific equipment for mural photographing, and also standardize the workflow. After mural photographing, color and shape adjustments are done to archive the highresolution digital image of mural.
Mural is the most important portion of Dunhuang art. Because of the oxidation of plumbum which is the ingredient of pigment for mural, the influence of the climate, sunlight, humidity, exhausted gas, bacterium and mildew, and destruction by people in addition for thousands years, the caves are suffering from many kinds of diseases, such as pulverization, falling off, fading, and color changing[1,2]. Latter-day experts of Dunhuang Academe did many works on environment monitoring and research on the causes of cave illness and mural fading,
0-7803-8566-7/04/$20.00 2004 IEEE This work is funded by Hi-tech Research and Development Program of China Project(No. 2003AA119020) and Zhejiang province science and technology department project(No. 2004C23035).
6434
2.2
2.3.1
Knowledge expression of mural
The knowledge of mural includes both vision character and semantic character. The vision character means those characters such as the color and layout of the mural, and the semantic character means the content and background description of the mural. The definition of semantic character is given by: Fs S No u S D u ST ,Where S No is the cave number in which the mural is, S D is the dynasty when the mural was painted, ST is the theme of the mural. The Knowledge of Dunhuang color is presented by three color rules: the rule of typical initial Dunhuang color, the rule of typical current Dunhuang color and the change rule of typical Dunhuang color.
Interactive dynamic programming
The description of disease marking by interactive dynamic programming [3] is as following: first of all, we get a weighted-graph from the mural image by defining an edge between every pair of neighboring pixels whose weight is determined by the image characteristic such as gradient. The interactive dynamic programming requires user to specify a start point and end point at edge of the disease area, then the algorithm can find the shortest path in the weighted-graph which form part of the disease edge. This procedure repeats with human’s judgment and interaction and finally can get an enclosed disease edge. The core of the algorithm is defining the weight of edges (We are using gradient). The shortest path is calculated by Dijkstra. A sample is shown in Figure 1.
Both the typical initial and current Dunhuang color rules are given by: ColorRule=<Time, Position, Value, C1, C2, Ă Ă , Cm>,Where Time is the dynasty, Position records where this color has been used, Value is given by a center point and a region, C1, C2, Ă Ă , Cm are harmonic colors of the current one expressed by Value. The change rule of typical Dunhuang color is expressed by some seriate colors according time order: ChgColorRule=, Where C1, C2,ĂĂ, Cn are the colors, Curve is the fit curve of those colors. We also use hierarchy structure to express the color knowledge of the mural. First the mural is divided into several classes such as construction, character, and decoration and so on by the mural theme. And each class is continuously divided until they are impartible color objects. The hierarchy expression gives us efficient color indices when we deal with the complicated mural color relationship.
2.3
Figure 1. Disease marking with interactive dynamic programming
2.3.2
Interactive region growing
Interactive region growing [3,7] can be also used for disease marking. The disease area here could be disjoint. Each pixel satisfying following formulas is regarded in the disease area. In the growing, the position and size of seed points are determined by the user interactively.
Interactive disease marking
In our protection and restoration system, disease marking is the foundation of disease inquiry and analysis. It could be archived completely by manual work, and actually we did disease marking manually in traditional method. But manual disease marking is a heavy work, and also the result is not satisfied in precision. On the other hand, considering of the complexity of Dunhuang mural it takes more time to find out a technique to mark the disease automatically. As a trade-off, we use manualautomatic marking methods in our system, which are acceptable in both speed and precision. Because in these manual-automatic methods, the capacity of judgment and computer are well combined.
d g 1 (C (i, j ), C ( m, n)) d G ptp °° ® d g 2 (C (i, j ), CR (i, j )) d G ptr ° °¯ d g 3 (C (i, j ), Cs (i, j )) d G pts
(1)
In the above formulas, C (i, j ) represents the color of pixel (i, j ) , C (m, n) , CR (i, j ) and Cs (i, j ) separately represent the color of neighboring pixel, the average color in the growing area and the color of the seed point. G ptp ,
6435
G ptr and G pts are thresholds and they satisfy G pts ! G ptr ! G ptp . In d gi () , the Euclidean distance of HIS is used
2.4.2
Color restoration
In Our experiments, we found some Dunhuang murals are kept rather well, but lost their original bright color in the long history. Color transfer can be used to restore the original color. Keeping the original image texture is necessary in color transfer to maintain the reality. Texture can be expressed in many forms, here we regard it as the color difference between neighboring pixels.
when s ! 10% , otherwise RGB distance is used. When the area size is smaller than 5 u 5 pixels, it’s considered as noise. For each joint sub area, at least one interaction is required to point one seed. When the color changes rapidly in the sub area, more interactions may be needed. A sample is shown in Figure 2.
Before the color transfer, we need to know the destination color. It can be obtained directly according to some existing knowledge or deduced from corresponding rules. The knowledge[4] that can be used includes the experience of the artists, the color fading rule and existing analogous murals. The color transfer algorithm[5] can be expressed as following:
H ' H a 'H ° ® S ' S b 'S °V ' V c 'V ¯
(2)
H , S ,V represent the color before transfer. H , S ,V represent the color after transfer. 'H , 'S , 'V
Where '
'
'
represent the difference between destination color and the average color of this object before transfer. a, b, c are coefficients for color transfer. Some of the murals are seriously damaged due to color fading or disease that some parts of the original arts are almost lost. Region replacement is used for these murals. It combines the figure of the replaced region and the color, texture of the replacing region together for the restoration purpose: the replacing region is warped to fit the outline of the replaced region and then filled into the replaced region.
Figure 2. Disease marking with interactive region growing
2.4
Virtual color restoration of Dunhuang mural
2.4.1 Object extracting The first step of virtual color restoration is to extract the object need restoration. The two methods mentioned in disease marking can be used for this purpose. Color separating based histogram[3] is another method.
2.5
Simulation of color evolution
The purpose of simulation the color evolution of Dunhuang mural is to provide a visible media for the communication of protection experts, and also to predict mural evolution from the current state by applying the evolution rule to the current mural.
The histogram curve of an image is usually not smooth, in order to process it effectively, it is divided into several regions according to wave crests and troughs of the histogram curve. Color values of each region are extracted from the image to generate a histogram table: Histogram˙. The forth column is the pixel count of the region. Finally the image is divided into multiple sub images according to the value of N_Pixel.
The process of simulation the color evolution[6] consists of image division according to the histogram, lengthways color fading and transverse color harmonization. The key techniques here are color fading curve and Color Harmonization.
6436
harmonization.The process of color harmonization is like following:
2.5.1 Curve of color fading The pigment of the mural experienced some chemical reactions caused by various environment factors such as sunshine, temperature, humidity and mildew. Its color changed from the original to what we see today. As the pigment fell off and dust adsorbed, the color of the mural faded. By considering the theories of color changing, fading and the drawing procedure of Dunhuang mural, the mural can be presented as following:
Ri Gi Bi o H i SiVi S
¦ Si
n
H i SVi o Ri Gi Bi
Figure 3. Color harmonization with saturation
3 3.1
x
mural consists of three layers, from bottom to top: clay layer, pigment layer and dust layer.
x
pigment and dust layers are translucent.
x
In the computer model, the image resolution is smaller than the original one.
System introduction System functions
The functions of computer aided cultural relic protection and restoration system include disease marking, disease inquiry and analysis, environment monitoring and data analysis, virtual color restoration and simulation the color evolution of Dunhuang mural etc.
Based on the above theories and mural model, two atomic color blending models can be drawn and other more complex ones can be constructed by combining these two. x
Blending model A: Two colors are blended together and place one above another.
x
Blending model B: Due to the limitation of the image resolution, multiple color points are blended into one pixel symmetrically. Figure 4. Disease inquiry
By applying the above blending models, the typical intermediate color values of the mural can be calculated. Taking these values as interpolation points to calculate a continuous color curve and limiting all the color values on the curve belong to the color space of Dunhuang, the color fading curve can be achieved. In the transform process, different colors are transformed on different color fading curve.
2.5.2 Color harmonization In the demonstration of color changing and fading, intermediate color harmonization should be considered if there are multiple colors in the same image are changing or fading. Color harmonization refers to that the color constitution of a image conforms to the principles of aesthetics harmonization. The human perception of color is determined by color's hue(H), lightness(L) and saturation(S). These three attributes are independent and constitute color cubic. When two or more colors don't harmonize, increase the same attribute of these colors can mitigate the difference. The more same attribute added, the more harmonization. Here the same attribute can be one of HLS. In the demonstration, the hue and lightness of intermediate color cannot be changed or the color curve will change, so only saturation is used for
Figure 5. Environment data analysis Users can navigate in the virtual cave just like they do in traditional work, and easily inquiry the murals with multi-resolution. Disease marking and register can be done visually during the navigation with the tools supplied by system. And also the disease records can be surveyed by many ways, such as by interested region or interested kind of disease. Sampling points for material analysis and probe points for environment data are recorded and marked in the system, and the corresponding data can be quickly surveyed in the navigation. Some analysis tools
6437
Disease inquiry module: this module supports disease inquiry in the 3D navigation.
are supplied for data analysis: the maximum, minimum and average calculators, distributing curve etc. Virtual color restoration and simulation the color evolution of Dunhuang mural are supported by the system.
3.2
Environment monitoring module: this module has the charge of environment data ingathering, inquiry, analysis and display.
System modules and workflow
The system modules of computer aided cultural relic protection and restoration consist of information management module, 3D navigation module, disease marking module, disease analysis module, disease inquiry module, environment monitoring module, virtual restoration module and virtual simulation module. The relationship of these modules is shown in Figure 6:
Virtual restoration module: this module has the charge of virtual color restoration of Dunhuang mural. Virtual simulation module: this module has the charge of virtual simulation of color evolution of Dunhuang mural. The system workflow is shown in figure 7.
Database
3D navigation
Information management
Disease marking Disease analysis
Virtual restoration
References
Virtual simulation
[1] Duan Wenjie, Corpus of Dunhuang research (cave protection), Gansu publishing company, Gansu, China, 1993.
Environment monitoring
Disease inquiry
[2] Li Zuixiong, Corpus of Cave Protection, Gansu publishing company, Gansu, China, 1994.
Figure 6. System modules Cave selection
Disease marking
[3] Wei Baogang, Pan Yunhe, “A frame and rule_based hybrid approach for color restoration of ancient mural”, Pattern recognition and artificial intelligenc, Vol 12 No.4, pp. 473-479, Dec. 1999.
Disease analysis 3D navigation
Disease inquiry Environment monitoring
[4] Wei Baogang, Pan Yunhe, Hua Zhong, “An analogy_based virtual approach for color restoration of wall painting”, Journal of computer research and development,Vol 36, No.11, pp. 1364-1368, Nov. 1999.
Virtual restoration Virtual simulation
Figure 7. System workflow
[5] Hua Zhong, Lu Dongming, Pan Yunhe, “Research on virtual color restoration and gradual changing simulation of Dunhuang fresco”, Journal of image and graphics, Vol 7(A), No. 2, pp. 181-186, Feb. 2002.
Information management module: this module has the charge of organization of all the data in the system, such as cave structure, image of mural, disease information, material analysis result and environment data etc.
[6] Lin Yi, Lu Dongming, “Technical research on virtual color shading of Dunhuang fresco”, Application research of computers, Vol 17, No. 12, pp.12-14, Dec. 2000.
3D navigation module: this module provides users a virtual cave for navigation, and also supplies some convenient user interfaces for mural protection and restoration.
[7] Wei Baogang, Lu Dongming, Pan Yunhe, “Interactive image segmentation using multiple color spaces”, Chinese journal of computer, Vol 24, No. 7, pp. 770-775, July 2001.
Disease marking module: this module supplies several manual-automatic methods for disease marking. And it uses different symbols for different kinds of disease, and also records the information such as the marking person and marked date. Disease analysis module: this module analyzes the disease according to the disease region characteristic (such as the shape and area of the disease region) and also monitors its status along the time.
6438
Hongfang Han College of Computer Science Zhejiang University Hangzhou, China [email protected]
and based on that they did scientific methods to protect and restore of the caves with right materials, and had achieved some fruits.
Abstract –The paper proposes a computer aided way for culture heritage protection and restoration, which is based on virtual navigation and image processing technologies. It first creates the database for the digital information of the cave construction and the mural, and also provides some operations, such as insert, query etc, for the database management. In a virtual environment of the cave, users can navigate, mark the disease area, inquiry the information about the mural, monitor and analyze the environment data, and also do virtual mural restoration and simulate the mural evolution without damaging the heritage itself. The article introduces the following technologies: virtual construction of the cave, knowledge expression of Mural, interactive disease marking, object extracting, color restoration with mural knowledge, color harmonization for simulating the mural evolution, and also the functions, modules, workflow of our system.
This paper proposes a computer aided way for culture heritage protection and restoration which is base on virtual navigation and image processing technologies. It helps the protection experts to mange information, monitor environment, inquiry disease, analyze data, and also do color restoration and evolution simulation. The integration of information technologies and traditional methods of protection and restoration supplies a more efficient method for Dunhuang mural which is one of the best world cultural heritages.
2 2.1
Virtual construction of the cave
Virtual cave modeling is the precondition of the computer aided cultural relic protection and restoration system, which is consisted of measurement, photograph and 3d modeling. The digitization of cave provides virtual scene not only for our system, but also for cave navigation and computer aided archaeology, and at the same time, it’s a new organization of cave construction and mural information.
Keywords: Cultural heritage digitization, protection, restoration, Dunhuang mural.
1
Key techniques
Introduction
The Mogao Caves in Dunhuang City is a treasure of Buddha Culture in China, and even in the world, which integrated with architecture, color statue and mural. It is in an importance place in the art history of China and the world.
The 3d cave model is constructed with 3dmax according to the measurement, and textured by the highresolution image of mural. The mural digitization is an important step in the process of cave digitization, which records the beautiful mural in the computer by photographing and image processing. We design a specific equipment for mural photographing, and also standardize the workflow. After mural photographing, color and shape adjustments are done to archive the highresolution digital image of mural.
Mural is the most important portion of Dunhuang art. Because of the oxidation of plumbum which is the ingredient of pigment for mural, the influence of the climate, sunlight, humidity, exhausted gas, bacterium and mildew, and destruction by people in addition for thousands years, the caves are suffering from many kinds of diseases, such as pulverization, falling off, fading, and color changing[1,2]. Latter-day experts of Dunhuang Academe did many works on environment monitoring and research on the causes of cave illness and mural fading,
0-7803-8566-7/04/$20.00 2004 IEEE This work is funded by Hi-tech Research and Development Program of China Project(No. 2003AA119020) and Zhejiang province science and technology department project(No. 2004C23035).
6434
2.2
2.3.1
Knowledge expression of mural
The knowledge of mural includes both vision character and semantic character. The vision character means those characters such as the color and layout of the mural, and the semantic character means the content and background description of the mural. The definition of semantic character is given by: Fs S No u S D u ST ,Where S No is the cave number in which the mural is, S D is the dynasty when the mural was painted, ST is the theme of the mural. The Knowledge of Dunhuang color is presented by three color rules: the rule of typical initial Dunhuang color, the rule of typical current Dunhuang color and the change rule of typical Dunhuang color.
Interactive dynamic programming
The description of disease marking by interactive dynamic programming [3] is as following: first of all, we get a weighted-graph from the mural image by defining an edge between every pair of neighboring pixels whose weight is determined by the image characteristic such as gradient. The interactive dynamic programming requires user to specify a start point and end point at edge of the disease area, then the algorithm can find the shortest path in the weighted-graph which form part of the disease edge. This procedure repeats with human’s judgment and interaction and finally can get an enclosed disease edge. The core of the algorithm is defining the weight of edges (We are using gradient). The shortest path is calculated by Dijkstra. A sample is shown in Figure 1.
Both the typical initial and current Dunhuang color rules are given by: ColorRule=<Time, Position, Value, C1, C2, Ă Ă , Cm>,Where Time is the dynasty, Position records where this color has been used, Value is given by a center point and a region, C1, C2, Ă Ă , Cm are harmonic colors of the current one expressed by Value. The change rule of typical Dunhuang color is expressed by some seriate colors according time order: ChgColorRule=, Where C1, C2,ĂĂ, Cn are the colors, Curve is the fit curve of those colors. We also use hierarchy structure to express the color knowledge of the mural. First the mural is divided into several classes such as construction, character, and decoration and so on by the mural theme. And each class is continuously divided until they are impartible color objects. The hierarchy expression gives us efficient color indices when we deal with the complicated mural color relationship.
2.3
Figure 1. Disease marking with interactive dynamic programming
2.3.2
Interactive region growing
Interactive region growing [3,7] can be also used for disease marking. The disease area here could be disjoint. Each pixel satisfying following formulas is regarded in the disease area. In the growing, the position and size of seed points are determined by the user interactively.
Interactive disease marking
In our protection and restoration system, disease marking is the foundation of disease inquiry and analysis. It could be archived completely by manual work, and actually we did disease marking manually in traditional method. But manual disease marking is a heavy work, and also the result is not satisfied in precision. On the other hand, considering of the complexity of Dunhuang mural it takes more time to find out a technique to mark the disease automatically. As a trade-off, we use manualautomatic marking methods in our system, which are acceptable in both speed and precision. Because in these manual-automatic methods, the capacity of judgment and computer are well combined.
d g 1 (C (i, j ), C ( m, n)) d G ptp °° ® d g 2 (C (i, j ), CR (i, j )) d G ptr ° °¯ d g 3 (C (i, j ), Cs (i, j )) d G pts
(1)
In the above formulas, C (i, j ) represents the color of pixel (i, j ) , C (m, n) , CR (i, j ) and Cs (i, j ) separately represent the color of neighboring pixel, the average color in the growing area and the color of the seed point. G ptp ,
6435
G ptr and G pts are thresholds and they satisfy G pts ! G ptr ! G ptp . In d gi () , the Euclidean distance of HIS is used
2.4.2
Color restoration
In Our experiments, we found some Dunhuang murals are kept rather well, but lost their original bright color in the long history. Color transfer can be used to restore the original color. Keeping the original image texture is necessary in color transfer to maintain the reality. Texture can be expressed in many forms, here we regard it as the color difference between neighboring pixels.
when s ! 10% , otherwise RGB distance is used. When the area size is smaller than 5 u 5 pixels, it’s considered as noise. For each joint sub area, at least one interaction is required to point one seed. When the color changes rapidly in the sub area, more interactions may be needed. A sample is shown in Figure 2.
Before the color transfer, we need to know the destination color. It can be obtained directly according to some existing knowledge or deduced from corresponding rules. The knowledge[4] that can be used includes the experience of the artists, the color fading rule and existing analogous murals. The color transfer algorithm[5] can be expressed as following:
H ' H a 'H ° ® S ' S b 'S °V ' V c 'V ¯
(2)
H , S ,V represent the color before transfer. H , S ,V represent the color after transfer. 'H , 'S , 'V
Where '
'
'
represent the difference between destination color and the average color of this object before transfer. a, b, c are coefficients for color transfer. Some of the murals are seriously damaged due to color fading or disease that some parts of the original arts are almost lost. Region replacement is used for these murals. It combines the figure of the replaced region and the color, texture of the replacing region together for the restoration purpose: the replacing region is warped to fit the outline of the replaced region and then filled into the replaced region.
Figure 2. Disease marking with interactive region growing
2.4
Virtual color restoration of Dunhuang mural
2.4.1 Object extracting The first step of virtual color restoration is to extract the object need restoration. The two methods mentioned in disease marking can be used for this purpose. Color separating based histogram[3] is another method.
2.5
Simulation of color evolution
The purpose of simulation the color evolution of Dunhuang mural is to provide a visible media for the communication of protection experts, and also to predict mural evolution from the current state by applying the evolution rule to the current mural.
The histogram curve of an image is usually not smooth, in order to process it effectively, it is divided into several regions according to wave crests and troughs of the histogram curve. Color values of each region are extracted from the image to generate a histogram table: Histogram˙. The forth column is the pixel count of the region. Finally the image is divided into multiple sub images according to the value of N_Pixel.
The process of simulation the color evolution[6] consists of image division according to the histogram, lengthways color fading and transverse color harmonization. The key techniques here are color fading curve and Color Harmonization.
6436
harmonization.The process of color harmonization is like following:
2.5.1 Curve of color fading The pigment of the mural experienced some chemical reactions caused by various environment factors such as sunshine, temperature, humidity and mildew. Its color changed from the original to what we see today. As the pigment fell off and dust adsorbed, the color of the mural faded. By considering the theories of color changing, fading and the drawing procedure of Dunhuang mural, the mural can be presented as following:
Ri Gi Bi o H i SiVi S
¦ Si
n
H i SVi o Ri Gi Bi
Figure 3. Color harmonization with saturation
3 3.1
x
mural consists of three layers, from bottom to top: clay layer, pigment layer and dust layer.
x
pigment and dust layers are translucent.
x
In the computer model, the image resolution is smaller than the original one.
System introduction System functions
The functions of computer aided cultural relic protection and restoration system include disease marking, disease inquiry and analysis, environment monitoring and data analysis, virtual color restoration and simulation the color evolution of Dunhuang mural etc.
Based on the above theories and mural model, two atomic color blending models can be drawn and other more complex ones can be constructed by combining these two. x
Blending model A: Two colors are blended together and place one above another.
x
Blending model B: Due to the limitation of the image resolution, multiple color points are blended into one pixel symmetrically. Figure 4. Disease inquiry
By applying the above blending models, the typical intermediate color values of the mural can be calculated. Taking these values as interpolation points to calculate a continuous color curve and limiting all the color values on the curve belong to the color space of Dunhuang, the color fading curve can be achieved. In the transform process, different colors are transformed on different color fading curve.
2.5.2 Color harmonization In the demonstration of color changing and fading, intermediate color harmonization should be considered if there are multiple colors in the same image are changing or fading. Color harmonization refers to that the color constitution of a image conforms to the principles of aesthetics harmonization. The human perception of color is determined by color's hue(H), lightness(L) and saturation(S). These three attributes are independent and constitute color cubic. When two or more colors don't harmonize, increase the same attribute of these colors can mitigate the difference. The more same attribute added, the more harmonization. Here the same attribute can be one of HLS. In the demonstration, the hue and lightness of intermediate color cannot be changed or the color curve will change, so only saturation is used for
Figure 5. Environment data analysis Users can navigate in the virtual cave just like they do in traditional work, and easily inquiry the murals with multi-resolution. Disease marking and register can be done visually during the navigation with the tools supplied by system. And also the disease records can be surveyed by many ways, such as by interested region or interested kind of disease. Sampling points for material analysis and probe points for environment data are recorded and marked in the system, and the corresponding data can be quickly surveyed in the navigation. Some analysis tools
6437
Disease inquiry module: this module supports disease inquiry in the 3D navigation.
are supplied for data analysis: the maximum, minimum and average calculators, distributing curve etc. Virtual color restoration and simulation the color evolution of Dunhuang mural are supported by the system.
3.2
Environment monitoring module: this module has the charge of environment data ingathering, inquiry, analysis and display.
System modules and workflow
The system modules of computer aided cultural relic protection and restoration consist of information management module, 3D navigation module, disease marking module, disease analysis module, disease inquiry module, environment monitoring module, virtual restoration module and virtual simulation module. The relationship of these modules is shown in Figure 6:
Virtual restoration module: this module has the charge of virtual color restoration of Dunhuang mural. Virtual simulation module: this module has the charge of virtual simulation of color evolution of Dunhuang mural. The system workflow is shown in figure 7.
Database
3D navigation
Information management
Disease marking Disease analysis
Virtual restoration
References
Virtual simulation
[1] Duan Wenjie, Corpus of Dunhuang research (cave protection), Gansu publishing company, Gansu, China, 1993.
Environment monitoring
Disease inquiry
[2] Li Zuixiong, Corpus of Cave Protection, Gansu publishing company, Gansu, China, 1994.
Figure 6. System modules Cave selection
Disease marking
[3] Wei Baogang, Pan Yunhe, “A frame and rule_based hybrid approach for color restoration of ancient mural”, Pattern recognition and artificial intelligenc, Vol 12 No.4, pp. 473-479, Dec. 1999.
Disease analysis 3D navigation
Disease inquiry Environment monitoring
[4] Wei Baogang, Pan Yunhe, Hua Zhong, “An analogy_based virtual approach for color restoration of wall painting”, Journal of computer research and development,Vol 36, No.11, pp. 1364-1368, Nov. 1999.
Virtual restoration Virtual simulation
Figure 7. System workflow
[5] Hua Zhong, Lu Dongming, Pan Yunhe, “Research on virtual color restoration and gradual changing simulation of Dunhuang fresco”, Journal of image and graphics, Vol 7(A), No. 2, pp. 181-186, Feb. 2002.
Information management module: this module has the charge of organization of all the data in the system, such as cave structure, image of mural, disease information, material analysis result and environment data etc.
[6] Lin Yi, Lu Dongming, “Technical research on virtual color shading of Dunhuang fresco”, Application research of computers, Vol 17, No. 12, pp.12-14, Dec. 2000.
3D navigation module: this module provides users a virtual cave for navigation, and also supplies some convenient user interfaces for mural protection and restoration.
[7] Wei Baogang, Lu Dongming, Pan Yunhe, “Interactive image segmentation using multiple color spaces”, Chinese journal of computer, Vol 24, No. 7, pp. 770-775, July 2001.
Disease marking module: this module supplies several manual-automatic methods for disease marking. And it uses different symbols for different kinds of disease, and also records the information such as the marking person and marked date. Disease analysis module: this module analyzes the disease according to the disease region characteristic (such as the shape and area of the disease region) and also monitors its status along the time.
6438
