Mural multi-spectral image online display platform and method

Abstract

The present application relates to a mural multispectral image online display platform and method, including three subsystems of multiscale pyramid data scheduling, multi-modal window linkage and superimposed display, digital achievement display and management. The scheduling subsystem gathers and manages pyramid slices, lightweight RGB interactive base map and metadata, dynamically loads based on the front-end engine on demand, realizes the rapid loading and smooth zooming of hundred billion pixel images on the Web, and avoids memory overflow and slow first screen. The linkage subsystem supports double window linkage, sliding comparison and transparency fusion of original images and multi-modal result layers, realizes unified superposition and efficient comparison of multispectral data. The display subsystem provides lossless zoom browsing, spectral curve specification display and operation record management, and provides intuitive and scientific collaborative support for the scientific protection of cultural heritage. Pyramid slices are used as data links to form a smooth data pipeline, providing system-level support for online display of hundred billion pixel murals.

Description

Technical Field

[0001] This invention relates to the field of digital cultural heritage protection, specifically to an online platform and method for displaying multispectral images of murals. Background Technology

[0002] Ancient murals, as precious cultural heritage, carry rich historical, artistic, and religious information. They have long faced multiple threats, including natural aging, environmental degradation leading to fading, and physical damage. High-precision digital recording and online display of murals have become crucial for cultural relic preservation and scientific research. With the development of technologies such as multispectral imaging, the data volume of digitized murals has increased dramatically. Efficiently and accurately displaying high-precision digitized murals on the web is of great significance for the digital preservation, scientific restoration, and preventative management of precious cultural relics.

[0003] Current technologies for digitally displaying murals still have significant shortcomings. On the one hand, rendering performance is limited. Ultra-high-precision mural images with billions of pixels exhibit extremely poor loading and rendering performance on the web. Existing solutions lack pyramid storage and on-demand scheduling mechanisms for large datasets, leading to browser memory overflows or excessively long initial loading times, hindering smooth, lossless zooming and severely restricting the online presentation of high-precision digital results. On the other hand, multimodal data lacks linkage. There is a lack of unified overlay and comparison mechanisms between multispectral channel data, implicit information result images, and color analysis result images. Cultural heritage researchers cannot simultaneously compare original mural images with various analysis results on the same interface, making it difficult to intuitively present the implicit information contained in multispectral data. Furthermore, spectral charts are not rigorously presented, lacking intuitiveness and scientific rigor, failing to effectively support the online collaborative research needs of cultural heritage researchers.

[0004] To address the aforementioned technical issues, there is currently no digital display technology solution for murals that can achieve ultra-fast rendering of massive amounts of data, multimodal layer overlay, and interactive visualization. There is an urgent need for an online mural multispectral image display solution that can simultaneously solve the problems of insufficient rendering performance, lack of linkage between multimodal data, and inaccurate spectral chart display. Summary of the Invention

[0005] The purpose of this invention is to provide an online display platform and method for multispectral images of murals, in order to solve the problems of insufficient loading and rendering performance of mural images with hundreds of billions of pixels on the Web, lack of unified overlay display and linkage comparison mechanism between multimodal data, and insufficient scientific rigor in the display of spectral charts in existing mural digital display technologies.

[0006] To achieve the above objectives, the following technical solution is adopted.

[0007] The online display platform for multispectral images of murals includes a multi-scale pyramid data scheduling subsystem, a multimodal window linkage and overlay display subsystem, and a digital achievement display and management subsystem. The multi-scale pyramid data scheduling subsystem is configured to aggregate and manage multi-scale pyramid slice data, lightweight RGB interactive base map data synthesized by dimensionality reduction of the top slice of the multi-scale pyramid slice data, and mural basic metadata, and to dynamically schedule and load the multi-scale pyramid slice data and the lightweight RGB interactive base map data on demand based on the front-end rendering engine. The multimodal window linkage and overlay display subsystem and the digital achievement display and management subsystem are respectively connected to the multi-scale pyramid data scheduling subsystem; The multimodal window linkage and overlay display subsystem is configured to perform simultaneous linkage of the original mural image and the multimodal analysis result layer data in two windows, sliding overlay comparison, and layer transparency fusion display; The digital achievement display and management subsystem is configured to perform lossless scaling and panning browsing of mural images with billions of pixels, standardized display of spectral curves, and management and presentation of operation records.

[0008] Optionally, the multi-scale pyramid slice data adopts a multi-band, multi-resolution hierarchical slicing structure, which is generated by multispectral mural images acquired by the acquisition device through multi-scale pyramid slicing; The multimodal analysis result layer data includes the draft line result map, ink line reconstruction result map, pigment identification distribution map, color tracing result map, and color restoration result map. Each of the multimodal analysis result layer data has a spatial registration relationship with the original mural image and is stored in the form of pyramid slices. The platform also includes top-level slice dimensionality reduction synthesis of the multi-scale pyramid slice data and lightweight RGB interactive base map data for initial front-end loading and fast preview.

[0009] Optionally, the multimodal window linkage and overlay display subsystem includes a dual-window linkage control module, a sliding overlay comparison module, a layer transparency fusion module, and a multimodal layer management module; The dual-window linkage control module is based on a front-end state machine to synchronize the translation and scaling of the left and right windows. The sliding overlay comparison module uses draggable dividing lines to display the upper and lower layers in the same window for comparison. The layer transparency blending module uses a transparency slider to display multimodal layers in a semi-transparent blending and overlay manner; The multimodal layer management module provides list-based management of multimodal result layers, and users can control the display and hiding of each layer by checking or toggling.

[0010] Optionally, the multi-scale pyramid data scheduling subsystem includes a window-aware scheduling module, a resolution adaptive module, and a progressive loading module; The window perception scheduling module is configured to load slice data within the window range as needed based on the front-end visible area, and not load slice data outside the window range; The resolution adaptive module is configured to automatically match the corresponding pyramid resolution level based on the current scaling level; The progressive loading module is configured to prioritize loading low-resolution slices when switching scaling levels, and then gradually load high-resolution slices.

[0011] Optionally, the digital achievement display and management subsystem includes a lossless browsing module with billions of pixels, a spectral curve display module, and an operation record management module; The billion-pixel lossless browsing module is based on the multi-scale pyramid slice data loaded by the multi-scale pyramid data scheduling subsystem, and supports users to perform lossless browsing on the billion-pixel mural image at any scale by using the mouse wheel to zoom and drag to pan. The spectral curve display module is configured to capture the two-dimensional coordinates of the user's click on the base map through the front end, send the two-dimensional coordinates to the back end to extract the corresponding one-dimensional spectral vector, and return them to the front end to draw the wavelength-reflectivity curve using the ECharts chart library. The minimum value of the vertical axis of the wavelength-reflectivity curve is fixed to 0, and the range of the vertical axis is dynamically determined by the maximum value of the current spectral vector. The operation record management module is configured to capture user interaction events on the details page in real time. These interaction events include mural scene switching, band carousel, color compositing, zooming, and panning operations, which are presented in chronological order. The module supports users in viewing historical operation trajectories and provides a one-click clear record function. The platform is built on the Django backend web framework, the Vue.js frontend framework, the Leaflet or Deck.gl frontend rendering engine, a relational database, and the ElementPlus frontend interface component library. The Django backend web framework is responsible for API services and the indexing and distribution of the multi-scale pyramid slice data. The relational database is used to store user data, basic metadata of cultural heritage sites, and index information of the multimodal analysis result layer data.

[0012] Optionally, the multi-scale pyramid data scheduling subsystem also includes a dynamic listener, and the platform performs a dynamic acquisition process of billions of pixel data: the front-end rendering engine by default only loads the top slice of the multi-scale pyramid slice data to generate a lightweight RGB image as an interactive base map. When the user performs zoom or pan operations, the dynamic listener calculates the boundary coordinates and resolution level of the current visible area in real time, and requests the deep high-resolution data of the corresponding spatial slice from the backend as needed for loading and stitching. The window perception scheduling module only loads slice data within the current window range, and does not load slice data outside the window range; The resolution adaptive module automatically matches the corresponding pyramid resolution level based on the current scaling level; The progressive loading module first loads low-resolution slices when switching scaling levels, and then gradually loads high-resolution slices. The multi-scale pyramid slice data is stored in a distributed storage terminal and dynamically acquired by the front-end rendering engine based on the current viewport range and scaling level.

[0013] Optionally, the dual-window linkage control module is equipped with a single linkage control button. When the single linkage control button is activated, the front-end state machine captures the change state of either the left or right window and applies the change matrix synchronously to the other window in real time, maintaining absolute synchronous linkage between the left and right windows. The sliding overlay comparison module divides the upper layer and the lower layer into left and right sections within a single window. The draggable dividing line displays the lower layer on the left and the upper layer on the right. Users can control the division position in real time by dragging the draggable dividing line with the mouse. The transparency slider of the layer transparency blending module is continuously adjustable within the range of 0% to 100%, adjusting the transparency parameters of the upper layer in real time. The multimodal layer management module displays all available multimodal result layers at the current data collection location in a list format. Users can control the display and hiding of each layer by checking or toggling.

[0014] The method for online display of multispectral images of murals includes the following steps: The process involves acquiring multi-scale pyramid slice data formed by storing original high-resolution mural images sliced ​​according to pyramid levels, and lightweight RGB interactive base map data synthesized by dimensionality reduction of the top slice of the multi-scale pyramid slice data. The lightweight RGB interactive base map data is then loaded to generate an initial display screen, resulting in a front-end rendering result containing the initial display screen. Monitor the scaling or translation operations in the front-end rendering results, calculate the boundary coordinates and resolution level of the current visible area in real time, and request the deep high-resolution data of the corresponding spatial slices from the multi-scale pyramid data scheduling subsystem on demand for loading and stitching based on the boundary coordinates and resolution level. When switching scaling levels, load low-resolution slices first and then gradually load high-resolution slices to obtain the slice rendering results after on-demand scheduling. Based on the slice rendering results, each result layer in the multimodal analysis result layer data is superimposed on the slice rendering results according to the spatial registration relationship with the original mural image. The display and hiding of each result layer are controlled by the check or switch operation of the multimodal layer management module to obtain the multimodal layer superimposed display result. Based on the multimodal layer overlay display result, the front-end state machine captures the transformation state of either the left or right window, and applies the transformation matrix synchronously to the other window in real time. The upper and lower layers are then compared or semi-transparently blended and overlaid using draggable dividing lines and transparency sliders to obtain the dual-window linkage and comparison display result. In response to a click operation on the target position in the slice rendering result, the two-dimensional coordinates of the target position are sent to the backend to extract the corresponding one-dimensional spectral vector. A wavelength-reflectivity curve is plotted based on the ECharts chart library, and the minimum value of the vertical axis of the wavelength-reflectivity curve is fixed to 0. The range of the vertical axis is dynamically determined by the maximum value of the current spectral vector, resulting in the spectral curve display result. The zoom, pan, click, check or switch, drag, and adjust operations are recorded as interactive operation events, arranged in chronological order, and can be cleared with one click to obtain the operation record management results.

[0015] Optionally, when monitoring scaling or translation operations in the front-end rendering results and calculating the boundary coordinates and resolution level of the current visible area in real time, the dynamic listener of the multi-scale pyramid data scheduling subsystem performs the real-time calculation, sends the boundary coordinates to the window perception scheduling module, and sends the resolution level to the resolution adaptation module. The window perception scheduling module requests deep high-resolution data of the corresponding spatial slice from the backend on demand based on the boundary coordinates and loads and stitches it. The resolution adaptive module matches the corresponding pyramid resolution level based on the resolution level. The progressive loading module loads low-resolution slices first and then gradually loads high-resolution slices when switching pyramid resolution levels. The multi-scale pyramid slice data is stored in a distributed storage terminal and dynamically retrieved on demand by the front-end rendering engine based on the current viewport range and scaling level.

[0016] Optionally, in response to a click operation on the target location in the slice rendering result, the front end captures the two-dimensional coordinates of the target location and sends the two-dimensional coordinates of the target location to the back end. The back end extracts a one-dimensional spectral vector based on the spatial location corresponding to the two-dimensional coordinates of the target location in the multi-scale pyramid slice data. The one-dimensional spectral vector is a sequence of reflectance values ​​of the mural pixels corresponding to the spatial location in each spectral band. After the one-dimensional spectral vector is returned to the front end, the ECharts chart library draws a wavelength-reflectance curve. The minimum value of the vertical axis of the wavelength-reflectance curve is fixed to 0, and the range of the vertical axis is dynamically determined by the maximum value of the current spectral vector. When the zoom, pan, click, check or switch, drag, and adjust operations are recorded as interactive operation events, the operation record management module captures the occurrence time and operation type of each interactive operation event in real time. The interactive operation events also include mural scene switching, band carousel, and color synthesis. The interactive operation events are arranged and presented in chronological order according to their occurrence time. Users can view historical operation trajectories and a one-click clear record function is provided, which clears all interactive operation events.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The online mural multispectral image display platform provided by this invention effectively solves the technical problems mentioned in the background art through the coordinated operation of a multi-scale pyramid data scheduling subsystem, a multi-modal window linkage and overlay display subsystem, and a digital achievement display and management subsystem. The multi-scale pyramid data scheduling subsystem aggregates and manages multi-scale pyramid slice data, lightweight RGB interactive base map data synthesized from the top-level slices of the multi-scale pyramid slice data through dimensionality reduction, and basic metadata of the mural. Based on a front-end rendering engine, it dynamically schedules and loads the multi-scale pyramid slice data and lightweight RGB interactive base map data on demand, enabling the rapid loading and smooth scaling of mural images with billions of pixels on the web, overcoming the browser memory overflow and excessively long first-screen loading time problems caused by the lack of a large data volume scheduling mechanism in existing technologies. The multi-modal window linkage and overlay display subsystem realizes the display of the original mural image... Features such as dual-window synchronous linkage with multimodal analysis result layers, sliding overlay comparison, and layer transparency fusion display enable unified overlay and efficient comparison of multimodal data, including multispectral channel data, implicit information result images, and color analysis result images, within the same interface. This overcomes the lack of linkage mechanisms in existing technologies for multimodal data. The digital results display and management subsystem enables lossless scaling and panning of mural images with billions of pixels, standardized display of spectral curves, and management and presentation of operation records. This makes the presentation of analysis results more intuitive and scientific, providing rigorous online collaborative research support for cultural heritage researchers. Simultaneously, the multimodal window linkage and overlay display subsystem and the digital results display and management subsystem are connected to the multi-scale pyramid data scheduling subsystem, forming a data pipeline linked by multi-scale pyramid slice data. Data flow between subsystems is smooth, and the overall platform is coordinated and orderly, providing system-level support for the online display of mural images with billions of pixels.

[0018] Furthermore, the multi-scale pyramid slice data adopts a multi-band, multi-resolution hierarchical slicing structure. The multispectral mural images acquired by the acquisition equipment are generated through multi-scale pyramid slicing. While preserving the complete information of each spectral band, the data is organized and stored hierarchically according to resolution. Combined with the lightweight RGB interactive base map data synthesized by dimensionality reduction of the top slice, the platform can quickly present the overall image during the initial loading stage. Subsequent acquisition of deep high-resolution slices is carried out as needed, balancing the speed of image presentation and the fidelity of image details. The multimodal analysis result layer data includes the draft line result map, ink line reconstruction result map, pigment identification distribution map, color tracing result map, and color restoration result map. Each result layer has a spatial registration relationship with the original mural image and is stored in the form of pyramid slices, so that various analysis results can be accurately aligned and superimposed with the original image, providing a spatially consistent display basis for multimodal comparison. The multimodal window linkage and overlay display subsystem features a dual-window linkage control module that uses a front-end state machine to synchronize the translation and scaling of the left and right windows. A single linkage control button triggers the front-end state machine to capture the transformation state of any window and synchronously apply the transformation matrix to the other window in real time, achieving absolute synchronization between the two windows. Combined with draggable dividing lines for left-right segmentation and a continuously adjustable transparency slider from 0% to 100% for semi-transparent fusion and overlay, it provides rich interactive methods for differential comparison of multimodal data. The multimodal layer management module displays all available multimodal result layers from the collection location in a list format, allowing flexible control of the display and hiding of each layer through checkboxes or toggles. The multi-scale pyramid data scheduling subsystem is equipped with a window-aware scheduling module, a resolution adaptive module, and a progressive loading module. These modules respectively enable on-demand loading of slice data within the window, automatic matching of scaling levels with pyramid resolution levels, and progressive loading from low to high when switching scaling levels. In addition, a dynamic listener calculates the boundary coordinates and resolution levels of the visible area in real time during scaling or panning operations, and requests deep high-resolution data of the corresponding spatial slices from the distributed storage terminal on demand. This ensures that the loading volume of massive amounts of data always matches the visual needs, fully utilizes system resources, and seamlessly connects image presentation with detail loading. The spectral curve display module of the digital achievement display and management subsystem sends the two-dimensional coordinates corresponding to the user's click location to the backend to extract a one-dimensional spectral vector. The ECharts chart library then draws a wavelength-reflectivity curve, fixing the minimum value of the ordinate of the wavelength-reflectivity curve to 0, while dynamically determining the range of the ordinate based on the maximum value of the current spectral vector. This ensures that the spectral data is presented in a standardized manner under a unified coordinate benchmark, providing a scientific and rigorous display standard for comparing spectral data from different locations and at different times. The operation record management module captures interactive operation events such as mural scene switching, band rotation, color synthesis, zooming, and panning in real time, presenting them in chronological order and providing a one-click clear function, making the research process traceable and reproducible.The platform is built upon the Django backend web framework, the Vue.js frontend framework, the Leaflet or Deck.gl frontend rendering engine, a relational database, and the ElementPlus frontend interface component library. The Django backend web framework is responsible for API services and the indexing and distribution of multi-scale pyramid slice data. The relational database stores user data, basic metadata of cultural heritage sites, and index information of multimodal analysis result layer data, achieving efficient separation between the frontend and backend and providing architectural support for the platform's stable operation and functional expansion. Based on this, the invention also provides an online display method for mural multispectral images. This method involves six tightly integrated steps: acquiring multi-scale pyramid slice data and lightweight RGB interactive base map data to generate an initial display screen; monitoring zoom or pan operations and scheduling deep, high-resolution data as needed; overlaying multimodal analysis result layers according to spatial registration relationships; implementing dual-window linkage and contrast fusion based on a frontend state machine; extracting one-dimensional spectral vectors and drawing spectral curves in response to click operations; and arranging interactive operation events in chronological order with one-click clearing support. This method fully implements the platform's functions at the methodological level, with each step progressing progressively based on the results of the previous step, providing a complete and executable technical path for the online display of multispectral mural images. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a module of an embodiment of the online mural multispectral image display platform of the present invention.

[0020] Figure 2 This is a schematic flowchart illustrating the steps of an embodiment of the online display method for multispectral images of murals according to the present invention. Detailed Implementation

[0021] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0022] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0023] The online display platform and method for multispectral images of murals provided by this invention are applied to the fields of digital cultural heritage protection, image processing, and geographic information. Specifically, it is used to display the multispectral digitization results of murals online in a web browser, enabling rapid loading and smooth scaling of mural images with billions of pixels, as well as the overlay display and synchronous comparison of multispectral data and multimodal analysis result layers. The platform runs between the front-end browser environment and the back-end server environment, and completes data scheduling, layer display, and interactive management through the collaborative cooperation between the front-end and the back-end. Example 1

[0024] like Figure 1 As shown, Example 1 provides an online display platform for multispectral images of murals, including a multi-scale pyramid data scheduling subsystem, a multimodal window linkage and overlay display subsystem, and a digital achievement display and management subsystem. These three subsystems respectively undertake three types of responsibilities: data scheduling, linkage display, and achievement management. The multi-scale pyramid data scheduling subsystem is responsible for the aggregation, management, and on-demand scheduling of the platform's underlying data, serving as the platform's data foundation. The multimodal window linkage and overlay display subsystem is responsible for the visualization and interactive linkage of multimodal data, serving as the platform's core display component. The digital achievement display and management subsystem is responsible for application-oriented achievement browsing, spectral analysis, and operation management, serving as the platform's achievement output. The multi-scale pyramid data scheduling subsystem is used to aggregate and manage multi-scale pyramid slice data, lightweight RGB interactive base map data synthesized from the top-level slices of the multi-scale pyramid slice data through dimensionality reduction, and basic mural metadata. Based on a front-end rendering engine, it dynamically schedules and loads the multi-scale pyramid slice data and lightweight RGB interactive base map data on demand. The front-end rendering engine is the front-end component in the platform responsible for loading, stitching, and rendering slice data. The basic mural metadata includes basic information such as the cultural heritage location to which the mural belongs and its acquisition records. The multimodal window linkage and overlay display subsystem and the digital achievement display and management subsystem are respectively connected to the multi-scale pyramid data scheduling subsystem. These three subsystems form a complete data pipeline with multi-scale pyramid slice data as the data link. The multi-scale pyramid slice data scheduled and loaded by the multi-scale pyramid data scheduling subsystem flows to both the multimodal window linkage and overlay display subsystem and the digital achievement display and management subsystem for display and processing. The multimodal window linkage and overlay display subsystem enables synchronous linkage, sliding overlay comparison, and layer transparency fusion display of the original mural image and the multimodal analysis result layer data. The original mural image is rendered from multi-scale pyramid slice data, and the multimodal analysis result layer data is the analysis result layer data generated during the mural's digital analysis process. The digital achievement display and management subsystem enables lossless scaling and panning of mural images with billions of pixels, standardized display of spectral curves, and management and presentation of operation records.

[0025] Multi-scale pyramid slice data employs a multi-band, multi-resolution hierarchical slicing structure, generated from multispectral mural images acquired by an acquisition device through multi-scale pyramid slicing. The acquisition device performs multispectral imaging on the mural, obtaining a raw high-resolution image containing multiple spectral bands. Each spectral band corresponds to the reflectance information of the mural surface within a specific wavelength range, and the reflectance information from multiple spectral bands collectively constitutes the multispectral image data of the mural. The raw high-resolution mural image is processed by multi-scale pyramid slicing, and sliced ​​and stored according to pyramid levels to form multi-scale pyramid slice data. Multi-scale pyramid slice data possesses organizational characteristics in two dimensions: multi-band and multi-resolution. Image information from each spectral band is fully preserved at each resolution level, with the resolution level increasing progressively from the top to the bottom. The top slice corresponds to the lowest resolution level and has the smallest data volume, while the bottom slice corresponds to the highest resolution level and has the largest data volume. The pyramid levels of the multi-scale pyramid slice data correspond to the spatial range of the mural image; the same spatial range corresponds to slice data of different resolutions at different resolution levels. When displaying mural images at different scaling scales, the front end retrieves the slice data of the corresponding spatial range from the corresponding resolution level. The multimodal analysis results layer data includes the draft line result map, ink line reconstruction result map, pigment identification distribution map, color tracing result map, and color restoration result map. The draft line result map corresponds to the analysis results of the extracted draft lines of the mural, reflecting the distribution information of the draft lines in the early stage of mural painting; the ink line reconstruction result map corresponds to the analysis results of the ink line reconstruction of the mural, reflecting the reconstruction information of the ink line outline of the mural; the pigment identification distribution map corresponds to the analysis results of the pigment identification of the mural, reflecting the spatial distribution information of the pigment components on the surface of the mural; the color tracing result map corresponds to the analysis results of the color tracing of the mural, reflecting the analysis information of the source and evolution of the mural's colors; and the color restoration result map corresponds to the analysis results of the color restoration of the mural, reflecting the analysis information of the restoration of the original colors of the mural. The five types of result maps reflect the analysis information of the mural in different dimensions. Each multimodal analysis result layer has a spatial registration relationship with the original mural image. The spatial registration relationship means that the pixel position of each result layer corresponds one-to-one with the pixel position of the original mural image in space and is precisely aligned. The multimodal analysis result layer data is stored in the form of pyramid slices and is displayed on demand by the front end. When displaying the results, the front end overlays the result layers on the original mural image according to the spatial registration relationship of each result layer, so that the analysis information reflected by the result layers corresponds accurately with the spatial position of the mural image.The platform also includes lightweight RGB interactive basemap data, which is synthesized from the top-level slices of the multi-scale pyramid slice data through dimensionality reduction. The top-level slices have the lowest resolution and the smallest data volume. The top-level slices are dimensionality reduced and synthesized into lightweight RGB three-channel images to form lightweight RGB interactive basemap data. Lightweight RGB interactive basemap data is used for initial front-end loading and quick preview, allowing users to immediately see the overall image of the mural when entering the platform page. After being loaded by the front-end, the lightweight RGB interactive basemap data is rendered into a lightweight RGB image, which is presented as the interactive basemap on the platform page. The interactive basemap and the mural image rendered from the multi-scale pyramid slice data have the same spatial range. The front-end performs zoom, pan, and click operations on the interactive basemap.

[0026] The multi-scale pyramid data scheduling subsystem includes a window-aware scheduling module, a resolution adaptive module, and a progressive loading module. The window-aware scheduling module loads slice data within the visible area of ​​the front-end as needed, excluding slice data outside the visible area. The visible area is the image region currently visible to the user in the browser page. The window-aware scheduling module only requests slice data covered by the visible area, avoiding loading data from areas invisible to the user, thus controlling the amount of data requested per request. The resolution adaptive module automatically matches the corresponding pyramid resolution level based on the current zoom level. When the user zooms in on the mural image, the current zoom level changes dynamically with the zoom operation. The resolution adaptive module matches the current zoom level with the resolution levels of the multi-scale pyramid slice data, selecting slice data corresponding to the resolution level appropriate for the current zoom level. The larger the current zoom level, the deeper the corresponding pyramid resolution level. The progressive loading module prioritizes loading low-resolution slices during zoom level switching to quickly render the image, then gradually loads high-resolution slices to achieve progressive image sharpening. During zooming, the user first sees the image rendered from low-resolution slices, followed by the gradual loading of high-resolution slices, revealing details step by step. The view-aware scheduling module, resolution adaptive module, and progressive loading module work together to form the on-demand scheduling mechanism of the multi-scale pyramid data scheduling subsystem. These three modules control the acquisition of slice data from three dimensions: spatial range, resolution level, and loading order. The view-aware scheduling module determines which spatial ranges of slice data to acquire, the resolution adaptive module determines which resolution level to acquire, and the progressive loading module determines the loading order of the slice data.

[0027] The multi-scale pyramid data scheduling subsystem also includes a dynamic listener, enabling the platform to dynamically acquire billions of pixel data points. By default, the front-end rendering engine only loads the top-level slice of the multi-scale pyramid slice data, generating a lightweight RGB image as an interactive base map. This interactive base map is the foundational image that users directly see on the platform page and use for interactive operations such as zooming, panning, and clicking. During the initial loading phase, the platform only needs to acquire a small amount of data from the top-level slice to complete the rendering of the interactive base map, eliminating the need to load all high-resolution data at once. A dynamic listener is configured at the platform's underlying layer. This listener continuously monitors the user's zoom and pan operations, triggering real-time calculations only when such operations are detected; otherwise, it maintains the existing slice data within the current viewport. When the user performs a zoom or pan operation, the dynamic listener calculates the boundary coordinates and resolution level of the current visible area in real time. It then requests the corresponding deep, high-resolution spatial slice data from the backend as needed for loading and stitching. The boundary coordinates determine the location range of the current visible area in the mural image space, and the resolution level determines the pyramid level appropriate for the current zoom level. The backend returns the corresponding slice data based on the boundary coordinates and resolution level, and the frontend rendering engine stitches the returned slice data into a complete image of the visible area. The viewport-aware scheduling module only loads slice data within the current viewport; slice data outside the viewport is not loaded. The resolution adaptive module automatically matches the corresponding pyramid resolution level based on the current zoom level. The progressive loading module loads low-resolution slices first, then gradually loads high-resolution slices when switching zoom levels. All tile data is stored in a distributed storage system. The front-end rendering engine dynamically retrieves the data on demand based on the current viewport size and zoom level, implementing lazy loading. The distributed storage system is a distributed storage facility that stores all tile data. The front-end rendering engine retrieves the corresponding tile data from the distributed storage system when needed. The distributed storage system responds to front-end requests in an on-demand manner, transmitting the tile data for the corresponding range only when a request occurs. A relational database is used to store user data, basic metadata of cultural heritage sites, and index information for multimodal result layers. The index information is used to locate the storage position of the multimodal result layer data in the multi-scale pyramid tile data. When the front-end overlays and displays the multimodal result layers, it retrieves the corresponding result layer data from the distributed storage system based on the index information.

[0028] The multimodal viewport linkage and overlay display subsystem is based on front-end state machine control, constructing a dual-view linkage comparison model to achieve synchronous linkage display and sliding overlay comparison of multimodal data in two views. The multimodal viewport linkage and overlay display subsystem includes a dual-viewport linkage control module, a sliding overlay comparison module, a layer transparency blending module, and a multimodal layer management module. The dual-viewport linkage control module, based on the front-end state machine, performs synchronous linkage of translation and scaling of the left and right views. The front-end state machine is a front-end state management component that manages the transformation states of the views, maintaining transformation state parameters such as the translation amount and scaling ratio of the left and right views. The left and right views are used to display the original mural image and the multimodal analysis result layer data, respectively. The original mural image is rendered from multi-scale pyramid slice data, and the multimodal analysis result layer data is overlaid on the original mural image according to spatial registration relationships before display. The dual-window linkage control module is equipped with a single linkage control button. When the single linkage control button is activated, the front-end state machine captures the transformation state of either the left or right window and applies the transformation matrix synchronously to the other window in real time, maintaining absolute synchronization between the left and right windows. The transformation matrix represents the translation and scaling transformation parameters of the window. After the transformation matrix is ​​synchronously applied to the other window, the other window translates and scales according to the same transformation parameters, thus ensuring absolute synchronization between the two windows. When the user performs a scaling or translation operation on either window, the other window synchronously performs the same scaling or translation with the same transformation parameters, and the two windows always present the same field of view. The sliding overlay comparison module uses a draggable dividing line to display the upper and lower layers in a single window for comparison. Within a single window, the module divides the upper and lower layers horizontally. The left side of the draggable dividing line displays the lower layer, and the right side displays the upper layer. Users can control the division position in real-time by dragging the dividing line with the mouse, allowing for horizontal comparison. As the dividing line is dragged, the display area of ​​the layers on the left and right sides of the dividing line changes accordingly, enabling users to perform position-by-position sliding comparisons between the upper and lower layers. The layer transparency blending module uses a transparency slider to achieve a semi-transparent blending and overlay display of multimodal layers. The transparency slider is continuously adjustable within the range of 0% to 100%. Users can adjust the transparency slider in real time to adjust the transparency parameter of the upper layer, achieving a semi-transparent blending and overlay display of multimodal layers. When the transparency of the upper layer is 0%, the upper layer is completely transparent and only the lower layer is displayed. When the transparency of the upper layer is 100%, the upper layer is completely opaque and completely covers the lower layer. Any value of the transparency parameter within the range of 0% to 100% corresponds to different display ratios of the upper and lower layers.The multimodal layer management module provides list-based management of multimodal result layers. The module displays all available multimodal result layers at the current collection location in a list format. Users can control the display and hiding of each layer by checking or toggling, enabling flexible overlay display of multimodal data. The checked result layers are overlaid on the mural image, while the unchecked result layers are hidden. Multiple checked result layers are overlaid and displayed according to the stacking order set by the multimodal layer management module.

[0029] The digital achievement display and management subsystem provides application-side functions such as lossless browsing at the billion-pixel level, spectral curve display, and operation record management. This subsystem includes a billion-pixel lossless browsing module, a spectral curve display module, and an operation record management module. The billion-pixel lossless browsing module utilizes multi-scale pyramid slice data scheduled and loaded by the multi-scale pyramid data scheduling subsystem. It allows users to zoom and drag on the billion-pixel mural image at any scale without loss of detail. Lossless browsing means that the image does not become pixelated or blurred during zooming; the image clarity increases synchronously with the zoom level. When the user zooms in, the platform retrieves slice data from a deeper resolution level, increasing image detail; when the user zooms out, the platform retrieves slice data from a shallower resolution level, clearly presenting the overall image outline. The spectral curve display module captures the two-dimensional coordinates of user clicks on the base map through the front end, sends the two-dimensional coordinates to the back end to extract the corresponding one-dimensional spectral vector. The one-dimensional spectral vector is a sequence of reflectance values ​​for the corresponding spatial location of the mural in each spectral band. After the one-dimensional spectral vector is returned to the front end, the ECharts chart library draws a wavelength-reflectance curve. The horizontal axis of the wavelength-reflectance curve corresponds to wavelength, and the vertical axis corresponds to reflectance. The minimum value of the vertical axis of the wavelength-reflectance curve is fixed to 0, and the range of the vertical axis is dynamically determined by the maximum value of the current spectral vector. This allows the spectral curves of different points and different periods to be presented in a standardized manner under the same coordinate reference. Users can obtain the wavelength-reflectance curve of the corresponding location by clicking on different positions on the mural image, which can be used for spectral analysis such as pigment identification. The operation log management module captures user interaction events on the details page in real time. These events include mural scene switching, band carousel, color compositing, zooming, and panning. Mural scene switching involves switching between different mural displays, including grottoes, temples, and tombs. Band carousel displays images in different spectral bands. Color compositing displays images created by combining multiple spectral bands. The operation log management module presents these interaction events in chronological order, allows users to view historical operation records, and provides a one-click clear record function. Once the one-click clear record function is triggered, all interaction events are cleared.

[0030] The platform's overall architecture is as follows: Django is used as the backend web framework. The Django backend web framework is responsible for API services and the distribution of multi-scale pyramid slice data indexes. The frontend initiates data requests to the backend through the API service. The backend distributes the multi-scale pyramid slice data indexes to the frontend according to the requests and provides backend data processing services such as spectral vector extraction. Vue.js is used as the frontend framework to build component-based large-screen visualization pages. Each visualization component is separated by function, independently developed, and reused. Leaflet or Deck.gl is used as the frontend rendering engine to dynamically load, stitch, and render multi-scale pyramid slice data. The frontend rendering engine requests slice data from the distributed storage end according to the current viewport range and zoom level, and stitches the returned slice data into a complete image according to spatial location. A relational database stores user data, basic metadata of cultural heritage sites, and index information of multimodal result layers. ElementPlus is used to build the frontend interface, and ECharts is used to build spectral curve charts. This system achieves efficient separation between the front-end and back-end using Django and Vue.js. The front-end is responsible for page rendering, interactive responses, and loading of sliced ​​data, while the back-end is responsible for data services such as data indexing and spectral vector extraction. The front-end rendering engine dynamically retrieves sliced ​​data from the distributed storage on demand based on the current viewport range and zoom level, while the back-end is responsible for data indexing, distribution, and spectral vector extraction. Example 2

[0031] like Figure 2 As shown, Example 2 provides a method for online display of multispectral images of murals. The following steps are performed using an online display platform for multispectral images of murals. The structure of the online display platform for multispectral images of murals is the same as that in Example 1, and each step is executed collaboratively between the front end and the back end of the platform.

[0032] The system acquires multi-scale pyramid slice data, formed by slicing the original high-resolution mural image into pyramid-level slices, and lightweight RGB interactive basemap data synthesized from the top-level slices of the multi-scale pyramid slice data through dimensionality reduction. It then loads the lightweight RGB interactive basemap data to generate the initial display screen, resulting in a front-end rendering result containing this initial display screen. During the initial loading phase, the platform only loads the top-level slices of the multi-scale pyramid slice data, dimensionality-reduced and synthesized them into lightweight RGB interactive basemap data, and loads this data to generate the initial display screen. The initial display screen presents the overall outline of the mural at a lower resolution, allowing users to see the complete image upon entering the platform page. After loading, the lightweight RGB interactive basemap data is rendered into a lightweight RGB image, which is then presented as the interactive basemap in the initial display screen.

[0033] The system monitors zooming or panning operations in the front-end rendering results, calculates the boundary coordinates and resolution level of the current visible area in real time, and requests deep high-resolution data of the corresponding spatial slices from the multi-scale pyramid data scheduling subsystem on demand for loading and stitching. When switching zoom levels, low-resolution slices are loaded first, followed by progressively loading high-resolution slices to obtain the slice rendering result after on-demand scheduling. When the user performs zooming or panning operations, the dynamic listener deployed at the bottom layer of the platform calculates the boundary coordinates and resolution level of the current visible area in real time, sends the boundary coordinates to the window-aware scheduling module, and sends the resolution level to the resolution adaptation module. The window-aware scheduling module only requests slice data within the window range based on the boundary coordinates, and does not request slice data outside the window range. The resolution adaptation module matches the pyramid resolution level corresponding to the current zoom level. When the pyramid resolution level switches, the progressive loading module first loads low-resolution slices to quickly render the image, and then progressively loads high-resolution slices to achieve progressive image sharpening. The front-end rendering engine loads and stitches the deep high-resolution data obtained on demand to obtain the slice rendering result after on-demand scheduling.

[0034] Based on the slice rendering results, each result layer in the multimodal analysis result layer data is superimposed on the slice rendering results according to its spatial registration relationship with the original mural image. The display and hiding of each result layer can be controlled by the selection or switching operation of the multimodal layer management module, resulting in a multimodal layer superimposed display result. The multimodal analysis result layer data includes the draft line result image, ink line reconstruction result image, pigment identification distribution map, color tracing result image, and color restoration result image. Each result layer has a spatial registration relationship with the original mural image. During superimposition, each result layer is precisely aligned with the original mural image according to the spatial registration relationship. Users can control the display and hiding of each result layer through the list selection or switching operation of the multimodal layer management module, realizing flexible superimposed display of multimodal data. The selected result layers are superimposed on the mural image, while the unselected result layers are hidden.

[0035] Based on the multimodal layer overlay display results, the front-end state machine captures the transformation state of either the left or right view, and applies the transformation matrix synchronously to the other view in real time. The upper and lower layers are then separated and compared left and right or semi-transparently blended and overlaid using draggable dividing lines and transparency sliders, resulting in a dual-view linkage and contrast fusion display result. The dual-window linkage control module is equipped with a single linkage control button. When the single linkage control button is activated, the front-end state machine captures the transformation state of either the left or right window and applies the transformation matrix synchronously to the other window in real time, maintaining absolute synchronization between the left and right windows. Within a single window, a dividing line can be dragged to separate the upper and lower layers for left and right display. The lower layer is displayed on the left and the upper layer on the right. Users can drag the dividing line with the mouse to control the division position in real time and slide left and right to compare. The transparency slider is continuously adjustable from 0% to 100%. Users can adjust the transparency slider to adjust the transparency parameter of the upper layer in real time, realizing the semi-transparent blending and overlay display of multi-modal layers.

[0036] In response to a click on a target location in the slice rendering result, the two-dimensional coordinates of the target location are sent to the backend to extract the corresponding one-dimensional spectral vector. A wavelength-reflectance curve is then plotted using the ECharts chart library, with the minimum value of the ordinate of the wavelength-reflectance curve fixed at 0, and the range of the ordinate dynamically determined by the maximum value of the current spectral vector, resulting in the spectral curve display. When a user clicks on a target location in the slice rendering result, the frontend captures the two-dimensional coordinates of the target location and sends them to the backend. The backend extracts a one-dimensional spectral vector based on the spatial location corresponding to the two-dimensional coordinates in the multi-scale pyramid slice data. The one-dimensional spectral vector is a sequence of reflectance values ​​for the mural pixels corresponding to the spatial location in each spectral band. The backend returns the one-dimensional spectral vector to the frontend, which then plots a wavelength-reflectance curve using the ECharts chart library, fixing the minimum value of the ordinate of the wavelength-reflectance curve at 0, and dynamically determining the range of the ordinate by the maximum value of the current spectral vector, resulting in the spectral curve display.

[0037] Zooming, panning, clicking, checking or toggling, dragging, and adjusting operations are recorded as interactive event records. These events are presented in chronological order and can be cleared with a single click, resulting in a comprehensive operation record management system. The module captures the occurrence time and operation type of each interactive event in real time. Interactive events also include mural scene switching, band rotation, and color compositing. The module arranges these events chronologically, allowing users to view historical operation data and providing a one-click clear function to erase all interactive events.

[0038] As is known from common technical knowledge, this invention can be implemented through other embodiments that do not depart from its spirit or essential characteristics. Therefore, the disclosed embodiments described above are merely illustrative in all respects and are not the only ones. All modifications within the scope of this invention or its equivalents are included in this invention.

Claims

1. An online display platform for multispectral images of murals, characterized in that, This includes a multi-scale pyramid data scheduling subsystem, a multi-modal window linkage and overlay display subsystem, and a digital achievement display and management subsystem; The multi-scale pyramid data scheduling subsystem is configured to aggregate and manage multi-scale pyramid slice data, lightweight RGB interactive base map data synthesized by dimensionality reduction of the top slice of the multi-scale pyramid slice data, and mural basic metadata, and to dynamically schedule and load the multi-scale pyramid slice data and the lightweight RGB interactive base map data on demand based on the front-end rendering engine. The multimodal window linkage and overlay display subsystem and the digital achievement display and management subsystem are respectively connected to the multi-scale pyramid data scheduling subsystem; The multimodal window linkage and overlay display subsystem is configured to perform simultaneous linkage of the original mural image and the multimodal analysis result layer data in two windows, sliding overlay comparison, and layer transparency fusion display; The digital achievement display and management subsystem is configured to perform lossless scaling and panning browsing of mural images with billions of pixels, standardized display of spectral curves, and management and presentation of operation records.

2. The online display platform for multispectral images of murals according to claim 1, characterized in that, The multi-scale pyramid slice data adopts a multi-band, multi-resolution hierarchical slicing structure, which is generated by multispectral images of murals acquired by the acquisition device through multi-scale pyramid slicing. The multimodal analysis result layer data includes the draft line result map, ink line reconstruction result map, pigment identification distribution map, color tracing result map, and color restoration result map. Each of the multimodal analysis result layer data has a spatial registration relationship with the original mural image and is stored in the form of pyramid slices. The platform also includes top-level slice dimensionality reduction synthesis of the multi-scale pyramid slice data and lightweight RGB interactive base map data for initial front-end loading and fast preview.

3. The online display platform for multispectral images of murals according to claim 1, characterized in that, The multimodal window linkage and overlay display subsystem includes a dual-window linkage control module, a sliding overlay comparison module, a layer transparency fusion module, and a multimodal layer management module; The dual-window linkage control module is based on a front-end state machine to synchronize the translation and scaling of the left and right windows. The sliding overlay comparison module uses draggable dividing lines to display the upper and lower layers in the same window for comparison. The layer transparency blending module uses a transparency slider to display multimodal layers in a semi-transparent blending and overlay manner; The multimodal layer management module provides list-based management of multimodal result layers, and users can control the display and hiding of each layer by checking or toggling.

4. The online display platform for multispectral images of murals according to claim 1, characterized in that, The multi-scale pyramid data scheduling subsystem includes a window-aware scheduling module, a resolution adaptive module, and a progressive loading module. The window perception scheduling module is configured to load slice data within the window range as needed based on the front-end visible area, and not load slice data outside the window range; The resolution adaptive module is configured to automatically match the corresponding pyramid resolution level based on the current scaling level; The progressive loading module is configured to prioritize loading low-resolution slices when switching scaling levels, and then gradually load high-resolution slices.

5. The online display platform for multispectral images of murals according to claim 1, characterized in that, The digital achievement display and management subsystem includes a lossless browsing module with billions of pixels, a spectral curve display module, and an operation record management module. The billion-pixel lossless browsing module is based on the multi-scale pyramid slice data loaded by the multi-scale pyramid data scheduling subsystem, and supports users to perform lossless browsing on the billion-pixel mural image at any scale by using the mouse wheel to zoom and drag to pan. The spectral curve display module is configured to capture the two-dimensional coordinates of the user's click on the base map through the front end, send the two-dimensional coordinates to the back end to extract the corresponding one-dimensional spectral vector, and return them to the front end to draw the wavelength-reflectivity curve using the ECharts chart library. The minimum value of the vertical axis of the wavelength-reflectivity curve is fixed to 0, and the range of the vertical axis is dynamically determined by the maximum value of the current spectral vector. The operation record management module is configured to capture user interaction events on the details page in real time. These interaction events include mural scene switching, band carousel, color compositing, zooming, and panning operations, which are presented in chronological order. The module supports users in viewing historical operation trajectories and provides a one-click clear record function. The platform is built on the Django backend web framework, the Vue.js frontend framework, the Leaflet or Deck.gl frontend rendering engine, a relational database, and the ElementPlus frontend interface component library. The Django backend web framework is responsible for API services and the indexing and distribution of the multi-scale pyramid slice data. The relational database is used to store user data, basic metadata of cultural heritage sites, and index information of the multimodal analysis result layer data.

6. The online display platform for multispectral images of murals according to claim 4, characterized in that, The multi-scale pyramid data scheduling subsystem also includes a dynamic listener. The platform performs a dynamic acquisition process of billions of pixel data: the front-end rendering engine by default only loads the top slice of the multi-scale pyramid slice data and generates a lightweight RGB image as an interactive base map. When the user performs zoom or pan operations, the dynamic listener calculates the boundary coordinates and resolution level of the current visible area in real time, and requests the deep high-resolution data of the corresponding spatial slice from the backend as needed for loading and stitching. The window perception scheduling module only loads slice data within the current window range, and does not load slice data outside the window range; The resolution adaptive module automatically matches the corresponding pyramid resolution level based on the current scaling level; The progressive loading module first loads low-resolution slices when switching scaling levels, and then gradually loads high-resolution slices. The multi-scale pyramid slice data is stored in a distributed storage terminal and dynamically acquired by the front-end rendering engine based on the current viewport range and scaling level.

7. The online display platform for multispectral images of murals according to claim 3, characterized in that, The dual-window linkage control module is equipped with a single linkage control button. When the single linkage control button is activated, the front-end state machine captures the change state of either the left or right window and applies the transformation matrix synchronously to the other window in real time, maintaining absolute synchronous linkage between the left and right windows. The sliding overlay comparison module divides the upper layer and the lower layer into left and right sections within a single window. The draggable dividing line displays the lower layer on the left and the upper layer on the right. Users can control the division position in real time by dragging the draggable dividing line with the mouse. The transparency slider of the layer transparency blending module is continuously adjustable within the range of 0% to 100%, adjusting the transparency parameters of the upper layer in real time. The multimodal layer management module displays all available multimodal result layers at the current data collection location in a list format. Users can control the display and hiding of each layer by checking or toggling.

8. A method for online display of multispectral images of murals, based on the online display platform for multispectral images of murals as described in any one of claims 1 to 7, characterized in that, Includes the following steps: The process involves acquiring multi-scale pyramid slice data formed by storing original high-resolution mural images sliced ​​according to pyramid levels, and lightweight RGB interactive base map data synthesized by dimensionality reduction of the top slice of the multi-scale pyramid slice data. The lightweight RGB interactive base map data is then loaded to generate an initial display screen, resulting in a front-end rendering result containing the initial display screen. Monitor the scaling or translation operations in the front-end rendering results, calculate the boundary coordinates and resolution level of the current visible area in real time, and request the deep high-resolution data of the corresponding spatial slices from the multi-scale pyramid data scheduling subsystem on demand for loading and stitching based on the boundary coordinates and resolution level. When switching scaling levels, load low-resolution slices first and then gradually load high-resolution slices to obtain the slice rendering results after on-demand scheduling. Based on the slice rendering results, each result layer in the multimodal analysis result layer data is superimposed on the slice rendering results according to the spatial registration relationship with the original mural image. The display and hiding of each result layer are controlled by the check or switch operation of the multimodal layer management module to obtain the multimodal layer superimposed display result. Based on the multimodal layer overlay display result, the front-end state machine captures the transformation state of either the left or right window, and applies the transformation matrix synchronously to the other window in real time. The upper and lower layers are then compared or semi-transparently blended and overlaid using draggable dividing lines and transparency sliders to obtain the dual-window linkage and comparison display result. In response to a click operation on the target position in the slice rendering result, the two-dimensional coordinates of the target position are sent to the backend to extract the corresponding one-dimensional spectral vector. A wavelength-reflectivity curve is plotted based on the ECharts chart library, and the minimum value of the vertical axis of the wavelength-reflectivity curve is fixed to 0. The range of the vertical axis is dynamically determined by the maximum value of the current spectral vector, resulting in the spectral curve display result. The zoom, pan, click, check or switch, drag, and adjust operations are recorded as interactive operation events, arranged in chronological order, and can be cleared with one click to obtain the operation record management results.

9. The method for online display of multispectral images of murals according to claim 8, characterized in that, When monitoring scaling or translation operations in the front-end rendering results and calculating the boundary coordinates and resolution level of the current visible area in real time, the dynamic listener of the multi-scale pyramid data scheduling subsystem performs the real-time calculation, sends the boundary coordinates to the window perception scheduling module, and sends the resolution level to the resolution adaptation module. The window perception scheduling module requests deep high-resolution data of the corresponding spatial slice from the backend on demand based on the boundary coordinates and loads and stitches it. The resolution adaptive module matches the corresponding pyramid resolution level based on the resolution level. The progressive loading module loads low-resolution slices first and then gradually loads high-resolution slices when switching pyramid resolution levels. The multi-scale pyramid slice data is stored in a distributed storage terminal and dynamically retrieved on demand by the front-end rendering engine based on the current viewport range and scaling level.

10. The method for online display of multispectral images of murals according to claim 8, characterized in that, In response to a click operation on the target location in the slice rendering result, the front end captures the two-dimensional coordinates of the target location and sends the two-dimensional coordinates of the target location to the back end. The back end extracts a one-dimensional spectral vector based on the spatial location corresponding to the two-dimensional coordinates of the target location in the multi-scale pyramid slice data. The one-dimensional spectral vector is the reflectance value sequence of the mural pixel corresponding to the spatial location in each spectral band. After the one-dimensional spectral vector is returned to the front end, the ECharts chart library draws a wavelength-reflectance curve. The minimum value of the vertical axis of the wavelength-reflectance curve is fixed to 0, and the range of the vertical axis is dynamically determined by the maximum value of the current spectral vector. When the zoom, pan, click, check or switch, drag, and adjust operations are recorded as interactive operation events, the operation record management module captures the occurrence time and operation type of each interactive operation event in real time. The interactive operation events also include mural scene switching, band carousel, and color synthesis. The interactive operation events are arranged and presented in chronological order according to their occurrence time. Users can view historical operation trajectories and a one-click clear record function is provided, which clears all interactive operation events.

Images