Layered material identification detection method for oil painting restoration process

Abstract

The disclosure provides a layered material identification detection method for oil painting restoration process, which comprises multi-scale imaging and local micro-area section detection of the oil painting surface to generate a pollution deposition space distribution map; based on the pollution deposition space distribution map, an interface continuity expression of the smoke deposition layer and the material layer is constructed to distinguish the penetrating diffusion structure, the interface adhesion structure and the transition mixed structure; based on the interface continuity expression, the penetration path of the pollution deposition to the varnish and pigment interface is identified to generate a pollution penetration path model; based on the pollution penetration path model, layered discrimination of the pollution pseudo-layer and the real restoration layer is performed to generate an oil painting real material layer sequence structure diagram and output a layered identification result, which can accurately distinguish the smoke deposition pseudo-layer and the real material layer, restore the real layer sequence structure of the oil painting, and improve the accuracy of layered judgment in the oil painting restoration process.

Description

Technical Field

[0001] This disclosure relates to the fields of image processing and material detection technology, and in particular to a method for layered material identification and detection in the process of oil painting restoration. Background Technology

[0002] In the field of oil painting restoration, accurately identifying the layered structure of oil painting materials is crucial for developing a restoration plan. Oil paintings are typically composed of layers of paint, varnish, and repair materials added during previous restorations. Restorers need to clearly distinguish the boundaries and properties of each layer in order to treat the affected areas effectively.

[0003] However, in actual restoration work, oil paintings exposed to specific environments for extended periods are often affected by smoke contamination. When an oil painting is placed in a furnace or gas environment for a long time, soot gradually seeps into the interface between the varnish layer and the paint layer, forming a contamination film that resembles an independent layer between the two layers. This contamination film is visually and structurally very similar to the actual historical restoration layers, making it difficult for existing detection systems to distinguish and easily misidentifying it as a historical restoration layer, thus affecting the accuracy of the restoration plan. This apparent pseudo-layer problem caused by smoke contamination interferes with the reconstruction of the true layered structure of the oil painting, making it impossible for restorers to accurately determine which areas contain historical traces that should be preserved and which areas contain contaminants that need to be cleaned.

[0004] Therefore, there is an urgent need for a layer identification and detection method that can accurately distinguish between smoke deposits and real material layers in order to restore the true layered structure of oil paintings and improve the accuracy of layer judgment during the restoration of oil paintings. Summary of the Invention

[0005] In view of this, in order to solve the problems brought about by the existing technology, this application provides a layered material identification and detection method for the oil painting restoration process.

[0006] In a first aspect, this disclosure provides a method for identifying and detecting layered materials in the oil painting restoration process, the method comprising: S1: Perform multi-scale imaging and local micro-area profile detection on the surface of the oil painting to generate a spatial distribution map of contamination deposition; S2: Based on the aforementioned spatial distribution map of contamination deposition, construct an expression of the interface continuity between the smoke deposition layer and the material layer, and distinguish between through-diffusion structures, interface attachment structures, and transitional mixed structures; S3: Based on the aforementioned interface continuity expression, identify the penetration path of contaminant deposition to the varnish-pigment interface and generate a contaminant penetration path model;

[0007] S4: Based on the pollution penetration path model, perform layer discrimination between the pollution pseudo layer and the real repair layer, generate a layer sequence structure diagram of the real oil painting materials, and output the layer recognition results.

[0008] Optionally, S1 includes: The surface of the oil painting is divided into continuous detection units. Multi-scale imaging and multi-angle illumination are performed on each detection unit to calculate the surface contamination candidate index and screen the surface candidate contamination anomaly area. Local micro-area profile detection is performed on the candidate surface contamination anomaly areas to statistically analyze the particle distribution intensity at each depth layer; The particle distribution intensity at different depths is sorted in the thickness direction to form a longitudinal particle distribution curve. The distribution state of particles in the surface, interior and interface is identified based on the peak position of the longitudinal particle distribution curve, the position of the particle interlayer transfer boundary is determined, the centroid of the longitudinal particle distribution is calculated, and it is converted into a particle penetration level identifier. The particle penetration level identifiers and interlayer transfer boundary positions of each detection unit are mapped back to the overall spatial position of the oil painting, generating a spatial distribution map of pollution deposition and a layered identification map of pollution particle depth.

[0009] Optionally, S2 includes: The spatial distribution map of the contamination deposition is subjected to secondary discretization processing, and the discrete detection units are aggregated into a set of adjacent segments of particles in the same layer and a set of candidate connection segments across layers. Calculate the intralayer aggregation continuity of the adjacent segments of the same-layer particles and the cross-layer continuity of the cross-layer candidate connection segments; Based on the same-layer aggregation continuity and the cross-layer continuity, the longitudinal diffusion attribution ratio is calculated, and the same-layer particle adjacent segment and the cross-layer candidate connection segment are divided into through-diffusion structure, interface attachment structure and transitional mixing structure. The three types of structure sets are remapped onto the overall space of the oil painting to generate a continuous expression of the contaminated interface that includes structural difference identifiers.

[0010] Optionally, the step of dividing the adjacent segments of the same-layer particles and the candidate cross-layer connection segments into a through-diffusion structure, an interface attachment structure, and a transitional hybrid structure includes: If the longitudinal diffusion attribution ratio is higher than the high-level threshold, it is classified into the through-diffusion structure set; If the longitudinal diffusion attribution ratio is lower than the low-level threshold and the same-layer aggregation continuity is higher than the preset continuity threshold, then it is classified into the interface attachment structure set. If the longitudinal diffusion attribution ratio is between the low-level threshold and the high-level threshold, it is classified into the transitional mixed structure set.

[0011] Optionally, S3 includes: The continuous regions in the interface continuity representation result are projected onto different depth domains and divided into partitioned penetration channel segments. The depth propulsion intensity of each partitioned penetration channel segment is calculated, and effective candidate channel segments are selected based on the depth propulsion intensity. For each valid candidate channel segment, a thickness-direction stratigraphic sequence is established, and the direction determination coefficient is calculated. Based on the direction determination coefficient, the direction determination result of a single channel is obtained. Multiple valid candidate channel segments in the same region are summarized, and a set of regional dominant permeation directions is formed based on the summary results. The boundary is expanded with the region of dominant penetration direction as the center, the regional diffusion consistency index is calculated, and the regional diffusion boundary results are generated. At the same time, path segment merging rules are established. The regional diffusion boundary results and the path segment merging rules are applied to the entire oil painting to generate a pollution infiltration path model.

[0012] Optionally, the contamination infiltration path model is constructed based on the path-level results and the contamination layer adhesion relationship results; The path level results include shallow path, medium path and deep interface path. The shallow path indicates that the contaminant mainly stays on the surface and shallow layer of the varnish. The medium path indicates that the contaminant has entered the interior of the varnish but has not yet accumulated stably at the interface. The deep interface path indicates that the contaminant has formed a significant residue at the varnish pigment interface. The contamination layer adhesion relationship results include no direct adhesion relationship, indirect interlayer adhesion relationship, and direct interface adhesion relationship. The absence of a direct adhesion relationship indicates that the path terminates at the surface of the varnish layer or the shallow layer of the varnish. The indirect interlayer adhesion relationship indicates that the path terminates inside the varnish layer. The direct interface adhesion relationship indicates that the path terminates at the varnish pigment interface and the path adhesion formation index is higher than a preset threshold.

[0013] Optionally, S4 includes: Based on the path termination layer, attachment range, attachment stability and regional formation mode of the pollution infiltration path model, the pollution pseudo-layer tendency index is calculated, and candidate pollution pseudo-layer areas, candidate real remediation layer areas and candidate mixed dispute areas are initially divided according to the pollution pseudo-layer tendency index. For the candidate contaminated pseudo-layer region, candidate real remediation layer region, and candidate mixed dispute region, the boundary closure characteristics and thickness abrupt change characteristics of each candidate region are analyzed, the regional thickness abrupt change index is calculated, and the boundary termination type is identified. Based on the layer boundary closure characteristics, the regional thickness abruptness index, and the boundary termination type, the regional structure discrimination value is calculated and used in conjunction with the pseudo-layer tendency index to determine the attribution of the contaminated pseudo-layer area, the real remediation layer area, and the area to be reviewed. The confirmed contaminated pseudo-layer areas and real restoration layers are mapped back to the spatial coordinates of the oil painting to generate a real material sequence structure diagram of the oil painting and output the area confirmation level. The confirmed areas to be reviewed are output as dispute markers. The area confirmation level is calculated by the area structure discrimination value and the contaminated pseudo-layer tendency index.

[0014] Optionally, the collaborative judgment includes: For the candidate mixed dispute area, when the pollution pseudo-layer tendency index is higher than its high-level threshold and the regional structure discrimination value is lower than its low-level threshold, if the path attachment formation index and the interface retention enhancement value are both significantly higher, it is classified as a pollution pseudo-layer confirmation area; otherwise, it is downgraded to a pending review area. When the pollution pseudo-layer tendency index is lower than its low-level threshold and the regional structure discrimination value is higher than its high-level threshold, if the intra-layer aggregation continuity is significantly higher than the inter-layer continuity, it is classified as a real remediation layer confirmation area; otherwise, it is downgraded to a pending review area.

[0015] In a second aspect, this disclosure provides an electronic device including a memory and at least one processor, the memory storing a computer program, and the processor executing the computer program to implement the method of the first aspect described above.

[0016] Thirdly, this disclosure provides a computer storage medium storing a computer program that, when executed, implements the method described in the first aspect.

[0017] The beneficial effects of this disclosure are that, compared with the prior art, this disclosure has the following advantages: 1) Addressing the technical problem that existing detection systems struggle to distinguish between smoke-contaminated films, which are visually and structurally very similar to genuine historical restoration layers, this disclosure constructs an interface continuity representation between the smoke-contaminated deposition layer and the material layer by segmenting the spatial distribution map of the contamination and calculating the continuity of aggregation within the same layer and the continuity of continuity across layers. This allows for the creation of comparable quantitative indicators for the structural characteristics of smoke contamination penetrating layer by layer from the surface to the interior along cracks or micropores, and the stable coverage and intact boundaries of genuine restoration layers. Furthermore, by identifying the penetration path of contamination deposition at the varnish and pigment interface and establishing a contamination penetration path model, this model clarifies the starting position, penetration layer, diffusion range, and termination interface of contamination advancing from the surface to the interior. During model construction, a path attachment index is used to assign a path level to each path, thereby fundamentally distinguishing the formation mechanism of smoke contamination advancing from the surface to the interior from the material stacking mechanism of genuine restoration layers at the path level. This accurately differentiates between the pseudo-smoke-contaminated layer and the genuine historical restoration layer, avoiding the technical defect of misjudging smoke deposits as historical restoration materials.

[0018] 2) Addressing the technical problem of apparent pseudo-layers formed by smoke contamination interfering with the reconstruction of the true stratification structure of oil paintings, making it difficult for restorers to accurately determine which areas are historical traces that should be preserved and which areas are contaminants that need to be cleaned, this disclosure obtains the distribution characteristics of contamination deposition on the oil painting surface and establishes stratification markers and particle penetration level markers at different depths of contamination particles. This provides basic data on spatial location and thickness direction for stratification reconstruction. Then, based on the contamination penetration path model and regional structural characteristics, it performs stratification discrimination between the contamination pseudo-layer and the true restoration layer, calculates the pseudo-layer tendency index and regional structure discrimination value for collaborative judgment, and finally generates a true material stratification structure diagram of the oil painting including the original pigment layer, varnish layer, contamination pseudo-layer, true restoration layer, and disputed areas. At the same time, it outputs the area confirmation level and dispute marker, enabling restorers to clearly identify which areas belong to the cleanable smoke pseudo-layer, which areas belong to the historical restoration traces that should be preserved, and which areas need manual verification. This solves the problem of smoke contamination pseudo-layers interfering with the reconstruction of the true stratification structure of oil paintings and significantly improves the accuracy of stratification judgment in the oil painting restoration process. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0020] Figure 1 A flowchart of a layered material identification and detection method for oil painting restoration provided in this embodiment is shown. Figure 2 A flowchart illustrating the contamination interface continuity structure discrimination provided in an embodiment of this disclosure is shown. Figure 3 A flowchart illustrating the construction process of the contamination infiltration path model provided in an embodiment of this disclosure is shown.

[0021] The accompanying drawings have illustrated specific embodiments of this disclosure, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concepts of this disclosure to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0022] The present disclosure will be further described below with reference to the accompanying drawings. The following embodiments are only used to illustrate the technical solutions of the present disclosure more clearly, and should not be used to limit the scope of protection of the present disclosure.

[0023] The components of the embodiments of the invention described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0024] In the following, the terms “comprising,” “having,” and their cognates, which may be used in various embodiments of the invention, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as excluding, firstly, the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more features, numbers, steps, operations, elements, components, or combinations thereof.

[0025] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the invention pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of the invention.

[0026] Figure 1 This is a flowchart of a layered material identification and detection method for oil painting restoration provided in this embodiment of the disclosure, as follows: Figure 1 As shown, the process may include the following steps: S1: Perform multi-scale imaging and local micro-area profile detection on the surface of the oil painting to generate a spatial distribution map of contamination deposits.

[0027] Multi-scale surface imaging and local micro-area profile detection are performed on the overall surface of the oil painting to identify the distribution of smoke particles on the surface of the varnish layer, inside the varnish layer, and at the interface of the pigment layer, generating a spatial distribution map of contamination deposition and establishing layered markers for contamination particles at different depths. This is achieved through the following sub-steps.

[0028] S1.1: Perform multi-scale imaging and multi-angle lighting on the oil painting surface, calculate the candidate surface contamination index, and screen the candidate surface contamination anomaly areas.

[0029] The surface of the target oil painting was divided into continuous detection units of fixed dimensions, each unit corresponding to a unique spatial location number, so that the detection results could be mapped back to the overall space of the painting. For each detection unit, low-magnification full-frame imaging, medium-magnification texture imaging, and high-magnification grain imaging were performed sequentially to obtain the surface morphology of the same area at different observation scales. Low-magnification full-frame imaging was used to determine the overall outline and macroscopic features of the detection unit; medium-magnification texture imaging was used to capture the aging texture of the varnish and the details of the paint brushstrokes; and high-magnification grain imaging was used to identify the aggregation state of tiny particles.

[0030] To avoid misinterpreting varnish aging textures as smoke deposits based on a single lighting condition, it is necessary to acquire reflective images of the same area under multiple incident angles. Specifically, images are taken under both oblique and perpendicular incident conditions to distinguish bright spots, dark lines, particle shadows, and crack boundaries. Under oblique incident conditions, tiny particles create noticeable shadows, while surface undulations are less pronounced under perpendicular incident conditions. By comparing images from both lighting conditions, misinterpretations caused by the angle of illumination can be effectively eliminated.

[0031] After obtaining multi-scale surface imaging results, surface particle aggregation response value, surface dark deposition response value, and fine crack occlusion response value are calculated for each detection unit. The surface particle aggregation response value is obtained by normalizing the number of high-contrast edge particles per unit area in the high-magnification image, reflecting the particle density in that area. The surface dark deposition response value is obtained by normalizing the area ratio of low-reflection areas in multi-angle images, reflecting the degree of color darkening caused by dust adhesion in that area. The fine crack occlusion response value is obtained by combining the brightness difference across the crack and the crack density, reflecting the occlusion effect of dust particles filling the crack on its features. The above three response values ​​are weighted and combined according to preset weights to obtain the surface contamination candidate index, which is calculated as follows: ; in, This represents a candidate index for surface contamination and is dimensionless. , , This represents the weighting coefficients, and the sum of the three is 1. The weighting coefficients are set based on the following criteria: if the oil painting surface is obviously thickly coated, the weight of the particle aggregation term is increased; if the varnish is obviously yellowed, the weight of the gray deposition term is increased; if the cracks are dense, the weight of the crack concealment term is increased. This represents the surface particle aggregation response value, which is dimensionless and is obtained by normalizing the number of high-contrast edge particles per unit area in a high-magnification image. This represents the surface gray deposition response value, which is dimensionless and is obtained by normalizing the area ratio of low-reflection regions in multi-angle images. This represents the shielding response value of a fine crack, which is dimensionless and is obtained by combining the brightness difference on both sides of the crack and the crack density. This represents the response enhancement coefficient, with a value ranging from 2 to 8; This represents the average reflected intensity under oblique incidence conditions. It is dimensionless and is obtained by normalizing the image grayscale. Its value ranges from 0 to 1. This represents the average reflected intensity under perpendicular incidence, is dimensionless, and is obtained after image grayscale normalization, with a value range of 0 to 1. The denominator adopts an exponential function form with the natural constant as the base, and its value range is (1,2). This is intended to address situations where the difference in reflected intensity between oblique and perpendicular incidence is significant. When the denominator is large, it approaches 1, thus enhancing the exponent; conversely, when the difference in reflection intensity is not significant, the denominator increases, thus suppressing the exponent and eliminating misjudgments caused by simple surface texture undulations.

[0032] This candidate index for surface contamination is constrained by three types of appearance features: surface particle aggregation, surface dark deposition, and crack occlusion. It also uses the difference in reflection at different angles to suppress false dark areas, thereby stably screening out suspected smoke deposit areas.

[0033] When the surface contamination candidate index exceeds the set threshold, the detection unit is marked as an abnormal surface contamination area. This sub-step first uses a large-scale, low-intervention method to screen out areas that are truly worthwhile for micro-area profile detection, avoiding wasting the operation window on a large number of irrelevant areas in subsequent profile detection, and also avoiding mistaking natural aging of varnish, darkening of the pigment body, and surface wear as smoke contamination.

[0034] S1.2: Perform local micro-area profile detection for candidate contamination anomaly areas on the surface and statistically analyze the particle distribution intensity at each depth layer.

[0035] For each candidate surface contamination anomaly area output in sub-step S1.1, its center location and boundary transition location are selected as local micro-area profile detection points. A non-destructive layer-by-layer scan is performed along the thickness direction for each detection point to obtain continuous depth responses from the surface into the varnish layer, from the varnish layer to the pigment layer, and close to the pigment layer interface. Non-destructive layer-by-layer scanning can be achieved using techniques such as optical coherence tomography or confocal microscopy, which can acquire structural information of each depth layer without damaging the painting.

[0036] To avoid misidentifying microbubbles, localized resin agglomerates, or pigment particle protrusions within the varnish as smoked particles, structural constraint identification of the continuous depth response is required. Specifically, the particle density, particle size distribution, neighborhood adhesion continuity, and intralayer attenuation gradient are statistically analyzed for each depth layer. If particles are mainly concentrated on the surface and decay rapidly downwards, they are classified as surface deposition; if particles extend discretely within the varnish layer, they are classified as internal penetration; if particles form banded aggregates at the varnish-pigment interface, they are classified as interfacial adhesion. To quantify the particle distribution intensity at each depth layer, the number of particles, average particle size, and neighborhood adhesion continuity between particles must be considered simultaneously. The particle distribution intensity is calculated as follows: ; in, It represents the intensity of particle distribution at a certain depth layer and is dimensionless. This indicates the number of particles identified in this depth layer, expressed in individual particles. This indicates the effective detection area corresponding to this depth layer, in square millimeters; to This represents the equivalent diameter of each particle within that depth layer, in millimeters. If n particles are identified in a certain layer, then each... The equivalent size of a specific particle; n represents the number of particles, which is a positive integer; This represents the continuity of particle neighborhood adhesion within this depth layer, is dimensionless, and is obtained by normalizing the degree of contact and spacing between adjacent particle boundaries. The calculated value... This will serve as the basis for the analysis of the longitudinal distribution characteristics of particles in step S1.3.

[0037] The core of this calculation method lies in the fact that when determining whether smoke pollution exists within a deep layer, one cannot only look at the number of particles, but also at the average size of the particles and whether they have an attachment and continuity relationship, because smoke infiltration usually manifests as continuous attachment of small particles, rather than isolated scattered large particles. This sub-step advances the candidate anomaly area on the surface from a planar suspicious area to a layered distribution result in the thickness direction, providing a direct basis for the subsequent generation of a spatial distribution map of pollution deposition.

[0038] S1.3: Sort the particle distribution intensity at different depths according to the thickness direction to form a longitudinal particle distribution curve. Identify the particle distribution status on the surface, inside and at the interface based on the peak position of the curve. Calculate the centroid of the longitudinal particle distribution and convert it into a particle penetration level identifier.

[0039] The particle distribution intensity at different depths obtained in sub-step S1.2 is reordered according to the thickness direction to form the longitudinal particle distribution curve corresponding to each detection unit. Then, the peak positions and decay rates of the curves in the surface region, middle region, and interface region are analyzed. If the peak is concentrated on the outermost surface and decreases monotonically downwards, it is determined that surface deposition is dominant; if the peak is located in the middle of the varnish layer and shows a bimodal or gradually decreasing state, it is determined that internal penetration is dominant; if the peak is close to the pigment interface and forms a local rise near the interface, it is determined that interface accumulation is dominant.

[0040] To establish layered identifiers that can be directly used in subsequent steps, the longitudinal distribution characteristics of each detection unit need to be converted into particle penetration level identifiers, and the locations of interlayer particle transfer boundaries need to be given. These boundary locations are not the actual thickness boundaries of the canvas, but rather the locations where particle distribution transitions from one dominant state to another. Specifically, the centroid of the longitudinal particle distribution is used to characterize the bias of contaminants in the thickness direction, and its calculation method is as follows: ; in, Indicates the position of the centroid of the longitudinal distribution of particles; dimensionless. to The particle distribution intensity, calculated in step S1.2, represents the particle distribution intensity at each depth from the surface to the pigment interface, all in millimeters per square millimeter. If the layer is divided into m depth layers, then each... It corresponds to a specific depth layer; to The value indicates the center depth of the corresponding depth layer, in millimeters; m represents the number of depth layers, a positive integer; H represents the total detection thickness of the current detection unit from the surface to the pigment interface, in millimeters. The closer the center of gravity of the longitudinal distribution of particles is to the surface, the more likely it is surface fumigation adhesion; the further inward the center of gravity moves, the more obvious the internal migration or interface accumulation of contaminants.

[0041] Based on this, a criterion for classifying particle penetration levels is established: when the centroid of the longitudinal distribution of particles is less than the lower threshold, it is classified as shallow deposition; when it is within the middle threshold range, it is classified as intermediate penetration; and when it exceeds the higher threshold, it is classified as deep interface accumulation. The specific values ​​of the lower and higher thresholds can be determined based on the average thickness of the varnish layer and previous restoration records of this type of oil painting.

[0042] S1.4: Map the particle penetration level identifiers and interlayer transfer boundary positions of each detection unit back to the overall spatial position of the oil painting to generate a spatial distribution map of pollution deposition and a layered identification map of pollution particle depth.

[0043] The particle penetration level identifiers and interlayer transfer boundary positions of each detection unit output from sub-step S1.3 are mapped back to the overall spatial position of the oil painting to reconstruct the spatial distribution map of contamination deposition in the entire oil painting. This distribution map contains at least three aspects: first, the distribution of abnormal contamination areas on the plane; second, the deposition location type in the thickness direction, i.e., whether each detection unit is dominated by surface deposition, internal penetration, or interface accumulation; and third, whether there are continuous smoke migration zones between different areas, i.e., whether the particle distribution characteristics of adjacent detection units show spatial continuity.

[0044] Based on this, the particle penetration level identifier of each detection unit is uniformly coded to form a contaminant particle depth stratification map. For surface deposition-dominant areas, this map is used to determine whether surface cleaning is sufficient to remove the influence of the contaminant pseudo-layer; for internal penetration-dominant areas, it is used to determine whether smoke particles have entered the interior of the clear coat layer; for interface accumulation-dominant areas, it is used to determine whether a pseudo-layer band may form at the clear coat pigment interface. Finally, the contaminant deposition spatial distribution map and the contaminant particle depth stratification map are combined as the final output of step S1 and directly provided to step S2 for constructing an interface continuity representation between the smoke deposition layer and the material layer.

[0045] In the technical solution of this disclosure, the distribution of contamination deposition characteristics on the surface of an oil painting is obtained by combining multi-scale surface imaging with local micro-area profile detection. This method can screen out candidate contamination anomaly areas on the surface under large-scale and low-intervention conditions, avoiding misjudging natural aging of varnish, darkening of the pigment body, and surface wear as smoke contamination. At the same time, it establishes layered identification and particle penetration level identification of contamination particles at different depths, providing dual basic data in terms of spatial location and thickness direction for subsequent continuous analysis and penetration path identification. This effectively solves the problem that traditional methods rely solely on surface visual features and cannot distinguish between smoke deposition and material layer boundaries.

[0046] S2: Based on the spatial distribution map of the pollution deposition, construct an expression of the interface continuity between the smoke deposition layer and the material layer, and distinguish between the through-diffusion structure, the interface attachment structure and the transitional mixing structure.

[0047] Based on the spatial distribution map of contamination deposits obtained in step S1, the continuity analysis of the aggregation state of sediment particles in different depth regions is performed to identify whether the sediment layer exhibits a through-diffusion structure or an interface attachment structure, and to form a continuous expression result of the contamination interface, so as to distinguish the structural feature differences between the sediment film and the actual repair material layer. Figure 2 A flowchart illustrating the contamination interface continuity structure discrimination process provided in an embodiment of this disclosure is shown, as follows: Figure 2 As shown, this is achieved through the following sub-steps.

[0048] S2.1: Perform secondary discretization on the spatial distribution map of contaminant deposition, and aggregate the discrete detection units into a set of adjacent segments of particles in the same layer and a set of candidate connection segments across layers.

[0049] The spatial distribution map of contaminant deposition output in step S1 is discretized a second time according to the location on the surface and the thickness of the oil painting, so that each detection unit simultaneously possesses four attributes: planar location, depth level, particle distribution intensity, and penetration level. The purpose of this second discretization is to integrate the originally scattered detection unit information into a data structure that facilitates continuous analysis, ensuring that subsequent analysis can simultaneously consider spatial adjacency relationships and inter-layer depth relationships.

[0050] Within the same depth layer, detection units with similar particle distribution intensities and consistent particle adhesion states are connected according to the boundary contact relationship between adjacent detection units, forming a set of adjacent particle segments in the same layer. In specific implementation, for each depth layer, all detection units marked as candidate pollution anomaly areas are traversed. If two units are adjacent in planar position and the difference in their particle distribution intensities is less than a preset tolerance, and their penetration levels are the same, they are grouped into the same adjacent segment. In this way, discrete detection units are aggregated into spatially continuous segments.

[0051] Between adjacent depth layers, it is checked whether the detection units of the upper and lower layers overlap or partially overlap in planar position. Combined with the position of the particle longitudinal distribution centroid and the interlayer transfer boundary in step S1, a set of cross-layer candidate connection segments that may constitute a continuation relationship between upper and lower layers is selected. In specific implementation, for each adjacent segment in the upper layer and each adjacent segment in the lower layer, if the overlapping area of ​​the two in the planar projection exceeds a preset threshold, and the particle longitudinal distribution centroid of the upper layer segment and the particle longitudinal distribution centroid of the lower layer segment show a progressive relationship in the thickness direction, then the upper and lower layer segments are marked as cross-layer candidate connection segments.

[0052] This sub-step does not directly determine whether the contamination layer is a smoke layer, but first clearly identifies which particle segments might be connected to each other. This is because subsequent continuity analysis does not directly judge individual particles, but rather the continuous structure composed of multiple detection units. Without this segmentation process, subsequent steps would only be able to describe points, failing to form an expression of interface continuity.

[0053] S2.2: Calculate the intra-layer aggregation continuity of adjacent segments of particles in the same layer and the cross-layer continuity of candidate connection segments across layers.

[0054] For each adjacent segment of particles in the same layer, its length continuity, width stability, and particle distribution intensity uniformity in the current depth layer are calculated to obtain the contiguous layer aggregation continuity result. The contiguous layer aggregation continuity is calculated as follows: ; in, It represents the continuity of clustering at the same level and is dimensionless. This indicates the actual connectivity length of adjacent segments of particles in the same layer, in millimeters; This indicates the theoretical maximum extendable length of the detection area corresponding to the layer where this section is located, in millimeters; to This represents n width samples collected sequentially along the connected length direction, in millimeters. This represents the width of the segment at the corresponding sampling location; n represents the number of width sampling points, which is a positive integer. This represents the width fluctuation normalization coefficient, in millimeters, with a value ranging from 0.05 to 0.50 millimeters, set according to the detection resolution and typical crack width; This represents the non-uniformity of particle distribution intensity within a layer, and is dimensionless. It is derived from the particle distribution intensity at adjacent sampling locations along the connected length within the same depth layer. The dispersion was normalized to obtain the value, which is used to characterize the uniformity of particle distribution within this section, and is consistent with the description of the adhesion relationship between particles in the same layer in S1.2. different.

[0055] The core of this calculation method lies in the fact that a truly continuous overburden layer must not only be long enough, but also relatively stable in width and relatively uniform in particle distribution intensity. Smoke deposits often have broken edges and uneven thickness in some areas, so the continuity of the same layer is generally not as high as that of artificially repaired layers.

[0056] For each cross-layer candidate connection segment, its positional offset, particle distribution attenuation, and contact boundary preservation between adjacent depth layers are calculated to obtain the cross-layer continuity result. The cross-layer continuity is calculated as follows: ; in, Indicates the continuity across layers; dimensionless. This represents the number of inter-layer segments connecting different layers, and is a positive integer. to It represents the contact boundary retention rate between each layer segment, is dimensionless, and is obtained by the ratio of the length of the overlapping boundary between the upper and lower layers to the length of the reference boundary. to It represents the planar position offset rate of each inter-layer segment, is dimensionless, and is obtained by dividing the geometric center position offset between the upper and lower layers by the reference length; This represents the particle distribution attenuation deviation, which is dimensionless and is obtained by normalizing the difference between the actual attenuation curve of particle distribution intensity and the ideal diffusion attenuation curve between each layer.

[0057] The core of this calculation method lies in the fact that a through-diffusion structure does not require the upper and lower layers to completely overlap, but it does require that it maintain a certain continuity of contact during interlayer transfer, while its offset and attenuation should conform to the basic laws of diffusion propagation. If the interlayer continuity is high, it indicates that the particles are more like infiltrating layer by layer along the channel, rather than being artificially and uniformly laid into a single layer.

[0058] The same-layer aggregation continuity and cross-layer continuity obtained from the above calculations are used to characterize the morphological features of each candidate structure, which are the adjacent segments of the same-layer particles and the candidate connection segments across layers.

[0059] S2.3: Based on the same-layer aggregation continuity and cross-layer continuity, the longitudinal diffusion attribution ratio is calculated, and the candidate structures are divided into through-diffusion structures, interface attachment structures and transitional hybrid structures.

[0060] Based on sub-step S2.2, a structure classification judgment is further performed on each candidate structure. If the cross-layer continuity of a candidate structure is significantly higher than the intra-layer aggregation continuity, and its particle longitudinal distribution center gradually shifts between multiple depth layers, then it is classified into the through-diffusion structure set. If the intra-layer aggregation continuity of a candidate structure is high, its width is stable and mainly concentrated at the interface between the varnish layer and the pigment layer, and its boundary morphology and particle distribution intensity uniformity conform to the characteristics of contamination adhesion, such as edge breakage and local uneven thickness, then it is classified into the interface adhesion structure set. If both types of continuity indicators are in the middle range, and the structure has both interface retention characteristics and local longitudinal diffusion characteristics, then it is classified into the transitional mixed structure set.

[0061] To ensure the feasibility of the discrimination results, it is necessary to set graded thresholds for both intralayer aggregation continuity and interlayer continuity. Generally, the lower threshold can be set to 0.25 to 0.40, and the higher threshold to 0.60 to 0.80. If the varnish layer of the oil painting is thick and has been exposed to heat and smoke for a long time in the early stages, the higher threshold for interlayer continuity can be appropriately lowered. If the oil painting has undergone large-area glaze repairs, the higher threshold for intralayer aggregation continuity should be increased to avoid misjudging the artificial repair layer as ordinary interface contamination. In practical applications, a mapping relationship between different oil painting types and threshold levels can be established in advance. A dynamic threshold adjustment library can be formed through a small number of samples or expert experience, enabling the discrimination process to automatically select appropriate threshold combinations based on the material characteristics of the oil painting itself, thereby further improving the robustness of interlayer continuity and intralayer aggregation continuity discrimination under different working conditions.

[0062] In the discrimination process, the longitudinal diffusion attribution ratio is used to quantify the structure attribution, and its calculation method is as follows: ; in, Indicates the longitudinal diffusion attribution ratio, dimensionless; Indicates the continuity across layers; dimensionless. denoted by , which represents the contiguousness of clustering within the same layer and is dimensionless; u represents the cross-layer continuity enhancement index, ranging from 1.20 to 2.50, used to enhance the discriminative sensitivity of longitudinal diffusion characteristics; v represents the contiguous clustering suppression index, ranging from 1.00 to 2.20, used to control the influence of covering structures on the attribution results. The term represents the mixed disturbance correction term, which is dimensionless and ranges from 0.05 to 0.30. It is calculated by combining the local interface retention ratio and the local crack diffusion ratio in the transitional mixed structure.

[0063] The core of this calculation method lies in converting the structure from resembling longitudinal diffusion to interfacial adhesion into a comparable attribution ratio. A higher attribution ratio indicates that the structure is closer to a contamination structure formed by the diffusion of smoke particles along the thickness direction; a lower attribution ratio indicates that the structure is closer to a stable interfacial adhesion or a real material layer.

[0064] If the longitudinal diffusion attribution ratio is higher than the high-level threshold, it is classified into the through-diffusion structure set; if the longitudinal diffusion attribution ratio is lower than the low-level threshold, and the same-layer aggregation continuity is higher than a preset continuity threshold, it is classified into the interface attachment structure set. This continuity threshold can be preset according to the type of oil painting and restoration history, and its value range is generally from 0.50 to 0.70. If the longitudinal diffusion attribution ratio is between the low-level threshold and the high-level threshold, it is classified into the transitional mixing structure set, and further refined in step S3 based on the penetration path.

[0065] S2.4: Remap the three types of structure sets to the overall space of the oil painting to generate a contamination interface continuity representation containing structural difference identifiers.

[0066] The three types of structure sets obtained in sub-step S2.3 are remapped into the overall space of the oil painting to form an expression of the interface continuity for the entire oil painting. This expression includes at least three aspects: first, which areas exhibit a through-diffusion structure; second, which areas exhibit an interface attachment structure; and third, which areas belong to a transitional mixed structure.

[0067] Based on this, structural differences are identified for different regions. For through-diffusion structures, the main diffusion initiation layer, main diffusion direction, and termination location are identified; for interface-attached structures, the main attachment interfaces, interface coverage length, and degree of continuous coverage are identified; for transitional mixing structures, the mixing region range and dominant tendency are identified. The final output of the contamination interface continuity expression results will be directly provided to step S3, which will further identify the penetration path of contamination deposits at the varnish-pigment interface.

[0068] The continuity representation here is not a simple classification diagram, but an engineered representation of how smoke particles form structural continuity between layers. Only then can we truly distinguish between pseudo-contaminated membranes and genuine historical repair layers. Genuine repair layers often exhibit relatively stable co-layer coverage and relatively clear interface boundaries, while smoke contamination is more likely to manifest as irregular, continuous structures along interfaces, cracks, and local channels.

[0069] In the technical solution of this disclosure, by segmenting the spatial distribution map of pollution deposition, an expression of the interface continuity between the smoke deposition layer and the material layer is constructed. This enables the discrete detection units to be aggregated into adjacent segments of particles in the same layer and candidate connection segments across layers with spatial continuity. The continuity of aggregation in the same layer and the continuity of continuity across layers are calculated, thereby classifying the through-diffusion structure, the interface attachment structure and the transitional mixing structure. This achieves an engineering-based quantitative expression of how smoke particles form structural continuity between layers, providing a structural feature-level discrimination basis for distinguishing between smoke pollution films and real historical repair layers.

[0070] S3: Based on the interface continuity expression, identify the penetration path of contaminant deposition to the varnish-pigment interface and generate a contaminant penetration path model.

[0071] Based on the continuity expression results of the contamination interface obtained in step S2, the structure of the penetration direction and diffusion range of the deposited layer in the interface between the varnish layer and the pigment layer is inferred, a contamination penetration path model is established, and the formation mode of the contamination layer in different regions and its adhesion relationship with the original material layer are determined. Figure 3 A flowchart illustrating the construction process of the contamination infiltration path model provided in this embodiment is shown, as follows: Figure 3 As shown, this is achieved through the following sub-steps.

[0072] S3.1: Project the continuous region onto different depth domains and divide it into partitioned penetration channel segments, calculate the depth propulsion intensity, and screen effective candidate channel segments.

[0073] Read the through-diffusion structure region, interface adhesion structure region, and transitional mixing structure region already marked in step S2, and reproject them onto three depth domains: the surface of the varnish layer, the interior of the varnish layer, and the varnish-pigment interface. Specifically, determine which depth domain each continuous region mainly exists in based on the position of its particle longitudinal distribution centroid, and record its spatial range.

[0074] Based on each continuous region, segmentation is performed along its length and thickness directions to obtain a set of partitioned penetration channel segments with clear start points, transition sections, and end points. During the segmentation process, three types of locations need to be identified: the first type is the high-deposition distribution area near the outer surface of the oil painting, which is usually the candidate penetration initiation area for smoke particles to enter the varnish system; the second type is the intermediate connecting area located inside the varnish layer and overlapping with cracks, pores, or locally aged loose zones, which is the key transition section for whether a penetration channel is established; the third type is the high-adhesion area close to the varnish-pigment interface, which corresponds to the candidate penetration termination area.

[0075] To ensure that subsequent judgments are based on quantifiable channel organization rather than empirical descriptions, it is necessary to calculate the depth-progression intensity, planar extension length, and interlayer retention for each zone's penetration channel segment. The depth-progression intensity is calculated as follows: ; in, It represents the depth advance intensity of a single candidate channel segment, and is dimensionless; This represents the effective depth of the candidate channel segment in the thickness direction, in millimeters. This value is obtained from the center depth difference between the starting and ending layers of the channel segment. This indicates the total detectable thickness of the area where the candidate channel segment is located, in millimeters; This indicates the connectivity length of the candidate channel segment in the planar direction, in millimeters; This indicates the average width of the candidate channel segment, in millimeters; This represents the interlayer interruption penalty value of the candidate channel segment. It is dimensionless and is obtained by normalizing the number of breaks that occur between adjacent depth layers in the channel segment.

[0076] The core of this calculation method lies in the fact that a true infiltration channel not only extends deep inwards but also possesses a certain continuity on a plane, without frequent interruptions. The higher the depth of penetration, the more likely the contaminant is to enter from the outside in along the actual structural channels, rather than randomly remaining on a local surface.

[0077] If a channel segment has significant propulsion capability in the thickness direction and is not completely scattered in the plane, but rather has a locally extending attachment zone, then this channel segment is preferentially retained as a valid candidate channel segment; if it only appears in isolation at a certain depth point and has no continuous connection between the preceding and following layers, it is excluded. Through the above processing, the final set of candidate infiltration initiation zones, candidate infiltration termination zones, and partitioned infiltration channel segments are obtained.

[0078] S3.2: Establish a thickness-direction stratigraphic sequence, calculate the direction determination coefficient, determine the permeation direction and diffusion gradient, and form a set of dominant permeation directions in the region.

[0079] For each valid candidate channel segment obtained from S3.1 screening, a layer sequence in the thickness direction is established according to its starting region, transition region, and ending region, and the centroid shift of particle distribution, interlayer cover attenuation, and interface adhesion enhancement between each layer are calculated.

[0080] If a channel segment exhibits dense outer-layer particles, gradually moving middle-layer particles down along fissures or loose structures, and localized reinforcement of interfacial particles, it indicates that the channel segment belongs to a typical path of infiltration from the outside inwards followed by accumulation at the interface. If a channel segment is not significant in the outer layer but is strongest at the interface and gradually weakens outwards, it is more likely that the interfacial attachment structure dominates rather than infiltration.

[0081] To avoid mistaking large-area surface deposition for seepage pathways, a direction determination coefficient is introduced, which is calculated as follows: ; in, This represents the single-channel direction determination coefficient, which is dimensionless. This indicates the centroid position of the longitudinal distribution of particles in the candidate termination region. It is dimensionless and is calculated from the stratification results in step S1. The closer to the inner layer, the larger the value. This indicates the centroid position of the longitudinal distribution of particles in the candidate initiation region; it is dimensionless. This represents the interlayer retention rate of particles in the intermediate transition section. It is dimensionless and is obtained from the retention of the contact boundary between adjacent depth layers. This represents the interlayer particle retention rate of adjacent layers in the initial region, and is dimensionless. This represents the particle retention enhancement value in the interface region. It is dimensionless and is obtained by normalizing the ratio of the particle distribution intensity of the interface layer to the particle distribution intensity of the layer above it. This represents the particle retention enhancement value at the end of the intermediate transition section, and is dimensionless. , , This represents a stability constant to prevent the denominator from becoming too small, and its value ranges from 0.01 to 0.10.

[0082] The core of this calculation method lies in the fact that the permeation direction is not only determined by the presence of particles, but also by whether the particle center of gravity migrates inward, whether the channel is maintained in the intermediate layer, and whether there is enhanced retention at the interface that conforms to the characteristics of permeation termination. A higher determination coefficient indicates that the channel belongs to a permeation path that propagates from the outside in. Based on the above direction determination coefficients, the single-channel direction determination result for each valid candidate channel segment can be obtained: when... When the permeation direction of the channel segment is determined to be from the outside to the inside; when At that time, it was determined that the channel segment did not have significant infiltration characteristics from the outside to the inside.

[0083] After obtaining the single-channel direction determination result for each valid candidate channel segment, a summary analysis is performed on multiple channel segments within the same region. The proportion of channel segments with a direction determination of inward penetration from the outside to the inside is calculated out of the total number of valid channel segments in the region, denoted as . .like If the proportion exceeds a first preset threshold, such as 70%, then the area is marked as the set of dominant penetration directions, with the dominant direction being penetration from the outside in; if If the percentage is less than the second preset threshold, for example, 30%, then the area is determined to lack a dominant direction of penetration from the outside in; if If the direction determination results of each channel segment are between the first and second preset ratio thresholds, and the distribution of these results is relatively scattered with significant positive and negative conflicts, then this region is retained as a complex mixing region and its diffusion range is further identified in sub-step S3.3. Simultaneously, the diffusion gradient in the thickness direction of each channel segment, i.e., the attenuation gradient of particle distribution intensity between adjacent layers, is recorded as the basis for subsequent merging judgment.

[0084] S3.3: Expand the boundary with the region of dominant penetration direction as the center, calculate the regional diffusion consistency index, and establish path segment merging rules.

[0085] After identifying the dominant infiltration direction in the region, it is necessary to further determine the extent to which pollution has spread. In practice, for each confirmed dominant infiltration direction area, a progressive boundary expansion assessment is performed on the surrounding adjacent monitoring units.

[0086] If adjacent detection units are consistent with the current region in terms of particle distribution intensity, stratigraphic advancement characteristics, and interface retention characteristics, they are included in the same diffusion range; if adjacent detection units are close in plan but their particle advancement depth is significantly insufficient, or they only have surface darkening without mid-layer transition and interface enhancement, they are excluded from the diffusion range.

[0087] During the boundary expansion process, special attention needs to be paid to the interception effect of actual repair material layers on the pollution path. Some historical repair layers themselves can form a dense covering structure, causing smoke particles to remain on top of them and making it difficult for them to infiltrate. Therefore, when the diffusion range suddenly terminates in a certain area and the termination boundary is straight and stable, it is important to first determine whether there is a material layer blocking that location, rather than simply assuming that smoke pollution has stopped naturally.

[0088] During the boundary expansion process, for each dominant penetration direction region, the location information of all detection units included in that region is recorded. Based on the spatial distribution of these detection units, the regional diffusion range result for that region can be generated. The regional diffusion range result includes at least the set of detection units covered by the region and its spatial boundary. Simultaneously, based on the geometry of this boundary, a regional diffusion boundary result is generated. The regional diffusion boundary result includes at least the boundary's location coordinates, boundary length, and straightness or tortuosity characteristics.

[0089] To quantify consistency within a region and guide path segment merging, a regional diffusion consistency index is calculated, which is performed as follows: ; in, It represents the uniformity of regional diffusion and is dimensionless. This represents the number of detection units participating in the evaluation within the area to be merged, and is a positive integer. to It represents the directional consistency of each detection unit, is dimensionless, and is obtained by the degree of agreement between the single-channel directional determination result of each unit and the dominant direction of the region; to This represents the layer-by-layer similarity of each detection unit. It is dimensionless and is obtained by combining the degree of similarity between the advancement depth ratio and the interface retention ratio of each unit. This represents the internal blocking penalty value, which is dimensionless and is calculated from the number of suspected material layer cutoff boundaries and the degree of boundary stability.

[0090] The core of this calculation method lies in the fact that the diffusion range is not determined by the size of the black area, but by whether the direction, penetration depth, and interface behavior within a region are sufficiently consistent. Only high consistency indicates that the entire area belongs to the effective range of the same penetration path or the same set of penetration paths.

[0091] Therefore, regional diffusion boundary results and regional diffusion range results are generated in each dominant region. At the same time, path segment merging rules are established: for multiple channel segments with the same direction, contiguous boundaries and continuously changing diffusion levels, merging is performed; for channel segments that are adjacent but have opposite diffusion trends or have a clear blocking interface in the middle, merging is prohibited.

[0092] S3.4: Apply the regional diffusion boundary and path segment merging rules to the entire oil painting to generate a pollution infiltration path model.

[0093] The regional diffusion boundaries and path segment merging rules obtained in sub-step S3.3 are uniformly applied to the entire oil painting to generate a complete pollution infiltration path model. This model is not a simple geometric connection, but rather provides a complete description of each pollution path, including its starting position, propagation layer, diffusion range, termination interface, and whether it is blocked. During the model construction process, the path-level results and pollution layer adhesion relationship results generated in subsequent steps are used to verify the rationality of the path classification and the accuracy of the path-to-material layer adhesion relationship, thereby ensuring that the model can truly reflect the actual situation of pollution infiltration.

[0094] After establishing the pollution infiltration path model, the formation methods of different regions are further determined. If a pollution path has a distinct surface initiation zone, a continuous mid-layer propagation section, and an interface termination accumulation zone, it is determined to be formed by infiltration along the loose structure of the varnish after surface deposition. If a path mainly unfolds along the varnish pigment interface, while its propagation in the thickness direction is limited, it is determined to be formed mainly by the interface adhesion structure. If a region has a large number of paths, dispersed directions, and local blockages, it is determined to be formed by a combination of mixed infiltration and local adhesion.

[0095] Based on the constructed pollution infiltration path model, the path adhesion formation index is further calculated to quantify the tendency of a certain area to eventually form a pseudo-layer. The calculation method is as follows: ; in, This represents the path attachment formation index, which is dimensionless. Indicates the length of the path coverage on the main attachment interface, in millimeters; This indicates the total thickness of the corresponding area, in millimeters. This represents the interface retention enhancement value, which is dimensionless and is obtained by normalizing the ratio of the particle distribution intensity of the interface layer to the average particle distribution intensity of the intermediate layer along the path. This represents the thickness propulsion dissipation value, which is dimensionless and is obtained by normalizing the total attenuation of particle distribution intensity during the propulsion process. This represents the path bifurcation disturbance value, which is dimensionless and is obtained by normalizing the number of branches and the branch offset magnitude during the path diffusion process. This represents the interface coverage enhancement index, with a value ranging from 1.10 to 2.20. This represents the interface dwell time enhancement index, with a value ranging from 1.20 to 2.80.

[0096] The core of this calculation method lies in why a pseudo-layer eventually forms in a certain area. It's not just because the interface is contaminated, but because it stops, spreads, and adheres to the interface. The higher the path adhesion formation index, the more likely that a contamination film resembling an independent material layer will form in that area.

[0097] To further facilitate the final differentiation between pseudo-layers and true repair layers in step S4, each path needs to be assigned a path-level result. Path levels can be categorized into at least three types: shallow paths, intermediate paths, and deep interface paths. Shallow paths indicate that contamination mainly remains on the varnish surface and in the shallow layer; intermediate paths indicate that contamination has penetrated the interior of the varnish but has not yet stably accumulated at the interface; deep interface paths indicate that contamination has formed significant residue at the varnish pigment interface. This result will directly determine the subsequent steps' assessment of the pseudo-layer risk level and the true sequence restoration strategy. When performing stratification in step S4, the pseudo-layer tendency index of the shallow path area will be assigned a lower weight because the contamination has not penetrated into the interface. The focus is on whether surface cleaning can completely remove the contamination without affecting the underlying structure. The middle path area needs to be combined with the calculation results of the cross-layer continuity in step S2. If the continuity is high, the pseudo-layer risk level will be increased, and vice versa. The path attachment formation index of the deep interface path area will be the main component of the pseudo-layer tendency index in step S4 because the contamination has formed significant residue at the interface. It will directly participate in the classification of contaminated pseudo-layers and real repair layers.

[0098] In practical applications, the determination of path level results can be implemented as follows: First, based on the path termination layer recorded in the contamination infiltration path model, if the path terminates within the first third of the varnish layer thickness, it is initially classified as a shallow path; if it terminates within the middle third, it is initially classified as a medium-layer path; if it terminates within the last third or reaches the varnish pigment interface, it is initially classified as a deep interface path. Then, it is corrected by combining the path adhesion formation index: if the path adhesion formation index is higher than 0.6, the path level is upgraded by one level, for example, from a medium-layer path to a deep interface path; if it is lower than 0.3, it is downgraded by one level. Through this graded adjustment mechanism, the path level results can more accurately reflect the actual depth of contamination infiltration and the degree of interface adhesion, providing a reliable basis for the pseudo-layer tendency judgment in the subsequent step S4.

[0099] Based on this, the positional relationship between each contamination penetration path and the original material layer is further determined, and the contamination layer adhesion relationship results are generated. If the path terminates on the surface or shallow layer of the varnish layer, there is no direct adhesion relationship between it and the original pigment layer; if the path terminates inside the varnish layer, there is an indirect interlayer adhesion relationship between it and the original pigment layer; if the path terminates at the varnish-pigment interface and the path adhesion formation index is higher than a preset threshold, there is a direct interface adhesion relationship between it and the original material layer. The adhesion relationships corresponding to each path are summarized by region to generate the contamination layer adhesion relationship results. This result includes at least the adhesion layer location, adhesion range, adhesion stability, and whether there is a risk of contamination pseudo-layer formation.

[0100] In practical applications, the determination of contamination layer adhesion results can be implemented as follows: For each contamination path, first extract its termination layer coordinates. If the termination layer is located within 0 to 5 micrometers below the surface of the varnish layer, it is determined that there is no direct adhesion relationship. If the termination layer is located at a depth of more than 5 micrometers inside the varnish layer but does not touch the pigment layer interface, it is determined to be an indirect interlayer adhesion relationship. If the difference between the termination layer coordinates and the pigment layer interface coordinates is less than the detection precision, such as 2 micrometers, and the path adhesion formation index is higher than 0.5, it is determined to be a direct interface adhesion relationship. The determination results of each path are summarized by region to form the contamination layer adhesion relationship results for that region. This result can be used to verify whether the description of the termination interface in the contamination penetration path model is accurate. If the adhesion relationship result of a certain path contradicts the termination interface description in the model, it is necessary to backtrack to S3.3 to re-evaluate the boundary expansion judgment of that region to ensure that the model and the adhesion relationship analysis results corroborate each other and are logically consistent.

[0101] Finally, the pollution infiltration path model constructed in this step is used as the final output of step S3, which can be directly called in step S4 to complete the final stratification of the pollution pseudo-layer and the real remediation layer. In the technical solution of this embodiment, based on the expression results of the pollution interface continuity, the infiltration path of pollution deposition to the varnish and pigment interface is identified. By extracting candidate infiltration channel segments, determining the infiltration direction and diffusion level difference, and determining the regional diffusion boundary and merging rules, a complete pollution infiltration path model is established. This model can clearly identify the starting position, advancement layer, diffusion range, termination interface, and whether the pollution is blocked as it advances from the surface to the interior. During the model construction process, the tendency of pseudo-layers is quantified by forming an index through path attachment, and each path is assigned a shallow, middle, or deep interface path level based on this index, thereby providing evidence support at the pollution formation mechanism level for subsequent stratification.

[0102] S4: Based on the pollution penetration path model, perform layer discrimination between the pollution pseudo layer and the real repair layer, generate a layer sequence structure diagram of the real oil painting materials, and output the layer recognition results.

[0103] Based on the contamination infiltration path model obtained in step S3, a comprehensive analysis of the layer boundary structure of each region on the surface and inside the oil painting is performed. Pseudo-layer regions with depositional diffusion characteristics are identified and distinguished from the real restoration layers with material stacking characteristics. Finally, a sequence diagram of the true material structure of the oil painting is generated, and the layer identification results are output. This is achieved through the following sub-steps.

[0104] S4.1: Based on the path termination layer, attachment range, attachment stability and regional formation mode of the pollution infiltration path model, calculate the pollution pseudo-layer tendency index, and preliminarily divide the candidate pollution pseudo-layer area, candidate real remediation layer area and candidate mixed dispute area.

[0105] Each contamination penetration path output in step S3 is remapped to the location on the oil painting surface, the location of the varnish layer, and the location of the varnish pigment interface. The entire oil painting is then divided into sections for the first time, based on the path termination layer, the adhesion range, the adhesion stability, and the area formation method.

[0106] To ensure that the initial zoning has a repeatable engineering basis, a contamination pseudo-layer tendency index needs to be introduced to compare diffusion termination characteristics and material stacking characteristics within a single region. The contamination pseudo-layer tendency index is calculated as follows: ; in, This represents the pollution pseudo-layer tendency index, which is dimensionless. The path attachment formation index is dimensionless and is obtained from step S3. This indicates regional diffusion consistency, is dimensionless, and is obtained from step S3; This value represents the stability of the interface coverage. It is dimensionless and is obtained by normalizing the smoothness of the coverage boundary, the uniformity of the coverage thickness, and the continuity of the interface stay within the current area. The higher the value, the closer it is to the stable coverage characteristics of the artificial repair layer. The value represents the integrity of the material stacking. It is dimensionless and is obtained by normalizing the degree of closure of the layer boundaries within the region, the local thickness plateau, and the neatness of the boundary termination. The higher the value, the closer it is to the stacking characteristics of the actual repair material layers. u represents the adhesion enhancement index, which ranges from 1.10 to 2.40. If the smoke pollution history is long and the interface retention is significant, a higher value can be taken. If the pollution is light, a lower value can be taken.

[0107] The core of this calculation method is that the contamination pseudo-layer is not simply a layer of black matter, but a false layer formed after it diffuses to the interface and stays there. Therefore, the higher the path adhesion formation index and regional diffusion consistency, the closer its formation mechanism is to the contamination pseudo-layer. The higher the interface coverage stability and material stacking integrity, the more it resembles a real repair layer.

[0108] Based on the calculated pollution pseudo-layer tendency index, the entire oil painting is initially divided into zones: if the pollution pseudo-layer tendency index of a certain area is higher than a first preset threshold, it is classified into a candidate pollution pseudo-layer zone; if it is lower than a second preset threshold, it is classified into a candidate real restoration layer zone; if it is between the two, it is classified into a candidate mixed dispute zone. The first preset threshold is greater than the second preset threshold.

[0109] S4.2: For the candidate contaminated pseudo-layer zone, candidate real remediation layer zone, and candidate mixed dispute zone, analyze the boundary closure characteristics and thickness abrupt change characteristics of each candidate zone, calculate the regional thickness abrupt change index, and identify the boundary termination type.

[0110] For each candidate region, including the candidate contaminated pseudo-layer region, the candidate real remediation layer region, and the candidate mixed dispute region, multiple thickness profiles are extracted along its contour boundary and internal region.

[0111] In the analysis of bedding boundary closure characteristics, the continuity of the bedding boundary profile is checked for each thickness profile: if the bedding boundary profile in the profile is complete and without obvious breaks, the bedding boundary at that profile is determined to be closed; if the bedding boundary profile shows breaks, misalignment, or local absence, it is determined to be unclosed. The proportion of closed profiles among all profiles in the candidate region is statistically analyzed and recorded as the bedding boundary closure degree. A true restoration typically manifests as a relatively independent and complete overburden layer within a certain area, and its boundary contours are more likely to form a continuous closed or semi-closed structure. The values ​​are relatively high; while the contamination pseudo-layer often adheres irregularly along interfaces, cracks, or micropore channels, and the layer boundaries often show localized adhesion and localized discontinuity. The value is low.

[0112] In the analysis of thickness abrupt change characteristics, a regional thickness abrupt change index is introduced to quantify the thickness change characteristics of each candidate region. Its calculation method is as follows: ; in, This represents the abrupt change index of regional thickness, and is dimensionless. This indicates the maximum layer thickness detected by the profile within the current candidate region, in millimeters; This indicates the minimum layer thickness detected by the profile within the current candidate region, in millimeters. This indicates the average layer thickness of the profile within the current candidate region, in millimeters. The boundary fading factor is dimensionless and is obtained by normalizing the degree of thickness decrease on both sides of the region boundary. The larger the value, the more gradual the boundary. This represents the number of significant thickness transitions after the section is segmented, and is a positive integer.

[0113] The core of this calculation method is that real remediation layers often have a more defined thickness plateau and a clearer thickness transition, so the thickness abrupt change index is usually higher; while contamination pseudo-layers are more like gradual deposition and gradual weakening, with a higher boundary fading factor, thus suppressing the overall index.

[0114] Based on the calculated abrupt change index of regional thickness Identify the boundary termination type of the candidate region: if the region thickness abruptly changes... If the mutation index exceeds the preset threshold, the boundary termination type is determined to be either truncated termination or covered edge termination, and this region is more inclined towards a true repair layer; if the region thickness mutation index... If the boundary termination type is below the preset fading threshold, it is determined to be a fading termination, and the region is more likely to be a contaminated pseudo-layer. If it falls between these two thresholds, the ambiguity of the boundary termination type is retained. Based on this, the proportion of the length of the boundary with a fading termination to the total boundary length of the candidate region is calculated and recorded as the fading termination ratio. The larger the value, the more likely the area is to exhibit characteristics of a contaminated pseudo-layer.

[0115] At this point, this step has yielded the layer boundary closure for each candidate region. Regional thickness abrupt change index and the proportion of the boundary fading to the end The structural attributes of candidate regions in terms of layer boundary morphology, thickness variation and boundary closing method were quantified, providing a basis for the subsequent step S4.3 to calculate the region structure discrimination value.

[0116] S4.3: Calculate the regional structure discrimination value and make a joint judgment with the pollution pseudo-layer tendency index to confirm the attribution of pollution pseudo-layer area, real remediation layer area and area to be reviewed.

[0117] In sub-step S4.2, the layer boundary closure is obtained. Regional thickness abrupt change index and the proportion of the boundary fading to the end Based on this, a regional structure discrimination value is calculated for each candidate region. This discrimination value comprehensively considers the degree of boundary closure, the thickness abrupt change index, and the boundary termination type. If a region has poor boundary closure, weak thickness abrupt change, and the boundary is mainly terminated in a fading manner, its structure discrimination value is biased towards a contaminated pseudo-layer; if a region has relatively complete boundary, obvious thickness abrupt change, and the boundary is terminated in a covering or truncated manner, its structure discrimination value is biased towards a true remediation layer.

[0118] The method for calculating the region structure discriminant value is as follows: ; in, This represents the region structure discriminant value and is dimensionless. It represents the degree of closure of the layer boundary, is dimensionless, and is obtained by normalizing the closure ratio of the layer boundary profile within the region and the length of continuous segments. This represents the abrupt change index of regional thickness, and is dimensionless. This represents the proportion of the boundary fading termination, is dimensionless, and is obtained by the proportion of the length of the boundary that fades in the current region to the total length of the boundary. It represents the local attachment dispersion, is dimensionless, and is obtained by normalizing the dispersion of the attachment point distribution within the region; This represents the closure enhancement index, with a value ranging from 1.00 to 2.00; This represents the thickness enhancement index, with a value ranging from 1.10 to 2.30.

[0119] The core of this calculation method lies in the fact that a real remediation layer resembles a material because it is more closed, has a thicker plateau, and has a more defined boundary; while a pseudo-contamination layer is a false layer because it is often not closed, its boundaries gradually disappear, and its attachments are discrete. The regional structure discrimination value quantifies the degree to which it resembles a real material.

[0120] For candidate mixed disputed areas, the discriminant value needs to be used in conjunction with the pollution pseudo-layer tendency index obtained in sub-step S4.1. Specifically, during the collaborative judgment, the numerical levels of the two are compared first: if the pollution pseudo-layer tendency index is higher than the first preset threshold and the structural discriminant value is lower than the second preset threshold, it indicates that the area is highly similar to a pollution pseudo-layer in terms of formation mechanism but exhibits abnormal integrity in structural morphology. In this case, the path attachment formation index and interface retention enhancement value of the area in step S3 need to be reviewed. If both are significantly higher, it is still classified as a confirmed pollution pseudo-layer area; if only one side is higher, it is downgraded to a pending review area. If the pseudo-layer tendency index is lower than the second preset threshold and the structural discriminant value is higher than the first preset threshold, it indicates that the area is highly similar to a real remediation layer in structural morphology but lacks evidence of material stacking in formation mechanism. In this case, the intra-layer aggregation continuity and cross-layer continuity of the area in step S2 need to be reviewed. If the intra-layer aggregation continuity is significantly higher than the cross-layer continuity, it is still classified as a confirmed real remediation layer area; if the two are close, it is downgraded to a pending review area. If the pseudo-layer tendency index is high and the structure discrimination value also leans towards the pseudo-layer, it is directly classified into the set of confirmed contamination pseudo-layers; if both lean towards the true remediation layer, it is classified into the set of confirmed true remediation layers; if the two directions are inconsistent, it is marked as the set of areas to be reviewed, for remediation personnel to focus on during subsequent manual review.

[0121] S4.4: Map the confirmed contaminated pseudo-layer areas and real restoration layers back to the spatial coordinates of the oil painting, generate a real material sequence structure diagram of the oil painting, and output the area confirmation level and dispute marker.

[0122] The identified contaminated pseudo-layers and genuine restoration layers are remapped onto the spatial coordinates of the entire painting, and disputed areas are marked with additional markers to generate a true material sequence diagram of the painting. This diagram should contain at least the following types of sequence information: the location of the original pigment layer, the location of the varnish layer, the location of the contaminated pseudo-layer, the location of the genuine restoration layer, and the location of the disputed area.

[0123] When generating the sequence structure diagram, for pseudo-layer confirmation areas, they should be removed from the historical repair layer candidate layers, and their true attachment relationships should be reconstructed so that they are marked as contaminated attachment layers rather than repair material layers in the sequence representation; for true repair layer confirmation areas, their sequence position as independent material layers should be retained, and their coverage and attachment interfaces should be recorded; for areas to be reviewed, they should be output as separate layers without being forcibly classified.

[0124] To quantify the credibility of the final output, the regional confirmation level is calculated as follows: ; in, Indicates the regional confirmation level; dimensionless. This represents the region structure discriminant value and is dimensionless. This represents the pseudo-layer tendency index, which is dimensionless. This represents the conflict correction value, which is dimensionless and is obtained by normalizing the degree of inconsistency between the structure discrimination direction and the formation mechanism discrimination direction. If the two are consistent, the value is lower; if the two are in conflict, the value is higher.

[0125] The core of this calculation method lies in the fact that the final output should not only include classification labels, but also the credibility of the judgment. The higher the region confirmation level, the more consistent the formation mechanism and structural performance of the region, and the more credible the conclusion. If the conflict correction value is too high, it indicates that although the region has a tendency, it still needs to be reviewed.

[0126] The final output of the layer identification results includes the area boundaries and area confirmation levels generated in this step, the contamination layer adhesion results generated in step S3.4, and the areas to be reviewed as dispute markers confirmed in step S4.3. The contamination layer adhesion results include adhesion layer location information; the area boundaries are obtained by mapping the confirmed pseudo-layer areas and real restoration layer areas back to the spatial coordinates of the oil painting; the area confirmation level is calculated in this step; and the dispute markers correspond to the spatial location of the areas to be reviewed. Restoration personnel can directly determine which areas can be treated for contamination, which areas should be preserved as historical restoration traces, and which areas require manual review based on these results, thus truly solving the technical problem of smoke-contaminated films being misjudged as historical restoration layers.

[0127] In the technical solution of this disclosure embodiment, a comprehensive pollution infiltration path model and regional structural characteristics are used to perform a layered discrimination between the pollution pseudo-layer and the real restoration layer. By calculating the pseudo-layer tendency index and the regional structure discrimination value, the candidate pollution pseudo-layer area, the candidate real restoration layer area, and the candidate mixed dispute area are judged collaboratively. In the mixed dispute area, a two-way verification logic of mechanism and structure is introduced. Finally, a real material layer sequence structure diagram of the oil painting is generated, which includes the original pigment layer, varnish layer, pollution pseudo-layer, real restoration layer and dispute area location. At the same time, the area confirmation level and dispute mark are output, so that restorers can clearly identify which areas belong to the cleanable smoke pseudo-layer and which areas belong to the historical restoration traces that should be retained. This fundamentally solves the technical problem of the smoke pollution film being misjudged as the historical restoration layer.

[0128] According to embodiments of this disclosure, an electronic device is also provided, which may include a processor, a communications interface, a memory, and a communication bus, wherein the processor, the communications interface, and the memory communicate with each other via the communication bus. The processor can invoke logical instructions stored in the memory to execute the methods provided in the above embodiments.

[0129] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0130] On the other hand, this disclosure also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the methods provided in the above embodiments.

[0131] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0132] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0133] It should be understood that the above embodiments are only used to illustrate the technical solutions of this disclosure, and not to limit them; although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure.

Claims

1. A method for identifying and detecting layered materials in the oil painting restoration process, characterized in that, The method includes: S1: Perform multi-scale imaging and local micro-area profile detection on the surface of the oil painting to generate a spatial distribution map of contamination deposition; S2: Based on the aforementioned spatial distribution map of contamination deposition, construct an expression of the interface continuity between the smoke deposition layer and the material layer, and distinguish between through-diffusion structures, interface attachment structures, and transitional mixed structures; S3: Based on the aforementioned interface continuity expression, identify the penetration path of contaminant deposition to the varnish-pigment interface and generate a contaminant penetration path model; S4: Based on the pollution penetration path model, perform layer discrimination between the pollution pseudo layer and the real repair layer, generate a layer sequence structure diagram of the real oil painting materials, and output the layer recognition results.

2. The method for layered material identification and detection in the oil painting restoration process according to claim 1, characterized in that, S1 includes: The surface of the oil painting is divided into continuous detection units. Multi-scale imaging and multi-angle illumination are performed on each detection unit to calculate the surface contamination candidate index and screen the surface candidate contamination anomaly area. Local micro-area profile detection is performed on the candidate surface contamination anomaly areas to statistically analyze the particle distribution intensity at each depth layer; The particle distribution intensity at different depths is sorted in the thickness direction to form a longitudinal particle distribution curve. The distribution state of particles in the surface, interior and interface is identified based on the peak position of the longitudinal particle distribution curve, the position of the particle interlayer transfer boundary is determined, the centroid of the longitudinal particle distribution is calculated, and it is converted into a particle penetration level identifier. The particle penetration level identifiers and interlayer transfer boundary positions of each detection unit are mapped back to the overall spatial position of the oil painting, generating a spatial distribution map of pollution deposition and a layered identification map of pollution particle depth.

3. The method for identifying and detecting layered materials in the oil painting restoration process according to claim 1, characterized in that, S2 includes: The spatial distribution map of the contamination deposition is subjected to secondary discretization processing, and the discrete detection units are aggregated into a set of adjacent segments of particles in the same layer and a set of candidate connection segments across layers. Calculate the intralayer aggregation continuity of the adjacent segments of the same-layer particles and the cross-layer continuity of the cross-layer candidate connection segments; Based on the same-layer aggregation continuity and the cross-layer continuity, the longitudinal diffusion attribution ratio is calculated, and the same-layer particle adjacent segment and the cross-layer candidate connection segment are divided into through-diffusion structure, interface attachment structure and transitional mixing structure. The three types of structure sets are remapped onto the overall space of the oil painting to generate a continuous expression of the contaminated interface that includes structural difference identifiers.

4. The method for identifying and detecting layered materials in the oil painting restoration process according to claim 3, characterized in that, The step of dividing the adjacent segments of the same-layer particles and the candidate cross-layer connection segments into a through-diffusion structure, an interface attachment structure, and a transitional hybrid structure includes: If the longitudinal diffusion attribution ratio is higher than the high-level threshold, it is classified into the through-diffusion structure set; If the longitudinal diffusion attribution ratio is lower than the low-level threshold and the same-layer aggregation continuity is higher than the preset continuity threshold, then it is classified into the interface attachment structure set. If the longitudinal diffusion attribution ratio is between the low-level threshold and the high-level threshold, it is classified into the transitional mixed structure set.

5. The method for identifying and detecting layered materials in the oil painting restoration process according to claim 1, characterized in that, S3 includes: The continuous regions in the interface continuity representation result are projected onto different depth domains and divided into partitioned penetration channel segments. The depth propulsion intensity of each partitioned penetration channel segment is calculated, and effective candidate channel segments are selected based on the depth propulsion intensity. For each valid candidate channel segment, a thickness-direction stratigraphic sequence is established, and the direction determination coefficient is calculated. Based on the direction determination coefficient, the direction determination result of a single channel is obtained. Multiple valid candidate channel segments in the same region are summarized, and a set of regional dominant permeation directions is formed based on the summary results. The boundary is expanded with the region of dominant penetration direction as the center, the regional diffusion consistency index is calculated, and the regional diffusion boundary results are generated. At the same time, path segment merging rules are established. The regional diffusion boundary results and the path segment merging rules are applied to the entire oil painting to generate a pollution infiltration path model.

6. The method for identifying and detecting layered materials in the oil painting restoration process according to claim 5, characterized in that, The contamination infiltration path model is constructed based on the path-level results and the contamination layer adhesion relationship results; The path level results include shallow path, medium path and deep interface path. The shallow path indicates that the contaminant mainly stays on the surface and shallow layer of the varnish. The medium path indicates that the contaminant has entered the interior of the varnish but has not yet accumulated stably at the interface. The deep interface path indicates that the contaminant has formed a significant residue at the varnish pigment interface. The contamination layer adhesion relationship results include no direct adhesion relationship, indirect interlayer adhesion relationship, and direct interface adhesion relationship. The absence of a direct adhesion relationship indicates that the path terminates at the surface of the varnish layer or the shallow layer of the varnish. The indirect interlayer adhesion relationship indicates that the path terminates inside the varnish layer. The direct interface adhesion relationship indicates that the path terminates at the varnish pigment interface and the path adhesion formation index is higher than a preset threshold.

7. The method for identifying and detecting layered materials in the oil painting restoration process according to claim 1, characterized in that, S4 includes: Based on the path termination layer, attachment range, attachment stability and regional formation mode of the pollution infiltration path model, the pollution pseudo-layer tendency index is calculated, and candidate pollution pseudo-layer areas, candidate real remediation layer areas and candidate mixed dispute areas are initially divided according to the pollution pseudo-layer tendency index. For the candidate contaminated pseudo-layer region, candidate real remediation layer region, and candidate mixed dispute region, the boundary closure characteristics and thickness abrupt change characteristics of each candidate region are analyzed, the regional thickness abrupt change index is calculated, and the boundary termination type is identified. Based on the layer boundary closure characteristics, the regional thickness abruptness index, and the boundary termination type, the regional structure discrimination value is calculated and used in conjunction with the pseudo-layer tendency index to determine the attribution of the contaminated pseudo-layer area, the real remediation layer area, and the area to be reviewed. The confirmed contaminated pseudo-layer areas and real restoration layers are mapped back to the spatial coordinates of the oil painting to generate a real material sequence structure diagram of the oil painting and output the area confirmation level. The confirmed areas to be reviewed are output as dispute markers. The area confirmation level is calculated by the area structure discrimination value and the contaminated pseudo-layer tendency index.

8. The method for identifying and detecting layered materials in the oil painting restoration process according to claim 7, characterized in that, The collaborative judgment includes: For the candidate mixed dispute area, when the pollution pseudo-layer tendency index is higher than its high-level threshold and the regional structure discrimination value is lower than its low-level threshold, if the path attachment formation index and the interface retention enhancement value are both significantly higher, it is classified as a pollution pseudo-layer confirmation area; otherwise, it is downgraded to a pending review area. When the pollution pseudo-layer tendency index is lower than its low-level threshold and the regional structure discrimination value is higher than its high-level threshold, if the intra-layer aggregation continuity is significantly higher than the inter-layer continuity, it is classified as a real remediation layer confirmation area; otherwise, it is downgraded to a pending review area.

9. An electronic device, characterized in that, The electronic device includes a memory and at least one processor, the memory storing a computer program, and the processor executing the computer program to implement the layered material identification and detection method for the oil painting restoration process according to any one of claims 1-8.

10. A computer storage medium, characterized in that, It stores a computer program, which, when executed, implements a layered material identification and detection method for the oil painting restoration process according to any one of claims 1-8.

Images