Building material surface quality identification method and system

Abstract

The invention discloses a method and system for identifying the surface quality of building materials, and relates to the technical field of building decoration engineering. The quality identification method comprises the following steps: recording the surface temperature and spraying pressure information of the target building material surface during the spraying process, and constructing a two-dimensional actual bonding index set of the target building material surface; analyzing the overall bonding effect of the coating on the target building material surface in combination with a statistical averaging algorithm, so as to identify the current target building material as a qualified bonding material, and issuing a surface quality analysis instruction; irradiating the surface of the qualified bonding material at multiple angles, and combining a sliding window convolution sampling mechanism to analyze the degree of reflection deviation of each pixel point on the surface of the qualified bonding material; and after performing comparative analysis, identifying the current qualified bonding material as a surface quality finished material. The invention significantly improves the real-time performance and accuracy of building material surface quality identification.

Description

Technical Field

[0001] The present invention relates to the technical field of building decoration engineering, in particular to a method and system for identifying the surface quality of building materials. Background Art

[0002] In modern building decoration, polymer-based spray coating materials are widely used in exterior walls, floors, and tunnel linings to enhance their anti-seepage, protective or aesthetic effects. During spray coating construction, the bonding strength and film-forming uniformity between the coating and the base surface directly affect the life and performance of the coating. Therefore, how to efficiently identify the coating bonding quality and accurately judge the spray surface quality at the construction site has become a key issue in current coating construction control. Especially with the gradual popularization of non-contact detection, intelligent recognition methods based on image acquisition and analysis are becoming an important development direction for surface quality identification of building materials.

[0003] To address the above-mentioned problem of spraying quality identification, existing methods mostly rely on manual sampling and post-peeling tests, making it difficult to achieve online quality identification during the spraying process and resulting in severe detection lag. Among them, some non-contact methods based on image recognition and infrared detection can preliminarily determine the surface film formation situation, but it is difficult to establish a theoretical prediction model for spraying adhesion performance based on the actual parameters of the spraying process, resulting in difficulty in achieving a coordinated evaluation of material properties and reflection response. At the same time, current multi-angle reflection image analysis has not yet formed a systematic method, and lacks quantitative analysis of the image reflection change trend under different polarization angles, resulting in unstable automatic identification results of building material surface quality, low accuracy, and poor adaptability. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method and system for identifying the surface quality of building materials, which solve the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for identifying the surface quality of building materials comprises the following steps:

[0006] S1. Using an image acquisition device to perform standard viewing angle imaging of the target building material surface, establish a two-dimensional pixel coordinate system, record the surface temperature and spraying pressure information of the target building material surface during the spraying process, and construct a two-dimensional actual bonding index set for the target building material surface;

[0007] S2. Extract features from the two-dimensional actual bonding index set and, combined with a statistical averaging algorithm, analyze the overall bonding effect of the coating on the surface of the target building material to identify the current target building material as a qualified bonding material and issue a surface quality analysis instruction;

[0008] S3. After receiving the surface quality analysis instruction, the qualified bonding material surface is illuminated at multiple angles. In combination with the sliding window convolution sampling mechanism, the overall deviation of the reflection intensity of each pixel point on the qualified bonding material surface at different polarization angles is analyzed to determine the degree of reflection deviation of each pixel point on the qualified bonding material surface.

[0009] S4. Compare and analyze the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with a preset reflection deviation threshold value, and identify the current qualified bonding material as a surface quality finished material.

[0010] Preferably, the specific steps of S1 include:

[0011] S11, using the actual sprayed surface of the building material to be sprayed as the target building material surface, imaging the target building material surface from a standard viewing angle using an image acquisition device, and establishing a two-dimensional pixel coordinate system for the target building material surface after geometric benchmark calibration and correction;

[0012] S12. Using a digital pressure sensor built into the spraying equipment to record the output pressure value during the spraying process in real time, with a set time length t as the sampling period, to generate pressure data with a timestamp, using a laser trajectory recorder to collect the spray gun position information of each pixel point during the spraying process, synchronously matching the timestamp of the pressure data with the spray gun position information, projecting the matched pressure data into a two-dimensional pixel coordinate system using an interpolation algorithm, and extracting the actual spray pressure value of each pixel point;

[0013] S13. Use the deployed infrared thermal imager to perform real-time temperature measurement and scanning on the surface of the target building material, collect the thermal radiation intensity of the target building material surface in a non-contact manner, and output a two-dimensional temperature distribution map with a resolution consistent with the two-dimensional pixel coordinate system. Combined with the image registration mechanism, the two-dimensional temperature distribution map and the two-dimensional pixel coordinate system are geometrically corrected to extract the actual surface temperature value of each pixel point.

[0014] Preferably, S14, conditionally indexing the experimental database according to the material type and coating specifications to obtain a standard spraying pressure value and a standard surface temperature value;

[0015] S15. Based on the actual spraying pressure value and the actual surface temperature value of each pixel point extracted, the actual bonding coefficient of each pixel point is determined, specifically: Where Bnj(x,y) represents the actual bonding coefficient of the corresponding pixel point, T0 represents the standard surface temperature value, Ts(x,y) represents the actual surface temperature value of the corresponding pixel point, P0 represents the standard spraying pressure value, Ps(x,y) represents the actual spraying pressure value of the corresponding pixel point, and (x,y) represents the coordinate position of the corresponding pixel point;

[0016] S16. Construct a two-dimensional actual bonding index set of the target building material surface based on the actual bonding coefficient of each pixel point and in combination with the coordinate position of each pixel point.

[0017] Preferably, the specific steps of S2 include:

[0018] S21. By extracting features from the two-dimensional actual bonding index set and combining it with a statistical averaging algorithm, the overall bonding effect of the coating on the surface of the target building material is analyzed to obtain the overall bonding coefficient of the coating on the surface of the target building material;

[0019] S22. Based on the value of the overall adhesion coefficient of the coating on the surface of the target building material, determine whether the overall adhesion level of the coating in the current spraying process is qualified, and issue corresponding surface quality analysis instructions, specifically:

[0020] When the overall adhesion coefficient of the paint on the surface of the target building material is equal to one, it means that the overall adhesion level of the paint in the current spraying process is qualified, and the current target building material is identified as a qualified adhesion material, and a surface quality analysis instruction is issued; otherwise, it means that the overall adhesion level of the paint in the current spraying process is unqualified, and the current target building material is identified as a surface quality defective material.

[0021] Preferably, the specific steps of S3 include:

[0022] S31. After receiving the surface quality analysis instruction, the qualified bonding material surface is illuminated at multiple angles to obtain a multi-angle polarized reflection frame grayscale image to form a multi-angle reflection vector, specifically:

[0023] After receiving the surface quality analysis instruction, the reflective light source emitted by the electrically controlled polarized LED array with polarization angle adjustment function is used to illuminate the surface of the qualified adhesive material at polarization angles of 30°, 60°, and 90°. A high-sensitivity CMOS image sensor is used with a polarization filter group to synchronously capture the reflected frame images at each polarization angle. After grayscale processing, the reflected frame grayscale images at each polarization angle are obtained.

[0024] After spatially calibrating the grayscale images of the reflection frames at each polarization angle using an image calibration plate, the grayscale images of the reflection frames at each polarization angle are perspective-corrected using a homography matrix model. The corrected grayscale images of the reflection frames at each polarization angle are then registered into a two-dimensional pixel coordinate system.

[0025] Based on the resolution of the image acquisition device, the grayscale images of the reflected frames at each polarization angle after registration are divided according to the imaging resolution, and the grayscale value of each pixel is converted into the reflection intensity value of the corresponding pixel through the image grayscale conversion model. The multi-angle reflection vector of the qualified bonding material surface is constructed in the form of: Rθ(x,y)=[R30(x,y),R60(x,y),R90(x,y)]; where Rθ(x,y) represents the multi-angle reflection vector, θ=30°, 60°, 90°, R30(x,y) represents the reflection intensity value of the pixel point (x,y) at a polarization angle of 30°, R60(x,y) represents the reflection intensity value of the pixel point (x,y) at a polarization angle of 60°, and R90(x,y) represents the reflection intensity value of the pixel point (x,y) at a perpendicular polarization angle.

[0026] Preferably, S32, constructing a third-order three-channel surface tensor of the qualified bonding material surface according to the multi-angle reflection vectors of the qualified bonding material surface, specifically:

[0027] The multi-angle reflection vectors of the qualified bonding material surface are normalized and combined with the sliding window convolution sampling mechanism to construct the third-order three-channel surface tensor of the qualified bonding material surface. The specific expression is:

[0028] ;

[0029] Where M(x,y,k) represents the third-order three-channel surface tensor, k represents the index number of the corresponding reflection channel, and k=1, 2, and 3 represent the channel index numbers corresponding to R30(x,y), R60(x,y), and R90(x,y), respectively.

[0030] Preferably, S33, based on constructing a third-order three-channel surface tensor of a qualified bonding material surface, summing the reflection intensity values ​​of all pixel points in the corresponding reflection channels in the third-order three-channel surface tensor, obtaining the total reflection intensity value of each reflection channel, and combining the statistical averaging algorithm to obtain the global average reflection intensity value of each reflection channel.

[0031] Preferably, S34, the global average reflection intensity value of each reflection channel is associated with the third-order three-channel surface tensor, and the overall deviation of the reflection intensity of each pixel point on the surface of the qualified bonding material at different polarization angles is analyzed to determine the degree of reflection deviation of each pixel point on the surface of the qualified bonding material, specifically: Where Py(x,y) represents the reflection deviation coefficient of the corresponding pixel point, M(x,y,k) represents the third-order three-channel surface tensor of the qualified bonding material surface, Ravg k Represents the global average reflection intensity value of the kth reflection channel.

[0032] Preferably, the specific steps of S4 include:

[0033] S41, comparing and analyzing the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with a preset reflection deviation threshold value to determine whether the surface of the qualified bonding material is in a flat state, specifically:

[0034] If the reflection deviation coefficient of the corresponding pixel exceeds the reflection deviation threshold, it means that the reflection deviation degree of the corresponding pixel is abnormal, and the corresponding pixel is divided into an abnormal deviation pixel set; otherwise, the corresponding pixel is divided into a normal deviation pixel set;

[0035] When the number of pixels in the abnormal deviation pixel set exceeds 5% of the total number of pixels on the surface of the qualified bonding material, it means that the surface of the qualified bonding material is not in a flat state, and the qualified bonding material is identified as a surface quality defective material;

[0036] When the number of pixels in the abnormal deviation pixel point set does not exceed 5% of the total number of pixels in the surface of the qualified bonding material, it means that the surface of the qualified bonding material is in a flat state, and the qualified bonding material is identified as a surface quality finished material.

[0037] Building material surface quality identification system, including adhesion analysis module, effect determination module, reflection deviation analysis module and quality identification module;

[0038] The bonding analysis module is used to establish a two-dimensional pixel coordinate system based on the standard viewing angle imaging of the target building material surface by the image acquisition device, record the surface temperature and spraying pressure information of the target building material surface during the spraying process, and construct a two-dimensional actual bonding index set of the target building material surface;

[0039] The effect determination module is used to extract features from the two-dimensional actual bonding index set and, combined with a statistical averaging algorithm, analyze the overall bonding effect of the coating on the surface of the target building material to identify the current target building material as a qualified bonding material and issue a surface quality analysis instruction;

[0040] The reflection deviation analysis module is used to, after receiving the surface quality analysis instruction, illuminate the qualified bonding material surface from multiple angles and, in combination with the sliding window convolution sampling mechanism, analyze the overall deviation of the reflection intensity of each pixel point on the qualified bonding material surface at different polarization angles to determine the degree of reflection deviation of each pixel point on the qualified bonding material surface;

[0041] The quality identification module is used to compare and analyze the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with the preset reflection deviation threshold value, and identify the current qualified bonding material as a surface quality finished material.

[0042] The present invention provides a method and system for identifying the surface quality of building materials, which has the following beneficial effects:

[0043] (1) The surface quality identification method and system of building materials provided by the present invention can realize the integrated judgment of multi-source parameter collection, theoretical evaluation of bonding performance and quantitative analysis of reflection image offset at the spraying construction site, and has non-contact, automatic, full coverage and refined identification capabilities; by establishing a spraying data space mapping model based on two-dimensional pixel coordinates, and comprehensively considering surface temperature, spraying pressure and polarized light reflection, it can dynamically identify the overall bonding effect of the coating on the surface of building materials during the spraying process and the overall flatness of the surface after spraying, thereby effectively avoiding the problems of traditional detection methods relying on manual labor, delayed response and frequent misjudgment, significantly improving the real-time, accuracy and construction closed-loop control capabilities of surface quality identification, and is suitable for the fine detection needs of spraying quality in various complex building environments.

[0044] (2) By synchronously recording the pressure data output by the spraying equipment and the surface temperature data collected by infrared thermal imaging, and combining the laser trajectory matching and image registration mechanism, the multi-temporal and spatial physical parameters of the spraying process can be accurately projected into the two-dimensional pixel coordinate system of the building material surface. On this basis, the standard construction parameters in the material experimental library are introduced to construct a two-dimensional actual bonding index set with spatial resolution; in order to realize the judgment of the overall bonding effect of the coating on the surface of the target building material, thereby identifying qualified bonding materials, truly realizing non-destructive, online and quantitative spraying performance prediction, and significantly improving the technical defects of the existing recognition methods such as insufficient recognition ability of coating adhesion, subjective judgment, and difficulty in dynamic prediction.

[0045] (3) By setting up an electrically controlled polarization angle light source and a multi-channel image sensor, multi-angle reflection acquisition is performed on the surface of qualified adhesive materials. Combined with image perspective correction and grayscale normalization processing technology, a third-order three-channel surface reflection tensor model is constructed in a two-dimensional pixel coordinate system. On this basis, a sliding window sampling mechanism and a reflection deviation coefficient calculation method are introduced to realize the reflection anomaly sensitivity judgment at the pixel level. This mechanism not only improves the detection and resolution ability of local inhomogeneous areas, but also solves the problem that existing methods are difficult to accurately quantify polarization response changes, making the surface quality analysis process more stable, accurate and robust, and is suitable for the identification of the quality of sprayed surface products in modular construction sites.

[0046] (4) By establishing a unified two-dimensional pixel coordinate system and introducing a high-precision image registration and perspective transformation correction mechanism, pixel-level consistent mapping and standard division are achieved in multi-angle imaging results, so that the quality assessment of the entire spray surface has fine-grained and highly aligned spatial expression capabilities; on this basis, each pixel point can be synchronously associated with the actual spray pressure, surface temperature and multi-source parameters of the reflected light response to form a high-dimensional data feature vector with clear spatial significance, providing a unified coordinate support for subsequent tensor construction and reflection deviation judgment; this pixel-level division mechanism significantly improves the recognition accuracy of micro-quality fluctuations, and at the same time has higher data versatility and regional recognition stability, solving the problem of easily ignoring edge effects and local anomalies in traditional judgment modes based on macro-regions and image blocks. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Schematic diagram of the process of the building material surface quality identification method of the present invention;

[0048] Figure 2 This is a block diagram of the building material surface quality identification system of the present invention;

[0049] Figure 3 This is a logical thinking diagram of the method for identifying the surface quality of building materials of the present invention;

[0050] Figure 4 Schematic diagram of the surface quality identification process of building materials according to the present invention. DETAILED DESCRIPTION

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0052] Example 1, please refer to Figure 1 and Figure 3 The present invention provides a method for identifying the surface quality of building materials, comprising the following steps:

[0053] S1. Using an image acquisition device to perform standard viewing angle imaging of the target building material surface, establish a two-dimensional pixel coordinate system, record the surface temperature and spraying pressure information of the target building material surface during the spraying process, and construct a two-dimensional actual bonding index set for the target building material surface;

[0054] S2. Extract features from the two-dimensional actual bonding index set and, combined with a statistical averaging algorithm, analyze the overall bonding effect of the coating on the surface of the target building material to identify the current target building material as a qualified bonding material and issue a surface quality analysis instruction;

[0055] S3. After receiving the surface quality analysis instruction, the qualified bonding material surface is illuminated at multiple angles. In combination with the sliding window convolution sampling mechanism, the overall deviation of the reflection intensity of each pixel point on the qualified bonding material surface at different polarization angles is analyzed to determine the degree of reflection deviation of each pixel point on the qualified bonding material surface.

[0056] S4. Compare and analyze the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with a preset reflection deviation threshold value, and identify the current qualified bonding material as a surface quality finished material.

[0057] In this embodiment, by constructing a unified two-dimensional pixel coordinate system during the spraying construction process and fusing the information collected by image acquisition equipment, pressure sensors, infrared thermal imagers and various sensing devices, the spatial one-to-one correspondence between spraying parameters and image features is effectively achieved, thereby generating a two-dimensional actual bonding index set with position resolution, which provides a physical basis and data support for subsequent bonding quality judgment; this method not only realizes the non-destructive, online and quantitative identification of the bonding performance of building material coatings, but also accurately evaluates the overall bonding level of the coating through the statistical mean analysis method, thereby avoiding the hysteresis and local errors caused by relying on manual sampling detection; by implementing multi-angle deviation on the surface of qualified bonding materials, the surface of the coating can be accurately identified. Vibration light irradiation is used, and a three-channel reflection tensor model is established based on the sliding window convolution sampling mechanism, which realizes multi-angle reflection intensity analysis and deviation evaluation at the pixel level, improves the ability to identify uneven areas on the surface of qualified adhesive materials, and can effectively detect microscopic abnormal phenomena such as coating thinning and spray hollowing, ensuring the overall quality stability and uniformity of the sprayed surface; by comparing with the preset reflection deviation threshold, this method can realize automatic classification and refined judgment of the quality of the finished surface of building materials, and has significant technical advantages of reasonable structure, fast response and high recognition accuracy. It is particularly suitable for modern architectural decoration projects with high requirements for high-quality spraying, and can achieve full-process, full-pixel and high-reliability quality control goals.

[0058] Example 2, please refer to Figure 1 and Figure 4 , specifically: S1 specific steps include:

[0059] S11, using the actual sprayed surface of the building material to be sprayed as the target building material surface, imaging the target building material surface from a standard viewing angle using an image acquisition device, and establishing a two-dimensional pixel coordinate system for the target building material surface after geometric benchmark calibration and correction;

[0060] The two-dimensional pixel coordinate system is based on the surface image of the target building material captured by the image acquisition device. It is established through standard perspective imaging and geometric calibration correction. It is a coordinate reference system used to identify the corresponding spatial position of each pixel point in the image on the sprayed surface. The establishment of this coordinate system enables the subsequent multi-source data of temperature, pressure, and reflection intensity collected to be accurately mapped to each pixel point in the actual spraying area, ensuring that each parameter has a unified spatial reference and achieving accuracy and consistency in data fusion analysis; the target surface is photographed at a standard angle using the image acquisition device; spatial geometric calibration is performed using a geometric reference calibration plate; the image is distortion corrected and spatially registered in combination with the perspective transformation model, thereby establishing a two-dimensional pixel coordinate system that is consistent with the image resolution and can accurately reflect the corresponding relationship of physical positions; this coordinate system plays a key role in data space alignment, pixel parameter mapping, and regional positioning analysis, and is the basic support framework for subsequent bonding index calculation, tensor construction, and quality identification;

[0061] S12. Using a digital pressure sensor built into the spraying equipment to record the output pressure value during the spraying process in real time, with a set time length t as the sampling period, to generate pressure data with a timestamp, using a laser trajectory recorder to collect the spray gun position information of each pixel point during the spraying process, synchronously matching the timestamp of the pressure data with the spray gun position information, projecting the matched pressure data into a two-dimensional pixel coordinate system using an interpolation algorithm, and extracting the actual spray pressure value of each pixel point;

[0062] The actual spraying pressure value refers to the actual output pressure intensity of the paint sprayed from the spray gun nozzle acting on the surface of the building material during the spraying construction process, which specifically reflects the force applied to the spray material in different pixel areas. This parameter directly affects the adhesion effect and film-forming quality of the coating, and is a key physical indicator for evaluating spraying consistency and bonding performance. The acquisition process is as follows: first, the digital pressure sensor integrated on the spraying equipment is used to collect output pressure data in real time with a set sampling period t and record the timestamp. At the same time, a laser trajectory recorder is used to synchronously obtain the spraying path and position information of the spray gun. The pressure data is matched with the spatial trajectory of the spray gun through the timestamp to identify the pixel area where the spray gun is located at the corresponding moment. The discrete pressure data is then smoothly mapped to the entire two-dimensional pixel coordinate system with the help of an interpolation algorithm, thereby assigning a corresponding actual spraying pressure value to each pixel point in the image. At the same time, it serves as a bridge connecting the spraying process and the image recognition model, and is an important basic data source for the subsequent construction of the actual bonding coefficient.

[0063] S13. Use the deployed infrared thermal imager to perform real-time temperature measurement and scanning on the surface of the target building material, collect the thermal radiation intensity of the target building material surface in a non-contact manner, and output a two-dimensional temperature distribution map with a resolution consistent with the two-dimensional pixel coordinate system. Combined with the image registration mechanism, the two-dimensional temperature distribution map and the two-dimensional pixel coordinate system are geometrically corrected to extract the actual surface temperature value of each pixel point.

[0064] It should be noted that the actual surface temperature value refers to the true surface temperature value reflected by each pixel point on the surface of the target building material under infrared thermal imaging monitoring during the spraying construction process, reflecting the heat conduction state and construction thermal stability of the corresponding pixel point during the coating film formation process; this temperature value directly affects the fluidity, curing rate and bonding performance of the coating, and is an important physical variable for constructing the actual bonding model; the infrared thermal imager deployed in the construction area continuously collects the surface thermal radiation signal in a non-contact manner, converts it into temperature information, and generates a two-dimensional temperature distribution map consistent with the resolution of the image acquisition system; then, the image registration mechanism is applied to geometrically correct and spatially align the thermal image with the original two-dimensional pixel coordinate system to achieve pixel-level correspondence; the actual surface temperature value of each position is extracted pixel by pixel from the corrected image;

[0065] Specifically, S14, according to the material type and coating specifications, conditionally index the experimental database to obtain a standard spraying pressure value and a standard surface temperature value;

[0066] S15. Based on the actual spraying pressure value and the actual surface temperature value of each pixel point extracted, the actual bonding coefficient of each pixel point is determined, specifically: Where Bnj(x,y) represents the actual bonding coefficient of the corresponding pixel point, T0 represents the standard surface temperature value, Ts(x,y) represents the actual surface temperature value of the corresponding pixel point, P0 represents the standard spraying pressure value, Ps(x,y) represents the actual spraying pressure value of the corresponding pixel point, and (x,y) represents the coordinate position of the corresponding pixel point;

[0067] The actual bonding coefficient of each pixel point defined by the formula in S15 is an indicator constructed based on the spraying process parameters and the thermal response characteristics of the material. Its purpose is to quantitatively compare the actual spraying pressure value and the actual surface temperature value during construction with the standard ideal parameters extracted from the experiment to calculate the bonding adaptability of the pixel point; in the formula, Indicates the degree of deviation between the current pixel surface temperature and the standard temperature. The closer the temperature is to the standard value, the closer this item is to 1. The second part It represents the degree of deviation between the current spraying pressure and the standard spraying pressure, and a value close to 1 indicates the optimal match. The mean of the two items is the final bonding coefficient, which reflects the influence of both heat and force on the bonding effect. This structural design is highly consistent with the pain point pointed out in the background technology that the existing non-contact identification method is difficult to introduce the actual parameters of the spraying process into the bonding performance analysis. The existing solutions mostly rely on the appearance judgment of the film-forming results themselves, and fail to reflect the inherent causal relationship between the construction process and the material properties. The introduction of this actual bonding coefficient model not only realizes the bridge between parameters and performance, but also lays the foundation for the subsequent construction of bonding index maps and the conduct of macro-coating quality analysis. The significance of the actual bonding coefficient is that it maps the ideal degree of fit of the micro-spraying conditions to a quantifiable feature of each pixel point. It is a core indicator reflecting whether the bonding state is stable, uniform, and up to standard, and provides a basic input for the construction of a two-dimensional actual bonding index set.

[0068] S16. Construct a two-dimensional actual bonding index set of the target building material surface based on the actual bonding coefficient of each pixel point and in combination with the coordinate position of each pixel point.

[0069] In this embodiment, an innovative path for achieving refined parameter mapping and theoretical performance modeling for the spraying construction process is proposed. By constructing a two-dimensional pixel coordinate system strictly aligned with the spraying area, full-surface resolution with spatial consistency is achieved; and a collaborative acquisition mechanism of pressure sensors and laser trajectory recorders is introduced to ensure high spatiotemporal matching accuracy between the spraying pressure data and the actual path of the spray gun, and then through interpolation projection, each surface pixel point has an actual pressure value expression; in addition, a non-contact two-dimensional temperature distribution map is established by deploying infrared thermal imagers, and accurately integrated with the pixel coordinate system with the help of image registration and geometric correction algorithms to form a pixel-level expression of the real temperature parameters; combined with the standard spraying database based on material The standard temperature and pressure thresholds for the type and coating specifications are retrieved, which further realize the pixel-by-pixel calculation of the actual bonding coefficient, forming a two-dimensional actual bonding index set that characterizes the spraying quality level; the effect of this process is that it not only breaks through the limitations of traditional manual sampling methods, but also significantly improves the coverage and granularity of data collection; it also constructs a precise bridge between the physical conditions, bonding performance and pixel response in the spraying process through a high-resolution parameter grid with consistent spatial distribution; this method is highly scalable and adaptable to different scenarios, and can achieve real-time construction quality prediction, spray parameter deviation warning and early identification of regional defects in actual engineering efficiency, significantly improving the intelligent management and quality assurance capabilities of the building spraying system.

[0070] Example 3, please refer to Figure 1 and Figure 4 , specifically: S2 specific steps include:

[0071] S21. By extracting features from the two-dimensional actual bonding index set and combining it with a statistical averaging algorithm, the overall bonding effect of the coating on the surface of the target building material is analyzed to obtain the overall bonding coefficient of the coating on the surface of the target building material;

[0072] The overall coating adhesion coefficient is a numerical indicator representing the average level of spray coating adhesion quality across the entire surface, calculated by statistically extracting the actual adhesion coefficients of all pixels on the target building material surface. It reflects the consistency of spray coating construction and the overall film-forming effect on the target building material surface. It serves as the core criterion for spray coating quality assessment, providing a unified value for macroscopically evaluating the adhesion status of the entire coating, facilitating a quick assessment of whether the spray coating process meets standards.

[0073] S22. Based on the value of the overall adhesion coefficient of the coating on the surface of the target building material, determine whether the overall adhesion level of the coating in the current spraying process is qualified, and issue corresponding surface quality analysis instructions, specifically:

[0074] When the overall adhesion coefficient of the paint on the surface of the target building material is equal to one, it means that the overall adhesion level of the paint in the current spraying process is qualified, and the current target building material is identified as a qualified adhesion material, and a surface quality analysis instruction is issued; otherwise, it means that the overall adhesion level of the paint in the current spraying process is unqualified, and the current target building material is identified as a surface quality defective material.

[0075] In this embodiment, through the bonding quality analysis process described in S21 and S22 above, the core quantitative indicator of the overall bonding coefficient of the coating is introduced into the surface quality identification process of building materials, thereby achieving quantitative improvement and logical closed-loop construction of the spraying construction effect from the pixel level to the overall surface level; the introduction of this indicator effectively solves the problem of the lack of global consistency criterion for bonding quality assessment in the existing technology; it is based on the statistical summary of the actual bonding coefficient of each pixel point, which not only retains the perception ability of subtle bonding differences, but also characterizes the overall homogeneity and consistency of the spraying results in a unified numerical method, significantly enhancing the clarity and execution efficiency of the judgment standard; especially in engineering applications, by setting the qualified threshold standard of the overall bonding coefficient of the coating to 1, it is possible to quickly screen out whether there are spraying areas with overall bonding abnormalities in a non-contact, image-driven manner, reducing the need for manual judgment and destructive testing; at the same time, this mechanism supports the automatic issuance of surface quality analysis instructions, providing a trigger signal for subsequent more refined surface reflection deviation analysis and quality grade classification, thereby improving the responsiveness and linkage intelligence level of the recognition system.

[0076] Example 4, please refer to Figure 1 , specifically: S3 specific steps include:

[0077] S31. After receiving the surface quality analysis instruction, the qualified bonding material surface is illuminated at multiple angles to obtain a multi-angle polarized reflection frame grayscale image to form a multi-angle reflection vector, specifically:

[0078] After receiving the surface quality analysis instruction, the reflective light source emitted by the electrically controlled polarized LED array with polarization angle adjustment function is used to illuminate the surface of the qualified adhesive material at polarization angles of 30°, 60°, and 90°. A high-sensitivity CMOS image sensor is used with a polarization filter group to synchronously capture the reflected frame images at each polarization angle. After grayscale processing, the reflected frame grayscale images at each polarization angle are obtained.

[0079] A reflection frame grayscale image is a two-dimensional image formed by the intensity variation of reflected light at each pixel on the surface of a qualified adhesive material when irradiated at a set polarization angle. The pixel grayscale value reflects the local reflective properties of the material at that polarization angle. This type of image is used to capture the differences in the microscopic optical response of the material surface under different lighting angles, thereby identifying the uniformity of the spray coating and local abnormal reflective behavior, providing a data foundation for the subsequent construction of multi-angle reflection vectors and reflection tensors.

[0080] After spatially calibrating the grayscale images of the reflection frames at each polarization angle using an image calibration plate, the grayscale images of the reflection frames at each polarization angle are perspective-corrected using a homography matrix model. The corrected grayscale images of the reflection frames at each polarization angle are then registered into a two-dimensional pixel coordinate system.

[0081] Perspective transformation correction is a process that uses the principles of projective geometry to convert image projection distortion caused by differences in camera viewing angle or position back to a standard view that is consistent with the true geometric shape of the measured plane. Its function is to uniformly map the reflection frame grayscale images taken at different polarization angles to the same two-dimensional pixel coordinate system, ensuring that the pixel coordinates of the same physical location in multiple images are completely corresponding, facilitating subsequent multi-angle data fusion and pixel-level comparison. The correction method usually first uses an image calibration plate to obtain the corresponding feature point pairs between each image and the reference plane, and then uses the least squares method and the homography matrix model to perform a perspective transformation. Each pixel coordinate of the reflection frame grayscale image is corrected and transformed. Finally, the transformed data is resampled to a uniform resolution to complete the perspective transformation correction and achieve accurate registration of multi-angle images.

[0082] Based on the resolution of the image acquisition device, the grayscale images of the reflected frames at each polarization angle after registration are divided according to the imaging resolution, and the grayscale value of each pixel is converted into the reflection intensity value of the corresponding pixel through the image grayscale conversion model. The multi-angle reflection vector of the qualified bonding material surface is constructed in the form of: Rθ(x,y)=[R30(x,y),R60(x,y),R90(x,y)]; where Rθ(x,y) represents the multi-angle reflection vector, θ=30°, 60°, 90°, R30(x,y) represents the reflection intensity value of the pixel point (x,y) at a polarization angle of 30°, R60(x,y) represents the reflection intensity value of the pixel point (x,y) at a polarization angle of 60°, and R90(x,y) represents the reflection intensity value of the pixel point (x,y) at a perpendicular polarization angle.

[0083] A multi-angle reflection vector is a feature vector constructed based on the difference in pixel reflection intensity at different polarization angles. It is used to describe the reflection behavior of a qualified adhesive surface under multi-angle illumination conditions. This vector uses an image grayscale conversion model to map grayscale images collected at 30°, 60°, and 90° polarization angles into reflection intensity values ​​for the corresponding pixels, thereby forming a local optical response description unit that reflects the spatial directional reflection characteristics. This structure is closely related to the problem pointed out in the background technology that existing methods lack systematic analysis and quantitative interpretation of multi-angle polarized reflection images. Due to the microstructural differences of the coating surface, the scattering and reflection characteristics of light vary at different angles. Relying solely on single-angle reflection images is prone to misjudgment, especially on complex surfaces such as smooth or porous surfaces. The construction of this vector not only enables multi-angle perception of reflection characteristics but also provides the core input basis for subsequent steps such as tensor model construction, identification of reflection deviation areas, and flatness quantification. This reflection vector can capture the anisotropic characteristics of the material surface, enhancing the sensitivity of quality recognition methods to local structural changes, thereby significantly improving the robustness and accuracy of surface recognition.

[0084] Specifically, S32, constructing a third-order three-channel surface tensor of the qualified bonding material surface according to the multi-angle reflection vectors of the qualified bonding material surface, specifically:

[0085] The multi-angle reflection vectors of the qualified bonding material surface are normalized and combined with the sliding window convolution sampling mechanism to construct the third-order three-channel surface tensor of the qualified bonding material surface. The specific expression is:

[0086] ;

[0087] Where M(x,y,k) represents the third-order three-channel surface tensor, k represents the index number of the corresponding reflection channel, and k=1, 2, and 3 represent the channel index numbers corresponding to R30(x,y), R60(x,y), and R90(x,y), respectively.

[0088] The third-order three-channel surface tensor is a spatial angle joint feature volume constructed by integrating multi-angle reflection vectors according to the reflection channel number, where k=1, 2, and 3 correspond to the reflection intensity channels of 30°, 60°, and 90°, respectively. Through normalization processing and sliding window convolution sampling mechanism, the tensor expands the multi-angle reflection data in the two-dimensional spatial coordinate system into a data volume with a three-dimensional structure, which has joint information of spatial distribution and angular response. This construction is a direct response to the problem that the background technology lacks quantitative analysis of the image reflection change trend under different polarization angles. The pixel-level reflection response expression is realized in the form of a tensor, which effectively preserves the spatial continuity and angular feature consistency of the reflection data. The third-order three-channel surface tensor not only provides structured data support for subsequent reflection deviation analysis, but also has good compatibility and scalability, and can support advanced processing tasks such as reflection difference classification, texture recognition or flatness analysis based on deep learning.

[0089] Specifically, S33, based on constructing a third-order three-channel surface tensor of the qualified bonding material surface, sum the reflection intensity values ​​of all pixel points in the corresponding reflection channel in the third-order three-channel surface tensor to obtain the total reflection intensity value of each reflection channel, and combine the statistical averaging algorithm to obtain the global average reflection intensity value of each reflection channel.

[0090] The global average reflection intensity value refers to the arithmetic mean value obtained by statistically analyzing the reflection intensity values ​​of all pixels in each reflection channel in the third-order three-channel surface tensor, reflecting the overall reflection brightness level of the entire material surface under the reflection channel. The role of this average value is to provide a unified reference benchmark for the subsequent calculation of the deviation degree of each pixel point from the overall reflection state, and to construct the reflection deviation coefficient to identify local abnormalities or inhomogeneous reflection areas on the surface. The acquisition method is: first, all pixel reflection intensity values ​​in each reflection channel are extracted from the constructed third-order three-channel tensor, and their sum is statistically calculated; using the averaging algorithm in statistics, that is, dividing the sum of all pixel reflection values ​​in each reflection channel by the total number of pixels in the reflection channel, the global average reflection intensity value of each reflection channel can be obtained.

[0091] Specifically, S34, the global average reflection intensity value of each reflection channel is correlated with the third-order three-channel surface tensor, and the overall deviation of the reflection intensity of each pixel point on the surface of the qualified bonding material at different polarization angles is analyzed to determine the degree of reflection deviation of each pixel point on the surface of the qualified bonding material, specifically: Where Py(x,y) represents the reflection deviation coefficient of the corresponding pixel point, M(x,y,k) represents the third-order three-channel surface tensor of the qualified bonding material surface, Ravg k Represents the global average reflection intensity value of the kth reflection channel.

[0092] The formula in S34 is used to calculate the reflection deviation coefficient of the pixel point. By taking the absolute value of the difference between the reflection intensity value under each channel of the third-order three-channel surface tensor and the global average reflection intensity of the corresponding channel and summing them, the overall deviation of the reflection intensity of the pixel at different polarization angles is measured. This calculation process is designed to address the problem that the background technology lacks a quantitative analysis of the image reflection change trend under different polarization angles. The formula realizes the quantitative characterization of the polarization response at the pixel level. The significance of the reflection deviation coefficient is to reveal the degree of abnormality of each pixel point under the standard polarization response distribution. The larger the reflection deviation coefficient, the more obvious the deviation of the reflection characteristics of the area at different angles from the global trend, indicating the presence of surface fluctuations, defects or structural abnormalities. This coefficient can not only be used for subsequent abnormal pixel identification and quality grading judgment, but also provides a high-dimensional and quantifiable representation basis for constructing a more robust building surface state discrimination model. It is a key bridge to realize the transformation from image brightness information to material structure state perception.

[0093] In this embodiment, by constructing multi-angle reflection vectors and third-order three-channel surface tensors, a deep and multi-dimensional analysis of the surface reflection response characteristics of qualified adhesive materials is achieved, which has intelligent recognition value and practical engineering applicability. After receiving the surface quality analysis instruction, the target surface is first illuminated at three polarization angles of 30°, 60°, and 90° using an electrically controlled LED array with polarization adjustment function, and the reflection frame image is synchronously collected by a high-sensitivity CMOS image sensor to form multi-angle grayscale information. After image calibration and perspective transformation registration, the images of each angle are uniformly mapped to a two-dimensional pixel coordinate system to ensure consistent alignment and correspondence of data under different channels. The image grayscale conversion model is used to convert the pixel grayscale value into a reflection intensity value with a clear physical meaning, and then a multi-angle reflection vector reflecting the reflection characteristics of the local material is constructed. The vector is further processed by normalization and sliding window convolution mechanism to form a third-order three-channel tensor with context-awareness, which not only captures the local texture in different The changing law of polarization response retains the integrity of its spatial distribution; by counting the total reflection intensity of each channel and calculating its global average value, it is further compared with the intensity value of the channel corresponding to each pixel in the tensor, and the degree of reflection deviation of each pixel in the current channel is quantified, thereby forming the reflection deviation coefficient of each pixel. The advantage is that it can not only realize the accurate judgment of surface homogeneity and micro-anomalies, but also can be used to check whether the spraying is flat even when the overall bonding is qualified; especially in dealing with multi-light reflection interference scenarios, the tensor-level multi-angle polarization response analysis method provided can effectively hedge the misjudgment caused by the difference in incident angle and surface micro-warping in traditional single-frame image analysis, and significantly improve the stability and adaptability of the judgment; therefore, this step not only provides a supplementary indicator of the reflection physical level for surface quality identification, but also lays the foundation for high-precision data for subsequent accurate classification and finished product screening. It is an indispensable core mechanism for realizing the quality identification of intelligent building materials.

[0094] Example 5, please refer to Figure 1 and Figure 4 , specifically: S4 specific steps include:

[0095] S41, comparing and analyzing the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with a preset reflection deviation threshold value to determine whether the surface of the qualified bonding material is in a flat state, specifically:

[0096] If the reflection deviation coefficient of the corresponding pixel exceeds the reflection deviation threshold, it means that the reflection deviation degree of the corresponding pixel is abnormal, and the corresponding pixel is divided into an abnormal deviation pixel set; otherwise, the corresponding pixel is divided into a normal deviation pixel set;

[0097] When the number of pixels in the abnormal deviation pixel set exceeds 5% of the total number of pixels on the surface of the qualified bonding material, it means that the surface of the qualified bonding material is not in a flat state, and the qualified bonding material is identified as a surface quality defective material;

[0098] When the number of pixels in the abnormal deviation pixel point set does not exceed 5% of the total number of pixels in the surface of the qualified bonding material, it means that the surface of the qualified bonding material is in a flat state, and the qualified bonding material is identified as a surface quality finished material.

[0099] The reflection deviation coefficient quantifies the difference between the reflection intensity of a single pixel in a certain polarization angle channel and the global average reflection intensity of that channel. It essentially reflects the degree of optical anomaly in the thickness, roughness, or local warping of this tiny surface area relative to the overall coating. When the spray coating is flat, uniform in thickness, and has no surface bumps, the reflection intensity of each pixel should maintain a small difference from the global mean, so its deviation coefficient is usually below the preset threshold. Once local ups and downs appear, or the spray coating is too thick or thin, the reflection direction and intensity will change significantly, causing the deviation coefficient of the pixels in this area to increase significantly. Once it exceeds the threshold, the pixels are marked as abnormally deviated pixels. The overall flatness can be determined by counting the proportion of abnormally deviated pixels in the entire image.

[0100] In this embodiment, in S4, the point-by-point comparison of the pixel-level reflection deviation coefficient and the preset threshold is used as the core criterion. By establishing an automatic classification mechanism for abnormal deviation pixel point sets and normal deviation pixel point sets, the judgment of whether the surface is flat is converted into an accurate, objective and quantifiable data decision-making process; compared with the traditional reliance on manual experience or macroscopic visual inspection, the outstanding advantages of this mechanism are: the fine-grained optical indicators obtained by polarization multi-channel tensor analysis are implemented as intuitive threshold judgments, which reduces the complexity of algorithm configuration and can be used by on-site construction personnel without deep image processing knowledge; secondly, the global proportion threshold of 5% abnormality is introduced to make local small surfaces flat. Accumulated defects will not cause overall rework, avoiding excessive intervention. At the same time, it can trigger defective product identification in time when the defect ratio reaches the lower risk limit, achieving a balance between quality and efficiency. The pixel ratio method has scale-adaptive characteristics. Regardless of whether the wall area is tens of square meters or hundreds of square meters, the judgment threshold will be dynamically adjusted based on the total number of pixels to ensure consistent recognition standards in large and small scenes. In summary, the recognition process not only completes the transformation from optical deviation to quality label, but also brings a lightweight, robust and easy-to-promote surface flatness intelligent judgment solution through the proportional threshold strategy, providing more reliable and more engineering-effective technical support for the quality identification of finished building coatings.

[0101] Example 6, please refer to Figure 1 and Figure 2,Specifically: Building material surface quality identification system, including bonding analysis module, ,effect determination module, reflection deviation analysis module and quality ,identification module;

[0102] The bonding analysis module is used to establish a two-dimensional pixel coordinate system based on the standard viewing angle imaging of the target building material surface by the image acquisition device, record the surface temperature and spraying pressure information of the target building material surface during the spraying process, and construct a two-dimensional actual bonding index set of the target building material surface;

[0103] The effect determination module is used to extract features from the two-dimensional actual bonding index set and, combined with a statistical averaging algorithm, analyze the overall bonding effect of the coating on the surface of the target building material to identify the current target building material as a qualified bonding material and issue a surface quality analysis instruction;

[0104] The reflection deviation analysis module is used to, after receiving the surface quality analysis instruction, illuminate the qualified bonding material surface from multiple angles and, in combination with the sliding window convolution sampling mechanism, analyze the overall deviation of the reflection intensity of each pixel point on the qualified bonding material surface at different polarization angles to determine the degree of reflection deviation of each pixel point on the qualified bonding material surface;

[0105] The quality identification module is used to compare and analyze the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with the preset reflection deviation threshold value, and identify the current qualified bonding material as a surface quality finished material.

[0106] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for identifying the surface quality of building materials, characterized by: The following steps are involved: S1. Using an image acquisition device to perform standard viewing angle imaging of the target building material surface, establish a two-dimensional pixel coordinate system, record the surface temperature and spraying pressure information of the target building material surface during the spraying process, and construct a two-dimensional actual bonding index set for the target building material surface; S2. Extract features from the two-dimensional actual bonding index set and, combined with a statistical averaging algorithm, analyze the overall bonding effect of the coating on the surface of the target building material to identify the current target building material as a qualified bonding material and issue a surface quality analysis instruction; S3. After receiving the surface quality analysis instruction, the qualified bonding material surface is illuminated at multiple angles. In combination with the sliding window convolution sampling mechanism, the overall deviation of the reflection intensity of each pixel point on the qualified bonding material surface at different polarization angles is analyzed to determine the degree of reflection deviation of each pixel point on the qualified bonding material surface. The specific steps of S3 include: S31. After receiving the surface quality analysis instruction, the qualified bonding material surface is illuminated at multiple angles to obtain a multi-angle polarized reflection frame grayscale image to form a multi-angle reflection vector, specifically: After receiving the surface quality analysis instruction, the reflective light source emitted by the electrically controlled polarized LED array with polarization angle adjustment function is used to illuminate the surface of the qualified adhesive material at polarization angles of 30°, 60°, and 90°. A high-sensitivity CMOS image sensor is used with a polarization filter group to synchronously capture the reflected frame images at each polarization angle. After grayscale processing, the reflected frame grayscale images at each polarization angle are obtained. After spatially calibrating the grayscale images of the reflection frames at each polarization angle using an image calibration plate, the grayscale images of the reflection frames at each polarization angle are perspective-corrected using a homography matrix model. The corrected grayscale images of the reflection frames at each polarization angle are then registered into a two-dimensional pixel coordinate system. Based on the resolution of the image acquisition device, the grayscale images of the reflected frames at each polarization angle after registration are divided according to the imaging resolution, and the grayscale value of each pixel is converted into the reflection intensity value of the corresponding pixel through the image grayscale conversion model to construct the multi-angle reflection vector of the qualified bonding material surface, which is in the form of: Rθ(x,y)=[R30(x,y),R60(x,y),R90(x,y)]; where Rθ(x,y) represents the multi-angle reflection vector, θ=30°, 60°, 90°, R30(x,y) represents the reflection intensity value of the pixel point (x,y) at a polarization angle of 30°, R60(x,y) represents the reflection intensity value of the pixel point (x,y) at a polarization angle of 60°, and R90(x,y) represents the reflection intensity value of the pixel point (x,y) at a perpendicular polarization angle; S32. Construct a third-order three-channel surface tensor of the qualified bonding material surface based on the multi-angle reflection vectors of the qualified bonding material surface, specifically: The multi-angle reflection vectors of the qualified bonding material surface are normalized and combined with the sliding window convolution sampling mechanism to construct the third-order three-channel surface tensor of the qualified bonding material surface. The specific expression is: ； Where M(x,y,k) represents the third-order three-channel surface tensor, k represents the corresponding reflection channel index number, and k=1, 2, and 3 represent the channel index numbers corresponding to R30(x,y), R60(x,y), and R90(x,y), respectively. S33. Based on the third-order three-channel surface tensor constructed for the qualified bonding material surface, summing the reflection intensity values ​​of all pixels in the corresponding reflection channels in the third-order three-channel surface tensor to obtain a total reflection intensity value of each reflection channel, and using a statistical averaging algorithm to obtain a global average reflection intensity value of each reflection channel; S34. Correlate the global average reflection intensity value of each reflection channel with the third-order three-channel surface tensor, and analyze the overall deviation of the reflection intensity of each pixel point on the surface of the qualified bonding material at different polarization angles to determine the degree of reflection deviation of each pixel point on the surface of the qualified bonding material, specifically: Where Py(x,y) represents the reflection deviation coefficient of the corresponding pixel point, M(x,y,k) represents the third-order three-channel surface tensor of the qualified bonding material surface, Ravg k Represents the global average reflection intensity value of the kth reflection channel; S4. Compare and analyze the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with a preset reflection deviation threshold value, and identify the current qualified bonding material as a surface quality finished material.

2. The method for identifying the surface quality of building materials according to claim 1, wherein: The specific steps of S1 include: S11, using the actual sprayed surface of the building material to be sprayed as the target building material surface, imaging the target building material surface from a standard viewing angle using an image acquisition device, and establishing a two-dimensional pixel coordinate system for the target building material surface after geometric benchmark calibration and correction; S12. Using a digital pressure sensor built into the spraying equipment to record the output pressure value during the spraying process in real time, with a set time length t as the sampling period, to generate pressure data with a timestamp, using a laser trajectory recorder to collect the spray gun position information of each pixel point during the spraying process, synchronously matching the timestamp of the pressure data with the spray gun position information, projecting the matched pressure data into a two-dimensional pixel coordinate system using an interpolation algorithm, and extracting the actual spray pressure value of each pixel point; S13. Use the deployed infrared thermal imager to perform real-time temperature measurement and scanning on the surface of the target building material, collect the thermal radiation intensity of the target building material surface in a non-contact manner, and output a two-dimensional temperature distribution map with a resolution consistent with the two-dimensional pixel coordinate system. Combined with the image registration mechanism, the two-dimensional temperature distribution map and the two-dimensional pixel coordinate system are geometrically corrected to extract the actual surface temperature value of each pixel point.

3. The method for identifying the surface quality of building materials according to claim 2, wherein: S14. Conditionally index the experimental database according to the material type and coating specifications to obtain standard spraying pressure values ​​and standard surface temperature values; S15. Based on the actual spraying pressure value and the actual surface temperature value of each pixel point extracted, the actual bonding coefficient of each pixel point is determined, specifically: Where Bnj(x,y) represents the actual bonding coefficient of the corresponding pixel point, T0 represents the standard surface temperature value, Ts(x,y) represents the actual surface temperature value of the corresponding pixel point, P0 represents the standard spraying pressure value, Ps(x,y) represents the actual spraying pressure value of the corresponding pixel point, and (x,y) represents the coordinate position of the corresponding pixel point; S16. Construct a two-dimensional actual bonding index set of the target building material surface based on the actual bonding coefficient of each pixel point and in combination with the coordinate position of each pixel point.

4. The method for identifying the surface quality of building materials according to claim 3, wherein: The specific steps of S2 include: S21. By extracting features from the two-dimensional actual bonding index set and combining it with a statistical averaging algorithm, the overall bonding effect of the coating on the surface of the target building material is analyzed to obtain the overall bonding coefficient of the coating on the surface of the target building material; S22. Based on the value of the overall adhesion coefficient of the coating on the surface of the target building material, determine whether the overall adhesion level of the coating in the current spraying process is qualified, and issue corresponding surface quality analysis instructions, specifically: When the overall adhesion coefficient of the paint on the surface of the target building material is equal to one, it means that the overall adhesion level of the paint in the current spraying process is qualified, and the current target building material is identified as a qualified adhesion material, and a surface quality analysis instruction is issued; otherwise, it means that the overall adhesion level of the paint in the current spraying process is unqualified, and the current target building material is identified as a surface quality defective material.

5. The method for identifying the surface quality of building materials according to claim 1, wherein: The specific steps of S4 include: S41, comparing and analyzing the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with a preset reflection deviation threshold value to determine whether the surface of the qualified bonding material is in a flat state, specifically: If the reflection deviation coefficient of the corresponding pixel exceeds the reflection deviation threshold, it means that the reflection deviation degree of the corresponding pixel is abnormal, and the corresponding pixel is divided into an abnormal deviation pixel set; otherwise, the corresponding pixel is divided into a normal deviation pixel set; When the number of pixels in the abnormal deviation pixel set exceeds 5% of the total number of pixels on the surface of the qualified bonding material, it means that the surface of the qualified bonding material is not in a flat state, and the qualified bonding material is identified as a surface quality defective material; When the number of pixels in the abnormal deviation pixel point set does not exceed 5% of the total number of pixels in the surface of the qualified bonding material, it means that the surface of the qualified bonding material is in a flat state, and the qualified bonding material is identified as a surface quality finished material.

6. A building material surface quality identification system, for implementing the building material surface quality identification method according to any one of claims 1 to 5, characterized in that: It includes adhesion analysis module, effect determination module, reflection deviation analysis module and quality identification module; The bonding analysis module is used to establish a two-dimensional pixel coordinate system based on the standard viewing angle imaging of the target building material surface by the image acquisition device, record the surface temperature and spraying pressure information of the target building material surface during the spraying process, and construct a two-dimensional actual bonding index set of the target building material surface; The effect determination module is used to extract features from the two-dimensional actual bonding index set and, combined with a statistical averaging algorithm, analyze the overall bonding effect of the coating on the surface of the target building material to identify the current target building material as a qualified bonding material and issue a surface quality analysis instruction; The reflection deviation analysis module is used to, after receiving the surface quality analysis instruction, illuminate the qualified bonding material surface from multiple angles and, in combination with the sliding window convolution sampling mechanism, analyze the overall deviation of the reflection intensity of each pixel point on the qualified bonding material surface at different polarization angles to determine the degree of reflection deviation of each pixel point on the qualified bonding material surface; The quality identification module is used to compare and analyze the reflection deviation coefficient of each pixel point on the surface of the qualified bonding material with the preset reflection deviation threshold value, and identify the current qualified bonding material as a surface quality finished material.

Images