Method and system for detecting apparent quality of lightweight partition walls based on deep learning

Abstract

The present invention relates to the field of partition wall quality detection, and proposes a method and system for realizing the apparent quality detection of lightweight partition walls based on deep learning. The method comprises: obtaining an apparent image of an installed partition wall, marking an image label, constructing an apparent image set, performing image smoothing processing on the apparent image set to obtain a smoothed image set, identifying grayscale and querying the pixel quantity, calculating an apparent histogram, calculating the cumulative distribution value of each grayscale in the apparent histogram, and based on these cumulative distribution values, performing regional segmentation on the partition wall apparent image using a pre-constructed partition wall apparent quality inspection model to obtain an apparent segmented region, and calculating a texture feature value, performing clustering processing on the texture feature to obtain a partition wall apparent quality category, performing quality marking on the apparent segmented region according to the apparent quality category, and generating an apparent quality inspection report of the installed lightweight partition wall according to the regional quality marking. The present invention can improve the apparent quality detection efficiency of lightweight partition walls.

Description

Technical Field

[0001] The present invention relates to the field of partition wall quality detection, and in particular to a method and system for realizing apparent quality detection of lightweight partition walls based on deep learning. Background Art

[0002] Lightweight partition walls are partition walls constructed with lightweight materials. They are usually made of lightweight bricks, lightweight steel keels, lightweight gypsum boards and other materials. Compared with traditional solid walls, lightweight partition walls have the advantages of light weight, fast construction, good sound insulation, and can be disassembled, adjusted and reassembled. They are used to divide space and define room functional areas. They can also be used for sound insulation and fire prevention.

[0003] At present, the commonly used apparent quality inspection contents of lightweight partition walls include: wall flatness, wall verticality, surface smoothness, burrs and damage, etc., and lightweight partition walls are often measured by using testing instruments, such as: laser rangefinder, level, inclinometer, etc. These instruments can measure the wall flatness, wall verticality, etc. However, since more testing instruments are needed to realize the apparent quality inspection of lightweight partition walls, the inspection efficiency is low. Therefore, a method and system for realizing the apparent quality inspection of lightweight partition walls based on deep learning is needed to improve the apparent quality inspection efficiency of lightweight partition walls. Summary of the invention

[0004] The present invention provides a method and system for detecting the apparent quality of lightweight partition walls based on deep learning, and its main purpose is to improve the efficiency of detecting the apparent quality of lightweight partition walls.

[0005] To achieve the above-mentioned purpose, the present invention provides a method for detecting the apparent quality of a lightweight partition wall based on deep learning, comprising:

[0006] Obtaining a partition wall appearance image corresponding to an installed lightweight partition wall, marking an image label corresponding to the partition wall appearance image, and constructing an appearance image set corresponding to the partition wall appearance image based on the image label;

[0007] Performing image smoothing processing on the apparent image set to obtain a smoothed image set, identifying gray levels corresponding to images in the smoothed image set, querying the number of pixels corresponding to the gray levels, and calculating an apparent histogram corresponding to the apparent image set based on the gray levels and the number of pixels;

[0008] Based on the apparent histogram, calculating the cumulative distribution value corresponding to each gray level in the apparent histogram;

[0009] Based on the cumulative distribution value, the partition wall appearance image is segmented using a pre-built partition wall appearance quality inspection model to obtain an appearance segmentation area, the texture feature value corresponding to the appearance segmentation area is calculated, and the texture feature is clustered to obtain an appearance quality category corresponding to the partition wall appearance image;

[0010] Based on the apparent quality category, the apparent segmented areas are quality marked to obtain regional quality marks, and based on the regional quality marks, an apparent quality inspection report corresponding to the installed lightweight partition wall is generated.

[0011] Optionally, the image tag corresponding to the partition wall apparent image includes:

[0012] Performing image adjustment on the partition wall surface image to obtain an adjusted image;

[0013] Extracting a partition wall region in the adjusted image; identifying partition wall features in the partition wall region;

[0014] Marking a feature sequence corresponding to the partition wall feature;

[0015] Based on the feature sequence, an image label corresponding to the partition wall appearance image is marked.

[0016] Optionally, constructing an apparent image set corresponding to the partition wall apparent image based on the image label includes:

[0017] Annotate a label sequence corresponding to the image label;

[0018] Based on the tag sequence, dividing the apparent image pairs corresponding to the partition wall apparent image;

[0019] Extract the apparent features in the apparent image pair, and identify the feature vectors corresponding to the apparent features; and construct the apparent image set corresponding to the partition wall apparent image based on the feature vectors.

[0020] Optionally, performing image smoothing processing on the apparent image set to obtain a smoothed image set includes:

[0021] Converting each image in the apparent image set into a grayscale image;

[0022] Performing image segmentation on the grayscale image to obtain small-block images;

[0023] Calculating the local pixel average of the small block image;

[0024] Taking the local pixel average as the central pixel value corresponding to the small block image;

[0025] Based on the central pixel value, the small block images are combined into a smooth image;

[0026] The smoothed images are stored to obtain a smoothed image set.

[0027] Optionally, calculating the local pixel average value of the small block image includes:

[0028] The local pixel average value of the small image is calculated using the following formula:

[0029]

[0030] Among them, PJ represents the local pixel average value of the small image, m and n respectively represent the number of rows and columns corresponding to the small image, and I(i,j) represents the pixel value of the i-th row and j-th column in the small image.

[0031] Optionally, the calculating, based on the grayscale and the pixel quantity, an apparent histogram corresponding to the apparent image set comprises:

[0032] The apparent histogram corresponding to the apparent image set is calculated using the following formula:

[0033]

[0034] Among them, C(gk) represents the number of occurrences of pixels in the apparent image set, gk represents the grayscale, n represents the pixel quantity, I represents the apparent image set, i represents the pixel index in the apparent image set, P(gk) represents the normalized frequency obtained by normalizing the number of occurrences of pixels of each grayscale level, and H(gk) represents the apparent histogram corresponding to the apparent image set.

[0035] Optionally, the calculating, based on the apparent histogram, the cumulative distribution value corresponding to each gray level in the apparent histogram includes:

[0036] The cumulative distribution value corresponding to each gray level in the apparent histogram is calculated using the following formula:

[0037]

[0038] Among them, S(gk) refers to the cumulative sum of the number of pixels in the apparent histogram, gk represents the gray level, k represents the index of the gray level, H(gk) represents the apparent histogram, N represents the total number of pixels in the apparent histogram, L represents the total number of gray levels, and D(gk) represents the cumulative distribution value corresponding to each gray level in the apparent histogram.

[0039] Optionally, the calculating the corresponding texture feature value in the apparent segmented area includes: identifying the corresponding grayscale value and average grayscale value in the apparent segmented area;

[0040] First, the difference between the grayscale value and the average grayscale value is calculated using the following formula:

[0041] D i (x,y)=E(x,y)-μ i

[0042] Among them, D i (x, y) represents the difference value of the pixel (x, y) in the region i corresponding to the apparent segmentation region, E(x, y) represents the grayscale value of the pixel (x, y) in the image corresponding to the apparent segmentation region, μ i represents the average gray value;

[0043] Secondly, the standard deviation corresponding to the difference value is calculated using the following formula:

[0044]

[0045] Among them, σ i represents the standard deviation corresponding to the difference value, W and H represent the width and height of the image corresponding to the apparent segmented area, respectively, and (x, y) represents the pixel position of the image corresponding to the apparent segmented area;

[0046] Furthermore, the grayscale difference mean corresponding to the apparent segmented area is calculated using the following formula:

[0047]

[0048] Wherein, M represents the mean grayscale difference corresponding to the apparent segmented region, N represents the number of regions corresponding to the apparent segmented region, μ i represents the average gray value of the ith region, μ j represents the average gray value of the jth region;

[0049] Finally, the corresponding texture feature value in the apparent segmentation area is calculated using the following formula:

[0050]

[0051] Wherein, F represents the texture feature value corresponding to the apparent segmentation area, (σ i / (1+M)) means normalizing the standard deviation of each region.

[0052] Optionally, the performing quality marking on the apparent segmented region based on the apparent quality category to obtain the region quality mark includes:

[0053] Identifying the quality standard corresponding to the apparent quality category;

[0054] extracting regional characteristics in the apparent segmented region based on the quality criterion;

[0055] Based on the regional characteristics, calculating a quality score corresponding to the apparent segmented region;

[0056] Based on the quality score, the apparent segmented region is quality marked to obtain a region quality mark.

[0057] In order to solve the above problems, the present invention also provides a system for detecting the apparent quality of lightweight partition walls based on deep learning, the system comprising:

[0058] An image set construction module, used to obtain an apparent image of a partition wall corresponding to an installed lightweight partition wall, mark an image tag corresponding to the apparent image of the partition wall, and construct an apparent image set corresponding to the apparent image of the partition wall based on the image tag;

[0059] a histogram calculation module, configured to perform image smoothing processing on the apparent image set to obtain a smoothed image set, identify gray levels corresponding to images in the smoothed image set, query the number of pixels corresponding to the gray levels, and calculate an apparent histogram corresponding to the apparent image set based on the gray levels and the number of pixels;

[0060] A distribution value calculation module, used for calculating the cumulative distribution value corresponding to each gray level in the apparent histogram based on the apparent histogram;

[0061] A feature clustering module, for performing image region segmentation on the partition wall appearance image based on the cumulative distribution value using a pre-built partition wall appearance quality inspection model to obtain an appearance segmentation region, calculating a texture feature value corresponding to the appearance segmentation region, and performing clustering processing on the texture feature to obtain an appearance quality category corresponding to the partition wall appearance image;

[0062] A report generation module is used to mark the quality of the apparent segmented areas based on the apparent quality categories to obtain regional quality marks, and to generate an apparent quality inspection report corresponding to the installed lightweight partition wall based on the regional quality marks.

[0063] The present invention obtains the partition wall surface image corresponding to the installed lightweight partition wall and marks the image label corresponding to the partition wall surface image, so as to improve the data processing efficiency, ensure the data accuracy and consistency, save time and labor costs, and provide valuable information and guidance for subsequent data analysis and application. The present invention performs image smoothing processing on the surface image set to obtain a smoothed image set, which can reduce the image size, simplify the image structure, reduce noise interference, provide more accurate and stable results, and thus reduce the calculation complexity. Based on the cumulative distribution value, the present invention uses a pre-constructed partition wall surface quality inspection model to perform image region segmentation on the partition wall surface image to obtain the surface segmentation region, which can be used as required. It is necessary to optimize and adjust to adapt to different image segmentation tasks, and the model can be improved by increasing training samples or changing the model structure, so as to improve the robustness and generalization ability of the algorithm. The present invention marks the quality of the apparent segmentation area based on the apparent quality category to obtain a regional quality mark, which can provide more accurate quality assessment, guide improvement measures, and optimize resource allocation, which is helpful to improve the quality and efficiency of partition wall construction. The present invention generates an apparent quality inspection report corresponding to the installed lightweight partition wall based on the regional quality mark, which can provide a basis for quality monitoring, analysis and traceability, help improve the level of quality management, reduce the occurrence of quality problems, and ensure the overall quality of the construction project. Therefore, the present invention proposes a method and system for realizing the apparent quality detection of lightweight partition walls based on deep learning to improve the efficiency of apparent quality detection of lightweight partition walls. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 A schematic diagram of a flow chart of a method for detecting the apparent quality of a lightweight partition wall based on deep learning provided in one embodiment of the present invention;

[0065] Figure 2 A schematic diagram of a module of a system for detecting the apparent quality of a lightweight partition wall based on deep learning provided by an embodiment of the present invention;

[0066] Figure 3 A schematic diagram of the internal structure of an electronic device for implementing a method for detecting the apparent quality of lightweight partition walls based on deep learning provided in one embodiment of the present invention.

[0067] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0068] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0069] The embodiment of the present application provides a method for detecting the apparent quality of lightweight partition walls based on deep learning. The execution subject of the method for detecting the apparent quality of lightweight partition walls based on deep learning includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for detecting the apparent quality of lightweight partition walls based on deep learning can be executed by software or hardware installed on a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0070] Reference Figure 1 FIG. 1 is a flow chart of a method for detecting the apparent quality of a lightweight partition wall based on deep learning according to an embodiment of the present invention. In this embodiment, the method for detecting the apparent quality of a lightweight partition wall based on deep learning includes:

[0071] S1. Obtain an apparent image of a partition wall corresponding to an installed lightweight partition wall, mark an image tag corresponding to the apparent image of the partition wall, and construct an apparent image set corresponding to the apparent image of the partition wall based on the image tag.

[0072] The present invention obtains the partition wall surface image corresponding to the installed lightweight partition wall and marks the image label corresponding to the partition wall surface image, so as to improve data processing efficiency, ensure data accuracy and consistency, save time and labor costs, and provide valuable information and guidance for subsequent data analysis and application.

[0073] The partition wall appearance image refers to the appearance image corresponding to the lightweight partition wall, including the shape, color, texture and other features of the partition wall; the image label refers to the key information or feature used to mark the partition wall appearance image.

[0074] As an embodiment of the present invention, the marking of the image label corresponding to the partition wall apparent image includes: performing image adjustment on the partition wall apparent image to obtain an adjusted image; extracting a partition wall area in the adjusted image; identifying partition wall features in the partition wall area; marking a feature sequence corresponding to the partition wall features; and marking the image label corresponding to the partition wall apparent image based on the feature sequence.

[0075] Among them, the adjusted image refers to an image that is processed or changed from the partition wall apparent image; the partition wall area refers to the area related to the partition wall extracted from the adjusted image; the partition wall features refer to feature points, textures, colors, etc. with unique features or attributes in the partition wall area; the feature sequence refers to a feature sequence formed according to the order or combination of the partition wall features, which is used to mark and identify the partition wall apparent image.

[0076] Furthermore, the adjusted image can be obtained through image editing tools, such as Adobe Photoshop, GIMP and other tools; the partition wall area can be obtained through an instance segmentation model, such as Mask R-CNN, YOLO and other models; the partition wall features can be obtained through feature extraction algorithms, such as MFCC algorithm, PLP algorithm, etc.; the feature sequence can be obtained through target detection algorithms, such as Faster R-CNN, SSD and other algorithms.

[0077] The present invention constructs an apparent image set corresponding to the apparent image of the partition wall based on the image label, which can improve the accuracy of the image label, improve the diversity of the apparent image, strengthen the image classification performance and enhance the reliability of image recognition, thereby improving the effect and accuracy in subsequent practical applications.

[0078] The apparent image set refers to a group of images with similar visual features, which are obtained by photographing or generating the same object under different viewing angles, lighting conditions and environments.

[0079] As an embodiment of the present invention, constructing an apparent image set corresponding to the partition wall apparent image based on the image label includes: marking a label sequence corresponding to the image label; dividing the apparent image pairs corresponding to the partition wall apparent image based on the label sequence; extracting the apparent features in the apparent image pairs; identifying the feature vector corresponding to the apparent feature; and constructing the apparent image set corresponding to the partition wall apparent image based on the feature vector.

[0080] Among them, the label sequence refers to the sequence that corresponds the input sequence to the corresponding label one by one in tasks such as sequence annotation or speech recognition; the appearance image pair refers to the features extracted from the image that represent the appearance, shape, texture, etc. of the image; the feature vector refers to the representation form of converting an image or other data into a numerical vector.

[0081] Furthermore, the label sequence can be obtained through sequence annotation tools, such as OpenCV, ffmpeg and other tools; the apparent image pair can be obtained through image pair generation tools, such as TensorFlow, PyTorch, Keras and the like; the feature vector can be obtained through feature extraction tools, such as scikit-learn, NLTK and other tools.

[0082] S2. Perform image smoothing processing on the apparent image set to obtain a smoothed image set, identify the grayscale corresponding to the image in the smoothed image set, and query the pixel quantity corresponding to the grayscale, and calculate the apparent histogram corresponding to the apparent image set based on the grayscale and the pixel quantity.

[0083] The present invention performs image smoothing processing on the apparent image set to obtain a smoothed image set, which can reduce the image size, simplify the image structure, reduce noise interference, provide more accurate and stable results, and thus reduce calculation complexity.

[0084] The smoothed image set refers to performing a smoothing operation on each image in the image set to obtain a set of smoothed images.

[0085] As an embodiment of the present invention, the image smoothing processing is performed on the apparent image set to obtain a smoothed image set, including: converting each image in the apparent image set into a grayscale image; performing image segmentation on the grayscale image to obtain small-block images; calculating the local pixel average of the small-block images; using the local pixel average as the central pixel value corresponding to the small-block image; combining the small-block images into a smoothed image based on the central pixel value; and performing image storage on the smoothed images to obtain a smoothed image set.

[0086] Among them, the grayscale image refers to an image in which each pixel is represented by a grayscale value; the small-block image refers to an image divided into several regions of equal size, each region is called a small-block image; the local pixel average value refers to the average value of the grayscale values ​​of all pixels in a small-block image; the central pixel value refers to the grayscale value of the pixel located at the center position in a small-block image; the smoothed image refers to the image obtained by the preprocessing step for image enhancement or feature extraction.

[0087] Furthermore, the grayscale image can be obtained through image processing tools, such as Adobe Photoshop, GIMP, ImageJ and other tools; the small block image can be obtained through the cv2.imread() function in OpenCV; the local pixel average value can be obtained through the following calculation formula; the central pixel value can be obtained through an image interpolation algorithm; the smoothed image can be obtained through image filtering tools, such as OpenCV, PIL and scikit-image and other tools.

[0088] As an embodiment of the present invention, the calculating of the local pixel average value of the small block image includes:

[0089] The local pixel average value of the small image is calculated using the following formula:

[0090]

[0091] Among them, PJ represents the local pixel average value of the small image, m and n respectively represent the number of rows and columns corresponding to the small image, and I(i,j) represents the pixel value of the i-th row and j-th column in the small image.

[0092] The present invention can provide understanding and accurate statistics of image brightness distribution by identifying the grayscale corresponding to the image in the smoothed image set and querying the pixel quantity corresponding to the grayscale, which has important benefits for applications such as image processing, classification and target detection.

[0093] The grayscale refers to the brightness level of each pixel in the image; the pixel quantity refers to the number of pixels in the image.

[0094] Optionally, the grayscale can be obtained through a CNN model, such as LeNet, AlexNet, VGG and other models; the pixel quantity can be obtained through an image processing tool, such as OpenCV, Photoshop and other tools.

[0095] Further, as an embodiment of the present invention, the calculating the apparent histogram corresponding to the apparent image set based on the grayscale and the pixel quantity includes:

[0096] The apparent histogram corresponding to the apparent image set is calculated using the following formula:

[0097]

[0098] Among them, C(gk) represents the number of occurrences of pixels in the apparent image set, gk represents the grayscale, n represents the pixel quantity, I represents the apparent image set, i represents the pixel index in the apparent image set, P(gk) represents the normalized frequency obtained by normalizing the number of occurrences of pixels of each grayscale level, and H(gk) represents the apparent histogram corresponding to the apparent image set.

[0099] S3. Based on the apparent histogram, calculate the cumulative distribution value corresponding to each gray level in the apparent histogram.

[0100] As an embodiment of the present invention, the calculating, based on the apparent histogram, the cumulative distribution value corresponding to each gray level in the apparent histogram includes:

[0101] The cumulative distribution value corresponding to each gray level in the apparent histogram is calculated using the following formula:

[0102]

[0103] Among them, S(gk) refers to the cumulative sum of the number of pixels in the apparent histogram, gk represents the gray level, k represents the index of the gray level, H(gk) represents the apparent histogram, N represents the total number of pixels in the apparent histogram, L represents the total number of gray levels, and D(gk) represents the cumulative distribution value corresponding to each gray level in the apparent histogram.

[0104] S4. Based on the cumulative distribution value, use a pre-constructed partition wall appearance quality inspection model to perform image region segmentation on the partition wall appearance image to obtain the appearance segmentation region, calculate the corresponding texture feature value in the appearance segmentation region, perform clustering processing on the texture feature value, and obtain the appearance quality category corresponding to the partition wall appearance image.

[0105] Based on the cumulative distribution value, the present invention uses a pre-built partition wall appearance quality inspection model to perform image region segmentation on the partition wall appearance image to obtain the appearance segmentation area, which can be optimized and adjusted as needed to adapt to different image segmentation tasks, and the model can be improved by adding training samples or changing the model structure, thereby improving the robustness and generalization ability of the algorithm.

[0106] Among them, the pre-constructed partition wall appearance quality inspection model refers to a model constructed by using machine learning or deep learning technology through labeled image data; the appearance segmentation area refers to dividing the image into different areas, each area contains pixels with similar appearance features.

[0107] Optionally, the partition wall appearance quality inspection model can be obtained through a deep learning model, such as support vector machine, random forest, decision tree and other models; the appearance segmentation area can be obtained through a superpixel segmentation algorithm, such as SLIC, SLICO and other algorithms.

[0108] As an embodiment of the present invention, the calculating of the corresponding texture feature value in the apparent segmented area includes: identifying the corresponding grayscale value and average grayscale value in the apparent segmented area;

[0109] First, the difference between the grayscale value and the average grayscale value is calculated using the following formula:

[0110] D i (x,y)=E(x,y)-μ i

[0111] Among them, D i (x, y) represents the difference value of the pixel (x, y) in the region i corresponding to the apparent segmentation region, E(x, y) represents the grayscale value of the pixel (x, y) in the image corresponding to the apparent segmentation region, μ i represents the average gray value;

[0112] Secondly, the standard deviation corresponding to the difference value is calculated using the following formula:

[0113]

[0114] Among them, σ i represents the standard deviation corresponding to the difference value, W and H represent the width and height of the image corresponding to the apparent segmented area, respectively, and (x, y) represents the pixel position of the image corresponding to the apparent segmented area;

[0115] Furthermore, the grayscale difference mean corresponding to the apparent segmented area is calculated using the following formula:

[0116]

[0117] Wherein, M represents the mean grayscale difference corresponding to the apparent segmented region, N represents the number of regions corresponding to the apparent segmented region, μ i represents the average gray value of the ith region, μ j represents the average gray value of the jth region;

[0118] Finally, the corresponding texture feature value in the apparent segmentation area is calculated using the following formula:

[0119]

[0120] Wherein, F represents the texture feature value corresponding to the apparent segmentation area, (σi / (1+M)) means normalizing the standard deviation of each region.

[0121] The present invention obtains the apparent quality category corresponding to the partition wall surface image by clustering the texture feature values, which can help to quickly and accurately evaluate the apparent quality, achieve quality management, guide repair work and data analysis optimization, and thus improve the efficiency and quality of partition wall construction.

[0122] The apparent quality category refers to the classification of the partition wall apparent image into different categories or grades after clustering the partition wall apparent image according to its texture feature value, and each category represents an apparent quality level or degree.

[0123] Optionally, the apparent quality category can be obtained through feature selection and dimensionality reduction models, such as PCA, LDA and other models.

[0124] S5. Based on the apparent quality category, the apparent segmented areas are quality marked to obtain regional quality marks, and based on the regional quality marks, an apparent quality inspection report corresponding to the installed lightweight partition wall is generated.

[0125] The present invention performs quality marking on the apparent segmented areas based on the apparent quality categories to obtain regional quality markings, which can provide more accurate quality assessments, guide improvement measures, optimize resource allocation, and help improve the quality and efficiency of partition wall construction.

[0126] The regional quality marking refers to evaluating and marking a specific region to determine the quality of the region in terms of a certain quality indicator.

[0127] As an embodiment of the present invention, the quality marking of the apparent segmented region based on the apparent quality category to obtain the regional quality marking includes: identifying the quality standard corresponding to the apparent quality category; extracting the regional characteristics in the apparent segmented region based on the quality standard; calculating the quality score corresponding to the apparent segmented region based on the regional characteristics; and quality marking the apparent segmented region based on the quality score to obtain the regional quality marking.

[0128] Among them, the quality standard refers to the standard for different apparent quality categories, which can be a pre-defined series of rules or indicators; the regional characteristics refer to the quality-related features extracted from the apparent segmented region; and the quality score refers to the quality level used to represent the apparent segmented region.

[0129] Furthermore, the quality standard can be obtained through a deep learning model, such as the Pix2PixGAN model; the regional characteristics can be obtained through an extraction and recognition algorithm, such as a spatial analysis algorithm, a convex hull algorithm, a K-means algorithm, etc.; the quality score can be obtained through a variety of algorithms, such as an average value algorithm, a weighted average value algorithm, a factor analysis algorithm, etc.

[0130] The present invention generates an apparent quality inspection report corresponding to the installed lightweight partition wall based on the regional quality mark, which can provide a basis for quality monitoring, analysis and traceability, help improve the quality management level, reduce the occurrence of quality problems, and ensure the overall quality of the construction project.

[0131] Among them, the appearance quality inspection report refers to a report on the quality inspection and evaluation of the appearance and appearance of the lightweight partition wall.

[0132] Optionally, the appearance quality inspection report can be obtained through a report generation tool, such as LaTeX, Tableau and other tools.

[0133] The present invention obtains the partition wall surface image corresponding to the installed lightweight partition wall and marks the image label corresponding to the partition wall surface image, so as to improve the data processing efficiency, ensure the data accuracy and consistency, save time and labor costs, and provide valuable information and guidance for subsequent data analysis and application. The present invention performs image smoothing processing on the surface image set to obtain a smoothed image set, which can reduce the image size, simplify the image structure, reduce noise interference, provide more accurate and stable results, and thus reduce the calculation complexity. Based on the cumulative distribution value, the present invention uses a pre-constructed partition wall surface quality inspection model to perform image region segmentation on the partition wall surface image to obtain the surface segmentation region, which can be used as required. It is necessary to optimize and adjust to adapt to different image segmentation tasks, and the model can be improved by increasing training samples or changing the model structure, so as to improve the robustness and generalization ability of the algorithm. The present invention marks the quality of the apparent segmentation area based on the apparent quality category to obtain a regional quality mark, which can provide more accurate quality assessment, guide improvement measures, and optimize resource allocation, which is helpful to improve the quality and efficiency of partition wall construction. The present invention generates an apparent quality inspection report corresponding to the installed lightweight partition wall based on the regional quality mark, which can provide a basis for quality monitoring, analysis and traceability, help improve the level of quality management, reduce the occurrence of quality problems, and ensure the overall quality of the construction project. Therefore, the present invention proposes a method and system for realizing the apparent quality detection of lightweight partition walls based on deep learning to improve the efficiency of apparent quality detection of lightweight partition walls.

[0134] like Figure 2As shown, it is a functional module diagram of a method and system for detecting the apparent quality of lightweight partition walls based on deep learning provided by one embodiment of the present invention.

[0135] The system 200 for detecting the apparent quality of lightweight partition walls based on deep learning of the present invention can be installed in an electronic device. According to the functions implemented, the system 200 for detecting the apparent quality of lightweight partition walls based on deep learning can include an image set construction module 201, a histogram calculation module 202, a distribution value calculation module 203, a feature clustering module 204, and a report generation module 205. The module of the present invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0136] In this embodiment, the functions of each module / unit are as follows:

[0137] The image set construction module 201 is used to obtain the partition wall appearance image corresponding to the installed lightweight partition wall, mark the image label corresponding to the partition wall appearance image, and construct the appearance image set corresponding to the partition wall appearance image based on the image label;

[0138] The histogram calculation module 202 is used to perform image smoothing processing on the apparent image set to obtain a smoothed image set, identify the grayscale corresponding to the image in the smoothed image set, and query the pixel quantity corresponding to the grayscale, and calculate the apparent histogram corresponding to the apparent image set based on the grayscale and the pixel quantity;

[0139] The distribution value calculation module 203 is used to calculate the cumulative distribution value corresponding to each gray level in the apparent histogram based on the apparent histogram;

[0140] The feature clustering module 204 is used to perform image region segmentation on the partition wall appearance image based on the cumulative distribution value using a pre-built partition wall appearance quality inspection model to obtain an appearance segmentation region, calculate the texture feature value corresponding to the appearance segmentation region, perform clustering processing on the texture feature, and obtain an appearance quality category corresponding to the partition wall appearance image;

[0141] The report generation module 205 is used to perform quality marking on the apparent segmented areas based on the apparent quality categories to obtain regional quality markings, and to generate an apparent quality inspection report corresponding to the installed lightweight partition wall based on the regional quality markings.

[0142] In detail, the modules described in the system 200 for detecting the apparent quality of lightweight partition walls based on deep learning described in the embodiment of the present invention adopt the same technical means as the method for detecting the apparent quality of lightweight partition walls based on deep learning described in the accompanying drawings when used, and can produce the same technical effects, which will not be repeated here.

[0143] like Figure 3 Shown is a schematic diagram of the structure of an electronic device that implements a method for detecting the apparent quality of lightweight partition walls based on deep learning according to the present invention.

[0144] The electronic device 1 may include a processor 30, a memory 31, a communication bus 32 and a communication interface 33, and may also include a computer program stored in the memory 31 and executable on the processor 30, such as an engineering safety supervision program based on artificial intelligence.

[0145] In some embodiments, the processor 30 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 30 is the control core (ControlUnit) of the electronic device 1, and uses various interfaces and lines to connect the various components of the entire electronic device, and executes or executes programs or modules stored in the memory 31 (for example, executing engineering safety supervision programs based on artificial intelligence, etc.), and calls data stored in the memory 31 to execute various functions of the electronic device and process data.

[0146] The memory 31 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 31 may be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 31 may also be an external storage device of an electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the memory 31 may also include both an internal storage unit of the electronic device and an external storage device. The memory 31 may not only be used to store application software and various types of data installed in the electronic device, such as the code of the database configuration connection program, but also be used to temporarily store data that has been output or is to be output.

[0147] The communication bus 32 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 31 and at least one processor 30, etc.

[0148] The communication interface 33 is used for communication between the above-mentioned electronic device 1 and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface, a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface.

[0149] Figure 3 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0150] For example, although not shown, the electronic device 1 may also include a power source (such as a battery) for supplying power to each component. Preferably, the power source may be logically connected to the at least one processor 30 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, and power status indicators. The electronic device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0151] It should be understood that the described embodiments are for illustration purposes only.

[0152] The database configuration connection program stored in the memory 31 in the electronic device 1 is a combination of multiple computer programs. When running in the processor 30, it can achieve:

[0153] Obtaining a partition wall appearance image corresponding to an installed lightweight partition wall, marking an image label corresponding to the partition wall appearance image, and constructing an appearance image set corresponding to the partition wall appearance image based on the image label;

[0154] Performing image smoothing processing on the apparent image set to obtain a smoothed image set, identifying gray levels corresponding to images in the smoothed image set, querying the number of pixels corresponding to the gray levels, and calculating an apparent histogram corresponding to the apparent image set based on the gray levels and the number of pixels;

[0155] Based on the apparent histogram, calculating the cumulative distribution value corresponding to each gray level in the apparent histogram;

[0156] Based on the cumulative distribution value, the partition wall appearance image is segmented using a pre-built partition wall appearance quality inspection model to obtain an appearance segmentation area, the texture feature value corresponding to the appearance segmentation area is calculated, and the texture feature is clustered to obtain an appearance quality category corresponding to the partition wall appearance image;

[0157] Based on the apparent quality category, the apparent segmented areas are quality marked to obtain regional quality marks, and based on the regional quality marks, an apparent quality inspection report corresponding to the installed lightweight partition wall is generated.

[0158] Specifically, the specific implementation method of the processor 30 for the above computer program can refer to Figure 1 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0159] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium. The storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0160] The present invention further provides a storage medium, wherein the readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:

[0161] Obtaining a partition wall appearance image corresponding to an installed lightweight partition wall, marking an image label corresponding to the partition wall appearance image, and constructing an appearance image set corresponding to the partition wall appearance image based on the image label;

[0162] Performing image smoothing processing on the apparent image set to obtain a smoothed image set, identifying gray levels corresponding to images in the smoothed image set, querying the number of pixels corresponding to the gray levels, and calculating an apparent histogram corresponding to the apparent image set based on the gray levels and the number of pixels;

[0163] Based on the apparent histogram, calculating the cumulative distribution value corresponding to each gray level in the apparent histogram;

[0164] Based on the cumulative distribution value, the partition wall appearance image is segmented using a pre-built partition wall appearance quality inspection model to obtain an appearance segmentation area, the texture feature value corresponding to the appearance segmentation area is calculated, and the texture feature is clustered to obtain an appearance quality category corresponding to the partition wall appearance image;

[0165] Based on the apparent quality category, the apparent segmented areas are quality marked to obtain regional quality marks, and based on the regional quality marks, an apparent quality inspection report corresponding to the installed lightweight partition wall is generated.

[0166] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0167] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0168] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0169] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0170] Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is therefore intended that all changes that fall within the meaning and range of equivalent elements of the claims are embraced in the present invention.

[0171] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0172] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for detecting the apparent quality of lightweight partition walls based on deep learning, characterized in that: The method comprises: Obtaining a partition wall appearance image corresponding to an installed lightweight partition wall, marking an image label corresponding to the partition wall appearance image, and constructing an appearance image set corresponding to the partition wall appearance image based on the image label; Performing image smoothing processing on the apparent image set to obtain a smoothed image set, identifying gray levels corresponding to images in the smoothed image set, querying the number of pixels corresponding to the gray levels, and calculating an apparent histogram corresponding to the apparent image set based on the gray levels and the number of pixels; Based on the apparent histogram, calculating the cumulative distribution value corresponding to each gray level in the apparent histogram; Based on the cumulative distribution value, the partition wall appearance image is segmented using a pre-built partition wall appearance quality inspection model to obtain an appearance segmentation area, the texture feature value corresponding to the appearance segmentation area is calculated, and the texture feature is clustered to obtain an appearance quality category corresponding to the partition wall appearance image; Based on the apparent quality category, the apparent segmented areas are quality marked to obtain regional quality marks, and based on the regional quality marks, an apparent quality inspection report corresponding to the installed lightweight partition wall is generated.

2. The method for detecting the apparent quality of a lightweight partition wall based on deep learning as claimed in claim 1, characterized in that: The image tag corresponding to the partition wall appearance image includes: Performing image adjustment on the partition wall surface image to obtain an adjusted image; Extracting a partition wall region in the adjusted image; identifying partition wall features in the partition wall region; Marking a feature sequence corresponding to the partition wall feature; Based on the feature sequence, an image label corresponding to the partition wall appearance image is marked.

3. The method for detecting the apparent quality of a lightweight partition wall based on deep learning as claimed in claim 1, characterized in that: The step of constructing an apparent image set corresponding to the partition wall apparent image based on the image label includes: Annotate a label sequence corresponding to the image label; Based on the tag sequence, dividing the apparent image pairs corresponding to the partition wall apparent image; Extract the apparent features in the apparent image pair, and identify the feature vectors corresponding to the apparent features; and construct the apparent image set corresponding to the partition wall apparent image based on the feature vectors.

4. The method for detecting the apparent quality of a lightweight partition wall based on deep learning according to claim 1, characterized in that: The step of performing image smoothing on the apparent image set to obtain a smoothed image set includes: Converting each image in the apparent image set into a grayscale image; Performing image segmentation on the grayscale image to obtain small-block images; Calculating the local pixel average of the small block image; Taking the local pixel average as the central pixel value corresponding to the small block image; Based on the central pixel value, the small block images are combined into a smooth image; The smoothed images are stored to obtain a smoothed image set.

5. The method for detecting the apparent quality of a lightweight partition wall based on deep learning as claimed in claim 4, characterized in that: The calculating the local pixel average value of the small block image includes: The local pixel average value of the small image is calculated using the following formula: Among them, PJ represents the local pixel average value of the small image, m and n respectively represent the number of rows and columns corresponding to the small image, and I(i,j) represents the pixel value of the i-th row and j-th column in the small image.

6. The method for detecting the apparent quality of a lightweight partition wall based on deep learning according to claim 1, characterized in that: The calculating, based on the grayscale and the pixel quantity, an apparent histogram corresponding to the apparent image set comprises: The apparent histogram corresponding to the apparent image set is calculated using the following formula: Among them, C(gk) represents the number of occurrences of pixels in the apparent image set, gk represents the grayscale, n represents the pixel quantity, I represents the apparent image set, i represents the pixel index in the apparent image set, P(gk) represents the normalized frequency obtained by normalizing the number of occurrences of pixels of each grayscale level, and H(gk) represents the apparent histogram corresponding to the apparent image set.

7. The method for detecting the apparent quality of a lightweight partition wall based on deep learning according to claim 1, characterized in that: The calculating of the texture feature value corresponding to the apparent segmented area comprises: identifying the gray value and the average gray value corresponding to the apparent segmented area; First, the difference between the grayscale value and the average grayscale value is calculated using the following formula: D i (x,y)=E(x,y)-μ i Among them, D i (x, y) represents the difference value of the pixel (x, y) in the region i corresponding to the apparent segmentation region, E(x, y) represents the grayscale value of the pixel (x, y) in the image corresponding to the apparent segmentation region, μ i represents the average gray value; Secondly, the standard deviation corresponding to the difference value is calculated using the following formula: Among them, σ i represents the standard deviation corresponding to the difference value, W and H represent the width and height of the image corresponding to the apparent segmented area, respectively, and (x, y) represents the pixel position of the image corresponding to the apparent segmented area; Furthermore, the grayscale difference mean corresponding to the apparent segmented area is calculated using the following formula: Wherein, M represents the mean grayscale difference corresponding to the apparent segmented region, N represents the number of regions corresponding to the apparent segmented region, μ i represents the average gray value of the ith region, μ j represents the average gray value of the jth region; Finally, the following formula is used to calculate the corresponding texture feature value in the apparent segmentation area: Wherein, F represents the texture feature value corresponding to the apparent segmentation area, (σ i / (1+M)) means normalizing the standard deviation of each region.

8. The method for detecting the apparent quality of a lightweight partition wall based on deep learning as claimed in claim 1, characterized in that: The step of marking the apparent segmented regions based on the apparent quality categories to obtain regional quality marks includes: Identifying the quality standard corresponding to the apparent quality category; extracting regional characteristics in the apparent segmented region based on the quality criterion; Based on the regional characteristics, calculating a quality score corresponding to the apparent segmented region; Based on the quality score, the apparent segmented region is quality marked to obtain a region quality mark.

9. A system for detecting the apparent quality of lightweight partition walls based on deep learning, characterized in that: The system is used to perform the method for detecting the apparent quality of a lightweight partition wall based on deep learning as described in any one of claims 1 to 8, the system comprising: An image set construction module, used to obtain an apparent image of a partition wall corresponding to an installed lightweight partition wall, mark an image tag corresponding to the apparent image of the partition wall, and construct an apparent image set corresponding to the apparent image of the partition wall based on the image tag; a histogram calculation module, configured to perform image smoothing processing on the apparent image set to obtain a smoothed image set, identify gray levels corresponding to images in the smoothed image set, query the number of pixels corresponding to the gray levels, and calculate an apparent histogram corresponding to the apparent image set based on the gray levels and the number of pixels; A distribution value calculation module, used for calculating the cumulative distribution value corresponding to each gray level in the apparent histogram based on the apparent histogram; A feature clustering module, for performing image region segmentation on the partition wall appearance image based on the cumulative distribution value using a pre-built partition wall appearance quality inspection model to obtain an appearance segmentation region, calculating a texture feature value corresponding to the appearance segmentation region, and performing clustering processing on the texture feature to obtain an appearance quality category corresponding to the partition wall appearance image; A report generation module is used to mark the quality of the apparent segmented areas based on the apparent quality categories to obtain regional quality marks, and to generate an apparent quality inspection report corresponding to the installed lightweight partition wall based on the regional quality marks.

Images