Visual inspection method for building external wall thermal insulation material

Abstract

A visual detection method for a building external wall thermal insulation material relates to the technical field of visual detection, and comprises the following steps: acquiring material attributes and external environment data of a to-be-detected building external wall material, judging whether to perform visual detection operation on the building external wall material, and acquiring shooting parameters of the to-be-detected building external wall material when the external environment data accords with an operation standard; whether infrared image detection operation can be performed on the external environment data of the to-be-detected building external wall material is judged, visual image detection operation is performed on the thermal insulation performance undetermined area, color abnormal pixel points and a texture analysis angle are obtained, local area texture analysis is performed on the thermal insulation performance undetermined area, and a texture abnormal area is obtained; according to the method, the Canny algorithm is used for carrying out edge detection to obtain the damaged abnormal area, and meanwhile, complexity level change trend prediction is carried out on the local area which is not marked as the damaged abnormal area to obtain the potential damaged area, so that the accuracy and reliability of detection are improved.

Description

Technical Field

[0001] The invention relates to the technical field of visual detection, in particular to a visual detection method for building exterior wall thermal insulation materials. Background Art

[0002] The prior art CN114882038A "A method and device for detecting building exterior wall insulation materials" obtains a surface grayscale image of the expanded perlite, divides it into multiple window areas of the same size, obtains an energy eigenvalue according to the grayscale co-occurrence matrix of each window area, obtains a grayscale pair of the grayscale level of the pixel with the largest autocorrelation in each window area as a characteristic grayscale pair, obtains the complexity of each window area according to the frequency of the characteristic grayscale pair and the grayscale difference of each grayscale pair, calculates the complexity correlation between each window area and other window areas according to the complexity of each window area, obtains a quality evaluation value of the expanded perlite according to the complexity of the window area with the largest complexity correlation and the energy eigenvalue of the grayscale co-occurrence matrix of each window area, and uses the quality evaluation value to detect the quality of the expanded perlite.

[0003] There are some problems with the existing visual inspection methods for building exterior wall insulation materials. For example, changes in lighting conditions during the visual inspection of building exterior wall insulation materials will cause changes in color and brightness, which may affect the effects of color detection and edge detection. Building exterior walls may use multiple types of insulation materials, which may have different colors and textures, which increases the difficulty of detection. High-level visual inspection algorithms may require a lot of computing resources, which may limit their practicality. These problems may affect the accuracy and efficiency of detection. Summary of the invention

[0004] In order to solve the above technical problems, the object of the present invention is to provide a visual inspection method for building exterior wall thermal insulation materials, comprising the following steps:

[0005] Step s1: obtaining material properties and external environment data of the building exterior wall material to be inspected, determining whether to perform a visual inspection operation of the building exterior wall material according to the external environment data, and obtaining shooting parameters of the building exterior wall material to be inspected when the external environment data meets the operation standard;

[0006] Step s2: determine whether the external environment data of the building exterior wall material to be detected can be subjected to infrared image detection operation, and obtain the thermal insulation performance unqualified area and the thermal insulation performance pending area according to the determination result;

[0007] Step s3: Perform visual image detection on the area whose thermal insulation performance is to be determined, obtain color abnormal pixel points and texture analysis angles, perform local area texture analysis on the area whose thermal insulation performance is to be determined according to the texture analysis angles, and obtain texture abnormal areas;

[0008] Step s4: Use the Canny algorithm to perform edge detection on each local area of ​​the area where the thermal insulation performance is to be determined, and obtain the damaged abnormal area. At the same time, predict the complexity level change trend of the local area that is not marked as the damaged abnormal area to obtain the potential damaged area.

[0009] Furthermore, the process of obtaining material properties and external environment data of the building exterior wall material to be inspected, and determining whether to perform a visual inspection operation of the building exterior wall material according to the external environment data includes:

[0010] Obtain material properties and external environment data of the building exterior wall material to be detected, wherein the material properties include color characteristics, reflective characteristics and texture structure, and the external environment data include indoor building temperature, outdoor building temperature, light intensity, wind speed and precipitation intensity, obtain the building outdoor temperature, wind speed and precipitation intensity corresponding to the building exterior wall material to be detected, and compare the building outdoor temperature, wind speed and precipitation intensity corresponding to the building exterior wall material to be detected with the preset temperature limit threshold, the preset wind speed limit threshold and the preset precipitation intensity limit threshold respectively; if the building outdoor temperature, wind speed and precipitation intensity corresponding to the building exterior wall material to be detected are all less than the corresponding temperature limit threshold, wind speed limit threshold and precipitation intensity limit threshold, then perform a visual inspection operation on the building exterior wall material to be detected; if the building outdoor temperature, wind speed or precipitation intensity corresponding to the building exterior wall material to be detected is greater than the corresponding temperature limit threshold, wind speed limit threshold and precipitation intensity limit threshold, then suspend the visual inspection operation on the building exterior wall material to be detected.

[0011] Furthermore, the process of obtaining the shooting parameters of the building exterior wall material to be inspected includes:

[0012] According to the wind speed and precipitation intensity corresponding to the building exterior wall materials to be tested as evaluation indicators, the indicator weights of the evaluation indicators are set, and the membership matrix of the building exterior wall materials to be tested for the preset interference level is obtained through fuzzy comprehensive evaluation;

[0013] The interference level of the building exterior wall material to be detected is obtained through the membership matrix and the indicator weight, and an adaptive shooting parameter comparison table is pre-constructed. The adaptive shooting parameter comparison table includes shooting parameters corresponding to different interference levels. The shooting parameters include resolution and frame number. The shooting parameters of the building exterior wall material to be detected are determined according to the interference level of the building exterior wall material to be detected.

[0014] Furthermore, the process of judging whether the external environment data of the building exterior wall material to be detected can be subjected to infrared image detection operation, and obtaining the thermal insulation performance unqualified area and the thermal insulation performance undetermined area according to the judgment result includes:

[0015] The temperature difference between the indoor temperature of the building and the outdoor temperature of the building is obtained according to the external environment data of the building exterior wall material to be detected, the temperature difference is marked as a first temperature difference, the first temperature difference is compared with a preset temperature difference threshold, when the temperature difference is greater than or equal to the temperature difference threshold, a threshold point is selected within the first temperature difference to divide the temperature difference sub-intervals into different thermal insulation performance levels, the emissivity of the drone camera terminal is set according to the material properties of the building exterior wall material to be detected, the drone camera terminal is used to perform infrared scanning on the building exterior wall material to be detected, an infrared image is obtained, and the temperature corresponding to the grayscale value of each pixel in the infrared image is obtained based on a preset temperature lookup table of the drone camera terminal. degree, obtaining the temperature difference between the temperature corresponding to each pixel and the outdoor temperature of the building, marking the temperature difference as a second temperature difference, matching the second temperature difference with the temperature difference sub-intervals of different thermal insulation performance levels, obtaining the thermal insulation performance level corresponding to each pixel, comparing the thermal insulation performance level corresponding to each pixel with a preset thermal insulation performance level threshold, marking the distribution area of ​​the pixels whose thermal insulation performance level is less than the thermal insulation performance level threshold as an unqualified thermal insulation performance area, marking the distribution area of ​​the pixels whose thermal insulation performance level is greater than or equal to the thermal insulation performance level threshold as an undetermined thermal insulation performance area, and performing a visual image detection operation on the undetermined thermal insulation performance area;

[0016] When the temperature difference is less than the temperature difference threshold, all areas covered by the exterior wall materials of the building to be inspected are marked as areas whose thermal insulation performance is pending, and visual image inspection operations are performed on the areas whose thermal insulation performance is pending.

[0017] Furthermore, the process of performing visual image detection operation on the area whose thermal insulation performance is to be determined and obtaining color abnormal pixel points and texture analysis angles includes:

[0018] A drone camera terminal is used to perform multi-angle image scanning on the area of ​​the building exterior wall material to be tested where the thermal insulation performance is to be determined, to obtain a multi-angle surface color image of the area where the thermal insulation performance is to be determined, and the multi-angle surface color image is converted from a preset BGR color space to an HSV color space, and then the color features of each pixel in the multi-angle surface color image are extracted, the color features including hue, saturation and brightness value, and the hue and saturation of each pixel are respectively compared with the hue and saturation corresponding to the material attribute to obtain the color difference value of each pixel, and the pixel whose color difference value is greater than the preset color difference value threshold is marked as a color abnormal pixel;

[0019] According to the corresponding brightness value of each pixel point in the multi-angle surface color image, the average brightness value and the standard deviation of the brightness value corresponding to different angles in the multi-angle surface color image are obtained, and the texture feature judgment value is obtained according to the brightness average value and the standard deviation of the brightness value, and the angle corresponding to the maximum texture feature judgment value at different angles is screened out as the texture analysis angle.

[0020] Furthermore, the process of performing local area texture analysis on the area whose thermal insulation performance is to be determined according to the texture analysis angle and obtaining the texture abnormal area includes:

[0021] The surface color image corresponding to the texture analysis angle in the multi-angle surface color image of the area with undetermined thermal insulation performance is denoised and grayscaled to obtain a grayscale image, and a region detection window and fixed parameters of a grayscale co-occurrence matrix are set. The grayscale co-occurrence matrix of each local region in the grayscale image is obtained by sliding the region detection window. The texture structure of each local region in the grayscale image is obtained according to the grayscale co-occurrence matrix, and the texture structure of each local region is compared with the corresponding texture structure in the material property for similarity to obtain the texture structure similarity of each local region, and the local region whose texture structure similarity is less than a preset similarity threshold is marked as a texture abnormality region.

[0022] Furthermore, the Canny algorithm is used to perform edge detection on each local area of ​​the area where the thermal insulation performance is to be determined, and the process of obtaining the damaged abnormal area includes:

[0023] The Canny algorithm is used to perform edge detection on each local area in the grayscale image to obtain the edge image of each local area, and the feature extraction is performed on the edge image of each local area to obtain the edge feature set of the edge image. The various types of edge features in the edge feature set include edge length, edge branch number, continuity measure, edge curvature, and edge roughness;

[0024] According to the various types of edge features in the edge feature set of the edge image as evaluation indicators, the indicator weight matrix of the evaluation indicators is set, and the membership matrix of the edge feature set for the preset complexity level is obtained through fuzzy comprehensive evaluation. The complexity level corresponding to the edge feature set is obtained according to the membership matrix and the indicator weight matrix, and the local area corresponding to the edge feature set with a complexity level greater than the preset complexity level threshold is marked as a damaged abnormal area. At the same time, the complexity level change trend of the local area that is not marked as a damaged abnormal area is predicted.

[0025] Furthermore, the complexity level change trend of the local area not marked as the damaged abnormal area is predicted, and the process of obtaining the potential damaged area includes:

[0026] Pre-construct a visual inspection database, the visual inspection database is used to store historical visual inspection records of a number of building exterior wall materials, obtain historical complexity level time series sequences of each local area not marked as damaged abnormal area according to the historical visual inspection records, construct a complexity level estimation model based on deep learning, use the historical complexity level time series sequences of each local area not marked as damaged abnormal area as a training set and a test set, input the training set into the complexity level estimation model for training until the loss function training is stable, save the model parameters, test the complexity level estimation model through the test set until it meets the preset requirements, and output the complexity level estimation model;

[0027] According to the complexity level estimation model, the estimated complexity level of each local area not marked as a damaged abnormal area is obtained, and the estimated complexity level is compared with the complexity level threshold. If the estimated complexity level of the local area not marked as a damaged abnormal area is greater than the complexity level threshold, the local area not marked as a damaged abnormal area is marked as a potential damaged area.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] The present invention provides a systematic method for detecting the thermal insulation performance and appearance quality of building exterior wall materials, especially for the detection of areas where the thermal insulation performance is to be determined. This method combines multiple technical means such as external environmental condition judgment, infrared image detection, visual image detection and edge detection, aiming to improve the accuracy and reliability of detection, wherein step s1: obtaining the material properties and external environmental data of the building exterior wall materials to be detected to ensure that the detection is carried out under a suitable external environment, avoiding misjudgment caused by environmental factors (such as light, temperature, etc.), reducing the detection error caused by changes in the external environment, and improving the reliability and accuracy of the detection result; step s2: judging whether the external environmental data can be subjected to infrared image detection operation, and using infrared images to detect areas with unqualified thermal insulation performance and areas to be determined. This is because in infrared images, the contrast between different temperature areas is high, which helps to distinguish areas with poor thermal insulation performance, and can effectively detect areas with poor thermal insulation performance, providing key areas for subsequent visual image detection. Domain, step s3: perform visual image detection operation on the area where the thermal insulation performance is to be determined, and identify the area where there may be problems through color abnormal pixels and texture analysis angles. The color abnormal pixels can be used to detect possible pollution, discoloration and other problems. Through the texture analysis angle, defects can be more clearly identified at a specific angle, thereby improving the pertinence of the detection. Step s4: Use the Canny algorithm to perform edge detection on each local area of ​​the area where the thermal insulation performance is to be determined. Use the Canny algorithm to detect edges to identify abnormal damaged areas and predict potential damaged areas. Predict the trend of complexity level changes in local areas that are not marked as abnormal damaged areas, which helps to discover possible problem areas in advance. By comprehensively considering external environmental conditions and multiple detection technologies, the accuracy and reliability of detection are improved. By pre-screening the area where the thermal insulation performance is to be determined, unnecessary detection workload is reduced and detection efficiency is improved. By predicting potential damaged areas, it helps to take measures in advance to avoid the deterioration of problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a schematic diagram of a visual inspection method for building exterior wall insulation materials according to an embodiment of the present application. DETAILED DESCRIPTION

[0031] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0032] like Figure 1 As shown, a visual inspection method for building exterior wall thermal insulation materials comprises the following steps:

[0033] Step s1: obtaining material properties and external environment data of the building exterior wall material to be inspected, determining whether to perform a visual inspection operation of the building exterior wall material according to the external environment data, and obtaining shooting parameters of the building exterior wall material to be inspected when the external environment data meets the operation standard;

[0034] Step s2: determine whether the external environment data of the building exterior wall material to be detected can be subjected to infrared image detection operation, and obtain the thermal insulation performance unqualified area and the thermal insulation performance pending area according to the determination result;

[0035] Step s3: Perform visual image detection on the area whose thermal insulation performance is to be determined, obtain color abnormal pixel points and texture analysis angles, perform local area texture analysis on the area whose thermal insulation performance is to be determined according to the texture analysis angles, and obtain texture abnormal areas;

[0036] Step s4: Use the Canny algorithm to perform edge detection on each local area of ​​the area where the thermal insulation performance is to be determined, and obtain the damaged abnormal area. At the same time, predict the complexity level change trend of the local area that is not marked as the damaged abnormal area to obtain the potential damaged area.

[0037] It should be further explained that, in the specific implementation process, the process of obtaining the material properties and external environment data of the building exterior wall material to be inspected, and determining whether to perform the visual inspection operation of the building exterior wall material according to the external environment data includes:

[0038] Obtain material properties and external environment data of the building exterior wall materials to be detected, wherein the material properties include color characteristics, reflective properties and texture structures, and the building exterior wall materials to be detected include metal plates, stones, tiles, polystyrene foam plastics, mineral wool boards, etc. Different types of exterior wall insulation materials will have different effects on visual detection due to different material properties. The external environment data include building indoor temperature, building outdoor temperature, light intensity, wind speed and precipitation intensity. Obtain the building outdoor temperature, wind speed and precipitation intensity corresponding to the building exterior wall materials to be detected, and compare the building outdoor temperature, wind speed and precipitation intensity corresponding to the building exterior wall materials to be detected with the preset temperature limit threshold, the preset wind speed limit threshold and the preset precipitation The visual inspection of the exterior wall materials to be inspected is performed by comparing the outdoor temperature, wind speed and precipitation intensity of the building to be inspected with the intensity limit thresholds. If the outdoor temperature, wind speed and precipitation intensity of the building to be inspected corresponding to the exterior wall materials to be inspected are all lower than the corresponding temperature limit thresholds, wind speed limit thresholds and precipitation intensity limit thresholds, the visual inspection of the exterior wall materials to be inspected is performed; if the outdoor temperature, wind speed or precipitation intensity of the building to be inspected corresponding to the exterior wall materials to be inspected are higher than the corresponding temperature limit thresholds, wind speed limit thresholds and precipitation intensity limit thresholds, the visual inspection of the exterior wall materials to be inspected is suspended. The inspection is performed under stable weather conditions and avoids testing under extreme temperatures or high wind speeds to reduce the impact of external factors on the visual inspection results, thereby ensuring the accuracy of the visual inspection results.

[0039] It should be further explained that, in the specific implementation process, the process of obtaining the shooting parameters of the building exterior wall material to be inspected includes:

[0040] According to the wind speed and precipitation intensity corresponding to the building exterior wall materials to be tested as evaluation indicators, the indicator weights of the evaluation indicators are set, and the membership matrix of the building exterior wall materials to be tested for the preset interference level is obtained through fuzzy comprehensive evaluation;

[0041] The interference level of the building exterior wall material to be detected is obtained through the membership matrix and the indicator weight, and an adaptive shooting parameter comparison table is pre-constructed. The adaptive shooting parameter comparison table includes shooting parameters corresponding to different interference levels. The shooting parameters include resolution and frame number. The shooting parameters of the building exterior wall material to be detected are determined according to the interference level of the building exterior wall material to be detected.

[0042] It should be further explained that, in the specific implementation process, the process of obtaining the interference level of the building exterior wall material to be tested according to the membership matrix and the index weight matrix includes:

[0043] The indicator weight matrix and the membership matrix of the evaluation indicators are integrated through a formula to obtain a fuzzy comprehensive evaluation matrix of the evaluation indicators, and the membership of the building exterior wall materials to be detected for different interference levels is obtained according to the fuzzy comprehensive evaluation matrix, and the interference level with the highest membership corresponding to the building exterior wall materials to be detected is screened out, and the interference level with the highest membership corresponding to the building exterior wall materials to be detected is used as the interference level of the building exterior wall materials to be detected;

[0044] Wherein, the formula is:

[0045] M = αM1 × βM2;

[0046] Among them, M is the fuzzy comprehensive evaluation matrix of the evaluation index, M1 is the indicator weight matrix of the evaluation index, M2 is the membership matrix, "×" represents the multiplication of the elements at the corresponding positions of the weight matrix of the evaluation index and the membership matrix, and α and β are weighting parameters for controlling the balance between the weight matrix and the membership matrix in the fuzzy comprehensive evaluation matrix of the evaluation index.

[0047] It should be further explained that, in the specific implementation process, the process of judging whether the external environment data of the building exterior wall material to be detected can be subjected to infrared image detection operation, and obtaining the thermal insulation performance unqualified area and the thermal insulation performance undetermined area according to the judgment result includes:

[0048] The temperature difference between the indoor temperature of the building and the outdoor temperature of the building is obtained according to the external environment data of the building exterior wall material to be detected, and the temperature difference is marked as a first temperature difference. The first temperature difference is compared with a preset temperature difference threshold. When the temperature difference is greater than or equal to the temperature difference threshold, a threshold point is selected within the first temperature difference to divide it into temperature difference sub-intervals with different thermal insulation performance levels. The emissivity of the drone camera terminal is set according to the material properties of the building exterior wall material to be detected. The emissivity is the ability of an object to emit infrared radiation. Different materials have different emissivities, which directly affects the temperature value displayed in the infrared image. For example, the emissivity of metals is usually low, while the emissivity of most non-metallic materials is high. If the emissivity is not set correctly, the temperature reading in the infrared image will be inaccurate. For example, if the emissivity of the metal is set too high, the measured temperature may be low. Therefore, the emissivity of the drone camera terminal is adjusted in advance according to the material properties to ensure the accuracy of the infrared image. The drone camera terminal performs an infrared scan on the building exterior wall materials to be inspected, obtains an infrared image of the building exterior wall materials to be inspected, and obtains the temperature corresponding to the grayscale value of each pixel in the infrared image based on a preset temperature lookup table of the drone camera terminal, the temperature lookup table includes temperatures corresponding to different grayscale values, obtains the temperature difference between the temperature corresponding to each pixel and the outdoor temperature of the building, marks the temperature difference as a second temperature difference, matches the second temperature difference with the temperature difference sub-intervals of different thermal insulation performance levels, obtains the thermal insulation performance level corresponding to each pixel, compares the thermal insulation performance level corresponding to each pixel with a preset thermal insulation performance level threshold, marks the distribution area of ​​pixels whose thermal insulation performance level is less than the thermal insulation performance level threshold as an unqualified thermal insulation performance area, marks the distribution area of ​​pixels whose thermal insulation performance level is greater than or equal to the thermal insulation performance level threshold as an area of ​​pending thermal insulation performance, and performs visual image detection operations on the area of ​​pending thermal insulation performance;

[0049] When the temperature difference is less than the temperature difference threshold, all areas covered by the building exterior wall materials to be inspected are marked as areas where the thermal insulation performance is pending, and visual image detection operations are performed on the areas where the thermal insulation performance is pending; the infrared thermal imager of the drone camera terminal is used to first detect areas with poor thermal insulation performance, and by comparing the temperature difference between different pixel points, it is determined whether the insulation layer is damaged or missing, and areas with unqualified thermal insulation performance are pre-screened, avoiding visual image detection of all areas of the building exterior wall materials to be inspected, reducing the consumption of shooting resources, and improving the efficiency, accuracy and comprehensiveness of visual inspection of building exterior wall insulation materials.

[0050] It should be further explained that, in the specific implementation process, the process of performing visual image detection operation on the area whose thermal insulation performance is to be determined and obtaining color abnormal pixel points and texture analysis angles includes:

[0051] A drone camera terminal is used to perform multi-angle image scanning on the area of ​​the building exterior wall material to be tested where the thermal insulation performance is to be determined, to obtain a multi-angle surface color image of the area where the thermal insulation performance is to be determined, and the multi-angle surface color image is converted from a preset BGR color space to an HSV color space, and then the color features of each pixel in the multi-angle surface color image are extracted, the color features including hue, saturation and brightness value, and the hue and saturation of each pixel are respectively compared with the hue and saturation corresponding to the material attribute to obtain the color difference value of each pixel, and the pixel whose color difference value is greater than the preset color difference value threshold is marked as a color abnormal pixel;

[0052] According to the corresponding brightness value of each pixel point in the multi-angle surface color image, the average brightness value and the standard deviation of the brightness value corresponding to different angles in the multi-angle surface color image are obtained, and the texture feature judgment value is obtained according to the brightness average value and the brightness standard deviation. The angle corresponding to the maximum texture feature judgment value at different angles is screened out as the texture analysis angle. A higher brightness average value usually means that the image is brighter as a whole, the contrast is better, and the texture feature is easier to identify, while a lower brightness average value may mean that the image is darker as a whole, the contrast is lower, and the texture feature may not be obvious enough. A larger brightness standard deviation indicates that there is a large brightness change in the image, which usually means that the image has rich details and contrast, which is conducive to texture analysis. A smaller brightness standard deviation indicates that the brightness change in the image is small, which may cause the texture feature to be unclear, which is not conducive to texture analysis. By analyzing the brightness average value and the brightness standard deviation corresponding to different angles in the multi-angle surface color image, the image angle most suitable for texture analysis is screened out, the influence of shadows and reflections on texture analysis is reduced, and the accuracy of texture analysis is improved;

[0053] The calculation formula for obtaining the color difference value of each pixel and the texture feature judgment value at different angles is:

[0054]

[0055] yp g =ω3*js g +ω4*rd g ;

[0056] Among them, sf i Represents the color difference value of pixel i, gh i represents the hue of pixel i, gh represents the hue corresponding to the material attribute, ur i represents the saturation of pixel i, ur represents the saturation corresponding to the material attribute, yp g Indicates the texture feature reading value corresponding to angle g, js g Represents the average brightness value corresponding to angle g, rd grepresents the standard deviation of the brightness value corresponding to the angle g, and ω1, ω2, ω3 and ω4 represent weight factors.

[0057] It should be further explained that, in the specific implementation process, the local area texture analysis is performed on the area whose thermal insulation performance is to be determined according to the texture analysis angle, and the process of obtaining the texture abnormal area includes:

[0058] The surface color image corresponding to the texture analysis angle in the multi-angle surface color image of the area with unknown thermal insulation performance is denoised and grayscaled to obtain a grayscale image. It should be further explained that in the specific implementation process, the multi-angle surface color image taken by the drone camera terminal may have noise problems. Therefore, in one embodiment of the present invention, after the image data is obtained, it is necessary to perform image preprocessing operations on it to improve the quality of the image data so as to obtain image data for subsequent analysis. Improving image quality through image preprocessing operations is an image processing method well known to those skilled in the art. In the embodiment of the invention, the noise in the image is removed by the Gaussian filtering method, and then the image data is equalized to improve the contrast within the image, and the image data is grayscaled to convert the color image into a grayscale image. The grayscale process is usually achieved by averaging the weights of the color channels, wherein denoising and equalization processing, semantic segmentation network and grayscale are technical means well known to those skilled in the art, and the specific steps are not repeated here. In other embodiments of the present invention, other image preprocessing means may also be used, which are not limited here. A region detection window and fixed parameters of a grayscale co-occurrence matrix are set, and a grayscale co-occurrence matrix of each local region in the grayscale image is obtained by sliding the region detection window. The texture structure of each local region in the grayscale image is obtained according to the grayscale co-occurrence matrix, and the texture structure of each local region is compared with the corresponding texture structure in the material property for similarity, to obtain the texture structure similarity of each local region, and mark the local region whose texture structure similarity is less than a preset similarity threshold as a texture abnormality region.

[0059] It should be further explained that, in a specific implementation process, the process of obtaining the texture structure of each local area in the grayscale image according to the gray level co-occurrence matrix includes:

[0060] The grayscale co-occurrence matrix of each local area in the grayscale image is obtained by sliding the area detection window. The specific details include quantizing the grayscale level of the surface image data into discrete grayscale levels, dividing the grayscale value range into several levels, for example, an 8-bit image can be divided into 16, 32, and 64 levels, and defining the parameters required for the grayscale co-occurrence matrix, including distance (d) and direction (θ), which are determined according to actual application requirements. The specific steps are not repeated here. Subsequently, based on the calculated grayscale co-occurrence matrix, a series of texture eigenvalues ​​are extracted to construct a texture structure. The texture eigenvalues ​​in the texture structure include: contrast, inverse variance, energy, homogeneity, and entropy.

[0061] The calculation formula for obtaining the texture structure similarity of each local area at the same time is:

[0062]

[0064] Among them, rl(e, v) represents the texture structure similarity between local area e and local area v, zf (e，v) represents the contrast between local area e and local area v, eh e represents the energy of the local area e, eh v represents the energy of the local region v, tk e represents the entropy of the local area e, tk v represents the entropy of the local region v, wp e represents the inverse variance of the local area e, wp v represents the inverse variance of the local region v, af e represents the homogeneity of the local region e, af v represents the homogeneity of the local region v, and ω3, ω4, ω5, ω6, and ω7 represent weight factors.

[0065] It should be further explained that, in the specific implementation process, the Canny algorithm is used to perform edge detection on each local area of ​​the area where the thermal insulation performance is to be determined, and the process of obtaining the damaged abnormal area includes:

[0066] Use the Canny algorithm to perform edge detection on each local area in the grayscale image, obtain the edge image of each local area, perform feature extraction on the edge image of each local area, and obtain the edge feature set of the edge image. The various types of edge features in the edge feature set include edge length (i.e., the total length of the edge, which can be used to evaluate the integrity of the edge), edge branch number (i.e., the number of branches of the edge, and the broken edge usually has more branches), continuity measure (used to define a measure to evaluate the continuity of the edge, the continuity measure is equal to the ratio of the total length of the edge break to the total length), edge curvature (i.e., the curvature distribution of the edge, a smooth edge usually has a smaller curvature change), and edge roughness (i.e., the standard deviation between edge pixels, used to evaluate the smoothness of the edge);

[0067] According to the various types of edge features in the edge feature set of the edge image as evaluation indicators, the indicator weight matrix of the evaluation indicators is set, and the membership matrix of the edge feature set for the preset complexity level is obtained through fuzzy comprehensive evaluation. The complexity level corresponding to the edge feature set is obtained according to the membership matrix and the indicator weight matrix, and the local area corresponding to the edge feature set with a complexity level greater than the preset complexity level threshold is marked as a damaged abnormal area. At the same time, the complexity level change trend of the local area that is not marked as a damaged abnormal area is predicted.

[0068] It should be further explained that, in the specific implementation process, the complexity level change trend of the local area not marked as the damaged abnormal area is predicted, and the process of obtaining the potential damaged area includes:

[0069] Pre-construct a visual inspection database, the visual inspection database is used to store historical visual inspection records of a number of building exterior wall materials, obtain historical complexity level time series sequences of each local area not marked as damaged abnormal area according to the historical visual inspection records, construct a complexity level estimation model based on deep learning, use the historical complexity level time series sequences of each local area not marked as damaged abnormal area as a training set and a test set, input the training set into the complexity level estimation model for training until the loss function training is stable, save the model parameters, test the complexity level estimation model through the test set until it meets the preset requirements, and output the complexity level estimation model;

[0070] Obtaining an estimated complexity level of each local area not marked as a damaged abnormal area according to the complexity level estimation model, comparing the estimated complexity level with a complexity level threshold, and if the estimated complexity level of the local area not marked as a damaged abnormal area is greater than the complexity level threshold, marking the local area not marked as a damaged abnormal area as a potential damaged area;

[0071] Subsequently, the distribution areas of abnormal color pixels, abnormal textures, abnormal damage and potential damage areas in the areas with unqualified thermal insulation performance and areas with pending thermal insulation performance are fed back to the relevant managers, who provide maintenance suggestions based on the feedback information. For example, for small-scale damage, direct repairs can be carried out. For large-area damage to the insulation material, overall replacement needs to be considered. For potential damage areas that are prone to problems, additional waterproofing or thermal insulation measures can be considered.

[0072] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, a person skilled in the art should understand that the technical method of the present invention may be modified or replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A visual inspection method for building exterior wall insulation materials, characterized in that: The following steps are involved: Step s1: obtaining material properties and external environment data of the building exterior wall material to be inspected, determining whether to perform a visual inspection operation of the building exterior wall material according to the external environment data, and obtaining shooting parameters of the building exterior wall material to be inspected when the external environment data meets the operation standard; Step s2: determine whether the external environment data of the building exterior wall material to be detected can be subjected to infrared image detection operation, and obtain the thermal insulation performance unqualified area and the thermal insulation performance pending area according to the determination result; Step s3: Perform visual image detection on the area whose thermal insulation performance is to be determined, obtain color abnormal pixel points and texture analysis angles, perform local area texture analysis on the area whose thermal insulation performance is to be determined according to the texture analysis angles, and obtain texture abnormal areas; Step s4: Use the Canny algorithm to perform edge detection on each local area of ​​the area where the thermal insulation performance is to be determined, and obtain the damaged abnormal area. At the same time, predict the complexity level change trend of the local area that is not marked as the damaged abnormal area to obtain the potential damaged area.

2. A visual inspection method for building exterior wall thermal insulation materials according to claim 1, characterized in that: The process of obtaining material properties and external environment data of the building exterior wall material to be inspected and determining whether to perform visual inspection of the building exterior wall material according to the external environment data includes: Obtain material properties and external environment data of the building exterior wall material to be detected, wherein the material properties include color characteristics, reflective characteristics and texture structure, and the external environment data include indoor building temperature, outdoor building temperature, light intensity, wind speed and precipitation intensity, obtain the building outdoor temperature, wind speed and precipitation intensity corresponding to the building exterior wall material to be detected, and compare the building outdoor temperature, wind speed and precipitation intensity corresponding to the building exterior wall material to be detected with the preset temperature limit threshold, the preset wind speed limit threshold and the preset precipitation intensity limit threshold respectively; if the building outdoor temperature, wind speed and precipitation intensity corresponding to the building exterior wall material to be detected are all less than the corresponding temperature limit threshold, wind speed limit threshold and precipitation intensity limit threshold, then perform a visual inspection operation on the building exterior wall material to be detected; if the building outdoor temperature, wind speed or precipitation intensity corresponding to the building exterior wall material to be detected is greater than the corresponding temperature limit threshold, wind speed limit threshold and precipitation intensity limit threshold, then suspend the visual inspection operation on the building exterior wall material to be detected.

3. A visual inspection method for building exterior wall thermal insulation materials according to claim 2, characterized in that: The process of obtaining the shooting parameters of the building exterior wall material to be inspected includes: According to the wind speed and precipitation intensity corresponding to the building exterior wall materials to be tested as evaluation indicators, the indicator weights of the evaluation indicators are set, and the membership matrix of the building exterior wall materials to be tested for the preset interference level is obtained through fuzzy comprehensive evaluation; The interference level of the building exterior wall material to be detected is obtained through the membership matrix and the indicator weight, and an adaptive shooting parameter comparison table is pre-constructed. The adaptive shooting parameter comparison table includes shooting parameters corresponding to different interference levels. The shooting parameters include resolution and frame number. The shooting parameters of the building exterior wall material to be detected are determined according to the interference level of the building exterior wall material to be detected.

4. A visual inspection method for building exterior wall thermal insulation materials according to claim 3, characterized in that: The process of judging whether the external environment data of the building exterior wall material to be inspected can be subjected to infrared image inspection operation, and obtaining the area with unqualified thermal insulation performance and the area with undetermined thermal insulation performance according to the judgment result includes: The temperature difference between the indoor temperature of the building and the outdoor temperature of the building is obtained according to the external environment data of the building exterior wall material to be detected, the temperature difference is marked as a first temperature difference, the first temperature difference is compared with a preset temperature difference threshold, when the temperature difference is greater than or equal to the temperature difference threshold, a threshold point is selected within the first temperature difference to divide the temperature difference sub-intervals into different thermal insulation performance levels, the emissivity of the drone camera terminal is set according to the material properties of the building exterior wall material to be detected, the drone camera terminal is used to perform infrared scanning on the building exterior wall material to be detected, an infrared image is obtained, and the temperature corresponding to the grayscale value of each pixel in the infrared image is obtained based on a preset temperature lookup table of the drone camera terminal. degree, obtaining the temperature difference between the temperature corresponding to each pixel and the outdoor temperature of the building, marking the temperature difference as a second temperature difference, matching the second temperature difference with the temperature difference sub-intervals of different thermal insulation performance levels, obtaining the thermal insulation performance level corresponding to each pixel, comparing the thermal insulation performance level corresponding to each pixel with a preset thermal insulation performance level threshold, marking the distribution area of ​​the pixels whose thermal insulation performance level is less than the thermal insulation performance level threshold as an unqualified thermal insulation performance area, marking the distribution area of ​​the pixels whose thermal insulation performance level is greater than or equal to the thermal insulation performance level threshold as an undetermined thermal insulation performance area, and performing a visual image detection operation on the undetermined thermal insulation performance area; When the temperature difference is less than the temperature difference threshold, all areas covered by the exterior wall materials of the building to be inspected are marked as areas whose thermal insulation performance is pending, and visual image inspection operations are performed on the areas whose thermal insulation performance is pending.

5. A visual inspection method for building exterior wall thermal insulation materials according to claim 4, characterized in that: The process of performing visual image detection on the area to be determined for thermal insulation performance and obtaining color abnormality pixels and texture analysis angles includes: A drone camera terminal is used to perform multi-angle image scanning on the area of ​​the building exterior wall material to be tested where the thermal insulation performance is to be determined, to obtain a multi-angle surface color image of the area where the thermal insulation performance is to be determined, and the multi-angle surface color image is converted from a preset BGR color space to an HSV color space, and then the color features of each pixel in the multi-angle surface color image are extracted, the color features including hue, saturation and brightness value, and the hue and saturation of each pixel are respectively compared with the hue and saturation corresponding to the material attribute to obtain the color difference value of each pixel, and the pixel whose color difference value is greater than the preset color difference value threshold is marked as a color abnormal pixel; According to the corresponding brightness value of each pixel point in the multi-angle surface color image, the average brightness value and the standard deviation of the brightness value corresponding to different angles in the multi-angle surface color image are obtained, and the texture feature judgment value is obtained according to the brightness average value and the standard deviation of the brightness value, and the angle corresponding to the maximum texture feature judgment value at different angles is screened out as the texture analysis angle.

6. A visual inspection method for building exterior wall thermal insulation materials according to claim 5, characterized in that: According to the texture analysis angle, local area texture analysis is performed on the area where the thermal insulation performance is to be determined. The process of obtaining the texture abnormal area includes: The surface color image corresponding to the texture analysis angle in the multi-angle surface color image of the area with undetermined thermal insulation performance is denoised and grayscaled to obtain a grayscale image, and a region detection window and fixed parameters of a grayscale co-occurrence matrix are set. The grayscale co-occurrence matrix of each local region in the grayscale image is obtained by sliding the region detection window. The texture structure of each local region in the grayscale image is obtained according to the grayscale co-occurrence matrix, and the texture structure of each local region is compared with the corresponding texture structure in the material property for similarity to obtain the texture structure similarity of each local region, and the local region whose texture structure similarity is less than a preset similarity threshold is marked as a texture abnormality region.

7. A visual inspection method for building exterior wall thermal insulation materials according to claim 6, characterized in that: The Canny algorithm is used to perform edge detection on each local area of ​​the area where the thermal insulation performance is to be determined. The process of obtaining the damaged abnormal area includes: The Canny algorithm is used to perform edge detection on each local area in the grayscale image to obtain the edge image of each local area, and the feature extraction is performed on the edge image of each local area to obtain the edge feature set of the edge image. The various types of edge features in the edge feature set include edge length, edge branch number, continuity measure, edge curvature, and edge roughness; According to the various types of edge features in the edge feature set of the edge image as evaluation indicators, the indicator weight matrix of the evaluation indicators is set, and the membership matrix of the edge feature set for the preset complexity level is obtained through fuzzy comprehensive evaluation. The complexity level corresponding to the edge feature set is obtained according to the membership matrix and the indicator weight matrix, and the local area corresponding to the edge feature set with a complexity level greater than the preset complexity level threshold is marked as a damaged abnormal area. At the same time, the complexity level change trend of the local area that is not marked as a damaged abnormal area is predicted.

8. A visual inspection method for building exterior wall thermal insulation materials according to claim 7, characterized in that: The process of predicting the complexity level change trend of the local area not marked as the damaged abnormal area and obtaining the potential damaged area includes: Pre-construct a visual inspection database, the visual inspection database is used to store historical visual inspection records of a number of building exterior wall materials, obtain historical complexity level time series sequences of each local area not marked as damaged abnormal area according to the historical visual inspection records, construct a complexity level estimation model based on deep learning, use the historical complexity level time series sequences of each local area not marked as damaged abnormal area as a training set and a test set, input the training set into the complexity level estimation model for training until the loss function training is stable, save the model parameters, test the complexity level estimation model through the test set until it meets the preset requirements, and output the complexity level estimation model; According to the complexity level estimation model, the estimated complexity level of each local area not marked as a damaged abnormal area is obtained, and the estimated complexity level is compared with the complexity level threshold. If the estimated complexity level of the local area not marked as a damaged abnormal area is greater than the complexity level threshold, the local area not marked as a damaged abnormal area is marked as a potential damaged area.

Images