A kind of building material waterproof detection device and method

Abstract

The present invention relates to the technical field of material waterproof performance testing, and specifically to a device and method for testing waterproof building materials. The present invention first obtains a connected domain in an image to be tested, and extracts the skeleton pixel points of each connected domain; obtains the humidity value of the unsprayed surface; further obtains the corrected texture possibility of each connected domain based on the distribution characteristics of the skeleton pixel points of each connected domain, combined with the similar characteristics of the overall extension characteristics of the skeleton pixel points between the connected domains; further obtains the corrected gradient value of each pixel point based on the corrected texture possibility, and obtains the water droplet contour, and finally obtains the waterproof performance parameters of the material to be tested in combination with the change in humidity value before and after spraying. The present invention analyzes the distribution of skeleton pixel points in the image and the similar characteristics of the overall extension, corrects the gradient value of the pixel point, accurately identifies the water droplet contour, and accurately and quantitatively evaluates the waterproof performance of the material in combination with the humidity dimension.

Description

Technical Field

[0001] The invention relates to the technical field of material waterproof performance testing, and in particular to a building material waterproof detection device and method. Background Art

[0002] Building waterproof materials are used to prevent water from penetrating and invading the building structure, so as to ensure the durability and service life of the building. These materials include coatings, membranes and sealants, which usually have good waterproof performance, weather resistance and adhesion, and can be applied to multiple parts such as roofs, walls, floors and underground structures to prevent damage to buildings caused by water, humidity and other environmental factors.

[0003] Existing testing of waterproof materials is usually qualitative testing, and the texture of the waterproof material itself will also affect the identification of water droplets, resulting in an inability to accurately evaluate the waterproof performance of the material, and thus unable to predict the long-term performance of the material in actual application and potential leakage risks, which is not conducive to ensuring the safety and reliability of the building structure. Summary of the invention

[0004] In order to solve the technical problem that the existing waterproof material detection is mostly qualitative detection and is limited by the influence of the material's own texture, the purpose of the present invention is to provide a building material waterproof detection device and method, and the technical scheme adopted is as follows:

[0005] A method for detecting waterproofness of building materials, the method comprising:

[0006] Obtain an image of the unsprayed surface of the material to be inspected after spraying; obtain the humidity value of the unsprayed surface before and after spraying; obtain a connected domain in the image to be inspected, and extract skeleton pixel points of each connected domain;

[0007] According to the distribution characteristics of the skeleton pixel points of each connected domain, the texture possibility of each connected domain is obtained; according to the overall extension characteristics of the skeleton pixel points of each connected domain, the skeleton extension angle of each connected domain is obtained; according to the similar characteristics of the skeleton extension angles of each connected domain and other connected domains, combined with the texture possibility, the corrected texture possibility of each connected domain is obtained; according to the gradient value of each pixel point in the image to be detected, combined with the corrected texture possibility of the connected domain to which it belongs, the corrected gradient value of each pixel point is obtained;

[0008] Based on the corrected gradient value of each pixel point, the water droplet contour of the image to be detected is obtained; according to the water droplet contour and the change in humidity value of the unsprayed surface before and after spraying, the waterproof performance parameters of the material to be detected are obtained.

[0009] Furthermore, the method for obtaining the texture possibility includes:

[0010] Obtain the number of skeleton pixel points in each of the connected domains; construct a two-dimensional coordinate system with any pixel point in the image to be detected as the origin; in any of the connected domains, take two adjacent skeleton pixel points as a group of skeleton pixel points, and obtain the angle between the connecting line of adjacent skeleton pixel points in all the skeleton pixel point groups and the positive direction of the horizontal axis of the two-dimensional coordinate system;

[0011] According to the similar features between the angles corresponding to all adjacent groups of skeleton pixel points in the connected domain and combined with the number of skeleton pixel points, the texture possibility corresponding to the connected domain is obtained; the similar features between the angles corresponding to all adjacent groups of skeleton pixel points and the number of skeleton pixel points are positively correlated with the texture possibility.

[0012] Furthermore, the method for obtaining the skeleton extension angle includes:

[0013] The skeleton extension angle corresponding to the connected domain is obtained at least according to the average angle of the angles corresponding to all the skeleton pixel point groups in the connected domain.

[0014] Furthermore, the method for obtaining the possibility of correcting the texture includes:

[0015] Select any of the connected domains as a target connected domain, and use the other connected domains as comparison connected domains; obtain a texture correction sub-coefficient between the target connected domain and each corresponding comparison connected domain based on similar features of the skeleton extension angle between the target connected domain and each corresponding comparison connected domain, combined with the texture possibility of each comparison connected domain; the similar features of the skeleton extension angle and the texture possibility are both positively correlated with the texture correction sub-coefficient;

[0016] The average value of the texture correction sub-coefficients between the target connected domain and all the corresponding comparison connected domains is used as the texture correction coefficient of the target connected domain; according to the texture correction coefficient and the texture possibility of the target connected domain, the corrected texture possibility of the target connected domain is obtained; the texture correction coefficient and the texture possibility are both positively correlated with the corrected texture possibility.

[0017] Furthermore, the method for obtaining the modified gradient value includes:

[0018] The modified texture possibility is negatively correlated and normalized as a gradient correction coefficient; the product of the gradient value of each pixel in the image to be detected and the gradient correction coefficient of the connected domain to which it belongs is used as the modified gradient value of each pixel.

[0019] Furthermore, the method for obtaining the water droplet contour includes:

[0020] Non-maximum suppression is performed based on the corrected gradient value and gradient direction of each pixel point, and then a preset double threshold detection method is used to obtain the water droplet contour of the image to be detected.

[0021] Furthermore, the method for obtaining the waterproof performance parameter includes:

[0022] The change in humidity of the unsprayed surface before and after spraying is used as a humidity change parameter; the water seepage parameter is obtained by taking the ratio of the area of ​​the water droplet contour region to the area of ​​the image to be detected;

[0023] The waterproof performance parameter of the material to be detected is obtained according to the humidity change parameter and the water seepage parameter corresponding to the material to be detected; the humidity change parameter and the water seepage parameter are both negatively correlated with the waterproof performance parameter.

[0024] Furthermore, the gradient value and gradient direction of the pixel point are obtained by using a Sobel operator.

[0025] Furthermore, the connected domain in the image to be detected is obtained by using a K-means clustering algorithm.

[0026] The present invention also proposes a building material waterproof detection device, comprising a device body, a spray unit, a humidity acquisition unit, an image acquisition unit and a data processing unit; wherein the spray unit is used to spray the spray surface of the material to be detected; the image acquisition unit is used to obtain the image to be detected of the unsprayed surface of the material to be detected after spraying; the humidity acquisition unit is used to obtain the humidity value of the unsprayed surface before and after spraying; the data processing unit is used to receive the image to be detected and the humidity value, and implement any one of the steps of the building material waterproof detection method.

[0027] The present invention has the following beneficial effects:

[0028] The present invention first obtains the image of the unsprayed surface of the material to be detected after spraying, providing an analysis basis; further obtains the humidity value of the unsprayed surface before and after spraying, providing more basis for subsequent evaluation of the waterproof performance of the material, and improving the reliability of the waterproof performance parameters; further, in order to avoid directly performing edge detection on the image to be detected to determine the edge of water drops, which is easily affected by the texture of the material to be detected itself, and some material textures are identified as water drop edges, affecting the accuracy of the waterproof performance test results of the material to be detected, a skeleton can be used to refine a connected area into a pixel width, which is helpful to extract more accurate features when performing pattern recognition and classification, obtain the connected domain in the image to be detected, and extract the skeleton pixel points of each connected domain; further use the texture of the waterproof material The invention analyzes the distribution characteristics of the skeleton pixels in each connected domain and the similar characteristics of the overall extension characteristics of the skeleton pixels between different connected domains, obtains the possibility of the corrected texture of each connected domain, and provides a basis for the subsequent correction of the gradient value of the pixel points to reduce the influence of the material texture during water drop recognition; further, according to the gradient value of each pixel point in the image to be detected, combined with the possibility of the corrected texture of the connected domain to which it belongs, obtains the corrected gradient value of each pixel point, and provides a basis for accurately identifying the contour of water drops; finally, based on the corrected gradient value of each pixel point, obtains the contour of water drops in the image to be detected, and then combines the change of the humidity value of the unsprayed surface before and after spraying to obtain the waterproof performance parameters of the material to be detected, and accurately and quantitatively evaluates the waterproof performance of the material. The invention analyzes the distribution of the skeleton pixels in the image and the similar characteristics of the overall extension, evaluates the possibility that each connected domain belongs to the material texture, corrects the gradient value of the pixel point, accurately identifies the contour of water drops, and accurately and quantitatively evaluates the waterproof performance of the material in combination with the humidity dimension. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 A schematic diagram of a building material waterproof detection device provided by an embodiment of the present invention;

[0031] Figure 2 A flowchart of a method for detecting waterproofness of building materials provided by one embodiment of the present invention;

[0032] Figure 3 An image to be detected provided by an embodiment of the present invention;

[0033] Figure 4 An edge detection image of an image to be detected is provided by an embodiment of the present invention.

[0034] The numbers in the figure are: 1. Sprayer; 2. Humidity sensor; 3. Camera; 4. Luminous base plate; 5. Material to be tested; 6. Spraying surface; 7. Unsprayed surface; 8. Spraying room; 9. Observation room. DETAILED DESCRIPTION

[0035] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a building material waterproof detection device and method proposed by the present invention, its specific implementation, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0036] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0037] The specific scheme of a building material waterproof detection device and method provided by the present invention is described in detail below with reference to the accompanying drawings.

[0038] The present invention provides a building material waterproof detection device, comprising a device body, a spray unit, a humidity collection unit, an image collection unit and a data processing unit.

[0039] See also Figure 1 , which shows a schematic diagram of a building material waterproof detection device provided by an embodiment of the present invention, Figure 1 The method comprises a sprayer 1, a humidity sensor 2, a camera 3, a light-emitting base plate 4, a material to be detected 5, a spray surface 6, an unsprayed surface 7, a spray chamber 8 and an observation chamber 9, wherein the spray chamber 8 is located above the observation chamber 9 and is separated by the material to be detected 5; the spray unit is installed in the spray chamber 8 for spraying the spray surface 6 of the material to be detected 5; the image acquisition unit is installed in the observation chamber 9 for obtaining an image to be detected of the unsprayed surface 7 of the material to be detected 5 after spraying; the humidity sensor 2 is installed in the observation chamber 9 for obtaining the humidity value of the unsprayed surface 7 before and after spraying.

[0040] Among them, the spray unit includes a sprayer 1 and a spray control module. The sprayer 1 is controlled by the spray control module. The spray control module can be programmed to adapt to different spray experiments. In order to reduce the image deformation of the material to be detected, two cameras 3 are installed in the image acquisition unit to capture images and then splice them, and a light-emitting base plate 4 is installed for fill light to obtain a clear image to be detected and provide a reliable analysis basis. At the same time, considering that the unsprayed surface 7 of the material to be detected 5 will affect the humidity in the observation room 9 when it seeps water, a humidity sensor is installed in the observation room 9 to obtain the humidity value of the unsprayed surface 7 before and after spraying. Finally, the data processing unit collects the experimental data, including the image to be detected and the humidity value, to realize a waterproof detection method for building materials.

[0041] It should be noted that image stitching technology is already an existing technology and will not be described in detail here; the humidity in the observation room 9 is collected by the humidity sensor 2 to represent the humidity of the unsprayed surface 7. In the spraying experiment, the observation room 9 and the spraying room 8 are separated by the material to be tested 5, and the observation room 9 is closed to ensure the accuracy of the humidity data; the camera 3 needs to have certain waterproof and moisture-proof properties. A protective shell with a pull ring can be placed at the camera lens. After the spraying is completed, the protective shell is removed by pulling the pull ring to collect the image, while keeping the observation room 9 closed and protecting the camera lens.

[0042] See also Figure 2 , which shows a flow chart of a method for detecting waterproofness of building materials provided by an embodiment of the present invention, specifically comprising:

[0043] Step S1: obtaining an image of the unsprayed surface of the material to be tested after spraying; obtaining the humidity value of the unsprayed surface before and after spraying; obtaining a connected domain in the image to be tested, and extracting skeleton pixel points of each connected domain.

[0044] First, obtain the image of the unsprayed surface 7 of the material 5 to be tested after spraying; obtain the humidity value of the unsprayed surface 7 before and after spraying to provide a data basis for subsequent analysis; please refer to Figure 3 , which shows an image to be detected provided by an embodiment of the present invention.

[0045] Considering that when the waterproof material seeps water, water droplets will seep out on the unsprayed surface 7, and considering that directly performing edge detection on the image to be detected to determine the edge of the water droplets is easily affected by the texture of the material to be detected 5 itself, please refer to Figure 4 , which shows an edge detection image of an image to be detected provided by an embodiment of the present invention, Figure 4In the process, while the water droplets are detected, some material textures are also identified as the edges of water droplets, which affects the accuracy of the test results of the waterproof performance of the material to be tested 5. Therefore, the connected domain in the image to be tested is extracted first, and the connected domain is analyzed more carefully, so as to finally accurately extract the water drop contour of the image to be tested; considering that the extracted skeleton can refine a connected area into a pixel width, thereby simplifying the shape representation while retaining the topological structure of the object, it is helpful to extract more accurate features during pattern recognition and classification, and eliminate the influence of the material's own texture, so the skeleton pixel points of each connected domain are extracted.

[0046] It should be noted that, in one embodiment of the present invention, the K-means clustering algorithm is used and the Process the image to be detected, and obtain the connected domain in the image to be detected according to the clustering result; the skeleton extraction technology of the connected domain is already an existing technology. For details, please refer to the Skeletonize function in the link https: / / www.cnblogs.com / denny402 / p / 5167414.html. Based on the connected domain in the image to be detected, the image to be detected is converted into a binary image, where the connected domain is 255 and the rest is 0, thereby extracting the skeleton and obtaining the skeleton pixel points. The K-means clustering algorithm is a technical means well known to those skilled in the art and will not be described in detail here.

[0047] It should be noted that the present invention is only applicable to opaque building waterproof materials, and the waterproof materials have regular and neatly arranged textures, that is, parallel grooves or stripes for guiding the flow of water; when designing the experiment, the spraying intensity or spraying time at least makes water droplets appear on the unsprayed surface 7 of the waterproof material.

[0048] Step S2: According to the distribution characteristics of the skeleton pixels of each connected domain, the texture possibility of each connected domain is obtained; according to the overall extension characteristics of the skeleton pixels of each connected domain, the skeleton extension angle of each connected domain is obtained; according to the similar characteristics of the skeleton extension angles of each connected domain and other connected domains, combined with the texture possibility, the corrected texture possibility of each connected domain is obtained; according to the gradient value of each pixel in the image to be detected, combined with the corrected texture possibility of the connected domain to which it belongs, the corrected gradient value of each pixel is obtained.

[0049] Considering the texture design of waterproof materials, which is usually used to help guide water flow, parallel grooves or stripes can effectively guide water flow into the drainage system, thereby reducing water retention and aggregation and reducing the waterproof pressure of the waterproof layer. Therefore, the texture of waterproof materials is relatively regular and arranged relatively neatly, so that the distribution of skeleton pixel points of the connected domain corresponding to the material texture is more regular. Therefore, according to the distribution characteristics of the skeleton pixel points of each connected domain, the texture possibility of each connected domain is obtained.

[0050] Preferably, in one embodiment of the present invention, considering that compared with the infiltrated water droplets, the material's own texture has strong continuity, a longer skeleton, and a greater number of skeleton pixels, the number of skeleton pixels in each connected domain is obtained, and the texture possibility of the connected domain is analyzed with the help of the number of skeleton pixels;

[0051] Considering that the material's own texture is more regular, the skeleton is straighter, and the corner of the line between the skeleton pixels is smaller, a two-dimensional coordinate system is constructed with any pixel in the image to be detected as the origin; in any connected domain, two adjacent skeleton pixels are used as a group of skeleton pixels, and the angle between the line connecting the adjacent skeleton pixels in all skeleton pixel groups and the positive direction of the horizontal axis is obtained;

[0052] According to the similar features between the angles corresponding to all adjacent skeleton pixel groups in the connected domain and the number of skeleton pixels, the texture possibility of the corresponding connected domain is obtained; the similar features between the angles corresponding to all adjacent skeleton pixel groups and the number of skeleton pixels are positively correlated with the texture possibility.

[0053] As an example, the formula for calculating texture likelihood includes:

[0054] ;

[0055] in, is the serial number of the connected domain; Indicates Texture possibility of connected domains; represents the linear normalization function; Indicates The number of skeleton pixels in a connected domain; Indicates The number of adjacent skeleton pixel groups in a connected domain; Indicates the serial number of adjacent skeleton pixel groups in the connected domain; Indicates For adjacent skeleton pixel groups, the angle corresponding to the first skeleton pixel group; Indicates For adjacent skeleton pixel groups, the angle corresponding to the second skeleton pixel group; Indicates taking the absolute value; Indicates preset division by zero positive parameter.

[0056] In the calculation formula of texture possibility, the length of the skeleton is represented by the number of skeleton pixels. The more skeleton pixels there are, the longer the skeleton in the connected domain is, the more consistent with the continuity of the material texture, and the greater the texture possibility is. The difference characteristics of the angle are represented by the absolute value of the difference, and then a negative correlation mapping is performed by taking the inverse, reflecting the similar characteristics between the angles corresponding to adjacent groups of skeleton pixels. The more similar the angles corresponding to adjacent groups of skeleton pixels are, the smaller the corner of the connecting line between the skeleton pixels is, the more likely the connected domain is to represent the material texture, and the greater the texture possibility is.

[0057] It should be noted that, for example, for four pixels with coordinates of A (20, 20), B (21, 20), C (22, 20) and D (23, 20), AB, BC and CD are each a group of skeleton pixels, AB and BC are a pair of adjacent skeleton pixel groups, and BC and CD are also a pair of adjacent skeleton pixel groups; when the skeleton has branches, for example, for four pixels with coordinates of A (20, 20), B (21, 20), C (22, 20) and D (21, 21), a branch appears at pixel point B (21, 20), then AB and BC, AB and BD are all adjacent skeleton pixel groups, and at this time, AB and BC, AB and BD all need to be calculated.

[0058] It should be noted that, in one embodiment of the present invention, ; The angle unit is radian; In other embodiments of the present invention, the implementer may also use degrees as the angle unit, or may use a negative correlation mapping function such as Negative correlation mapping is performed, where Expressed as a natural constant The exponential function with base ; represents the independent variable; the implementer can also use the characteristic parameters such as variance or standard deviation to represent the characteristics of drastic data fluctuations to obtain the texture possibility, such as obtaining the variance of the angle of all skeleton pixel groups in the connected domain, , Indicates The variance of the angles of all skeleton pixel groups in a connected domain.

[0059] Due to the influence of lighting and the shape of water drops, the arc of the shadow part of some water drops is small. After being identified as a separate connected domain during clustering, it may be mistakenly identified as a texture part, so it is necessary to further identify the texture characteristics based on the texture possibility; considering that the textures of the material itself are parallel to each other or close in angle, it is more conducive to guiding the water to flow out, so the skeleton extension angle of each connected domain is obtained according to the overall extension characteristics of the skeleton pixels of each connected domain; according to the similar characteristics of the skeleton extension angles of each connected domain and other connected domains, combined with the texture possibility, the corrected texture possibility of each connected domain is obtained; further based on the similar characteristics of the overall skeleton angles between the connected domains, the texture possibility is further corrected, and the texture attribution of the connected domain is more accurately evaluated, which is convenient for the subsequent correction of the gradient value of the pixel point to obtain the accurate water drop edge.

[0060] Preferably, in one embodiment of the present invention, the skeleton extension angle of the corresponding connected domain is obtained at least according to the average value of the angles corresponding to all skeleton pixel point groups in the connected domain.

[0061] As an example, the extension characteristics of the skeleton pixel points are represented by the angle between the connecting line between adjacent pixel points and the positive direction of the horizontal axis, and the overall extension characteristics of the skeleton pixel points of the connected domain are represented by the average value. The average value of the angles corresponding to all skeleton pixel point groups in the connected domain is taken as the skeleton extension angle of the corresponding connected domain.

[0062] In another embodiment of the present invention, the skeleton extension angle is obtained by the average, median and mode of the angles corresponding to all skeleton pixel point groups in the connected domain, for example, the weighted weights are 0.4, 0.3 and 0.3 respectively.

[0063] Preferably, in one embodiment of the present invention, considering that the more similar the skeleton extension angles of a certain connected domain are to other connected domains, the more obvious the parallel features shown by the connected domain and other connected domains are, and the more consistent with the material's own texture features are, and the greater the texture possibility of other connected domains is, the stronger the reference of the skeleton extension angles of other connected domains is. Based on this, any connected domain is selected as the target connected domain, and other connected domains are used as comparison connected domains; according to the similar features of the skeleton extension angles between the target connected domain and each corresponding comparison connected domain, combined with the texture possibility of each comparison connected domain, the texture correction sub-coefficient between the target connected domain and each corresponding comparison connected domain is obtained; the similar features of the skeleton extension angles and the texture possibility are both positively correlated with the texture correction sub-coefficient;

[0064] The average value of the texture correction sub-coefficients between the target connected domain and all corresponding comparison connected domains is taken as the texture correction coefficient of the target connected domain; according to the texture correction coefficient and texture possibility of the target connected domain, the corrected texture possibility of the target connected domain is obtained; the texture correction coefficient and texture possibility are both positively correlated with the corrected texture possibility.

[0065] As an example, the formula for calculating the modified texture likelihood includes:

[0066] ;

[0067] in, is the serial number of the connected domain; Indicates Texture possibility of connected domains; represents the linear normalization function; Indicates Corrected texture likelihood of connected domains; Represents the number of all connected domains in the image to be detected; Indicates When a connected domain is used as the target connected domain, the corresponding serial number of the comparison connected domain; Indicates The connected domain corresponds to The texture likelihood of the contrasting connected domains; Indicates The extension angle of the connected domain skeleton; Indicates The connected domain corresponds to The skeleton extension angle of the contrast connected domain; Indicates The connected domain and Texture modifier coefficients between contrasting connected domains; Indicates Texture correction coefficient of connected domains; Indicates taking the absolute value; Indicates the preset zero-dividing positive parameter, in this example .

[0068] In the calculation formula of the corrected texture possibility, the similar characteristics of the skeleton extension angles between the two connected domains are expressed by the absolute value of the difference, and then a negative correlation mapping is performed by taking the inverse; the smaller the absolute value of the difference in the skeleton extension angle between the target connected domain and the corresponding comparison connected domain, the more similar the skeleton extension angles are, and the more obvious the parallel characteristics of the skeletons of the connected domains are; at the same time, the greater the texture possibility of the comparison connected domain, the more likely the comparison connected domain is to be a texture area, the stronger the reference of the skeleton extension angle of the comparison connected domain, the more similar the skeleton extension angles between the target connected domain and the comparison connected domain with a greater texture possibility are, and the more likely the connected domain is to be a material texture.

[0069] In other embodiments of the present invention, implementers may also use a negative correlation mapping function such as Negative correlation mapping.

[0070] By analyzing the distribution of skeleton pixel points in the connected domain and the similar features of skeleton extensions between different connected domains, and obtaining the corrected texture possibility that characterizes each connected domain as a material texture, it is necessary to correct the gradient value of the pixel point in the image to be detected. According to the gradient value of each pixel point in the image to be detected and the corrected texture possibility of the connected domain to which it belongs, the corrected gradient value of each pixel point is obtained to avoid the subsequent identification of the pixel point of the material texture as the outline of water droplets, thereby improving the accuracy of the waterproof detection results of building materials.

[0071] Preferably, in one embodiment of the present invention, considering that the greater the possibility of the corrected texture of the connected domain is, the more likely it is that it belongs to the material texture, the more it is necessary to reduce the gradient value of the pixel points in the connected domain to reduce the possibility of the pixel points being identified as the edge of a water droplet. Therefore, the corrected texture possibility is negatively correlated mapped and normalized as a gradient correction coefficient; the product of the gradient value of each pixel point in the image to be detected and the gradient correction coefficient of the connected domain to which it belongs is used as the corrected gradient value of each pixel point.

[0072] As an example, the calculation formula for the modified gradient value includes: ;

[0073] in, Indicates the sequence number of the pixel in the image to be detected; Indicates The gradient value of each pixel; Indicates Corrected texture probability of the connected domain to which the pixel points belong; Indicates The gradient correction coefficient of each pixel.

[0074] In the calculation formula of the corrected gradient value, since the corrected texture probability is normalized, it can be calculated by In this way, the possibility of corrected texture is negatively mapped and normalized, so as to obtain the gradient correction coefficient to correct the gradient value of the pixel point.

[0075] Step S3: based on the corrected gradient value of each pixel point, obtain the water droplet contour of the image to be detected; according to the water droplet contour and the change in humidity value of the unsprayed surface before and after spraying, obtain the waterproof performance parameters of the material to be detected.

[0076] After correcting the gradient value of the pixel point and eliminating the influence of material texture, the water droplet outline of the image to be detected can be obtained based on the corrected gradient value of each pixel point, and the water droplet area can be accurately identified, providing a basis for obtaining waterproof performance parameters.

[0077] Preferably, in one embodiment of the present invention, non-maximum suppression is performed based on the corrected gradient value and gradient direction of each pixel, and then a preset dual-threshold detection method is used to obtain the water droplet contour of the image to be detected.

[0078] It should be noted that, in one embodiment of the present invention, the gradient value and gradient direction of the pixel point are obtained by the Sobel operator, the high threshold is obtained by the Otsu method, and half of the high threshold is used as the low threshold to obtain the water drop contour of the image to be detected; wherein, the Sobel operator, Otsu method and non-maximum suppression are all technical means well known to those skilled in the art, and specific reference may be made to the link https: / / blog.csdn.net / weixin_42491648 / article / details / 131348643. In this implementation scenario, the detected edge contour is the water drop contour, which will not be described in detail here.

[0079] In the embodiment of the present invention, considering that water seepage occurs on the unsprayed surface 7 of the material to be tested 5, the humidity of the observation chamber 9 will be affected. Therefore, the waterproof performance parameters of the material to be tested 5 are obtained according to the change of the water droplet contour and the humidity value of the unsprayed surface 7 before and after spraying. Combined with the humidity change angle, the waterproof performance of the material to be tested 5 is more accurately evaluated.

[0080] Preferably, in one embodiment of the present invention, the change in humidity value of the unsprayed surface 7 before and after spraying is used as a humidity change parameter; the water seepage parameter is obtained by taking the ratio of the area of ​​the water droplet contour region to the area of ​​the image to be detected;

[0081] The waterproof performance parameter of the material to be detected 5 is obtained according to the humidity change parameter and the water seepage parameter corresponding to the material to be detected 5; the humidity change parameter and the water seepage parameter are both negatively correlated with the waterproof performance parameter.

[0082] As an example, the calculation formula for the waterproof performance parameter includes:

[0083] ;

[0084] in, Indicates the waterproof performance parameter of the material 5 to be tested; Represents the area of ​​all water droplet contour regions in the image to be detected; Represents the area of ​​the image to be detected; Indicates the humidity value of the unsprayed surface 7 before spraying; Indicates the humidity value of the unsprayed surface 7 after spraying; Indicates water seepage parameters; Indicates humidity change parameter; represents the hyperbolic tangent function.

[0085] In the calculation formula of the waterproof performance parameter, the area of ​​the water droplet contour area or the area of ​​the image to be detected is represented by the number of pixels; the larger the area of ​​all water droplet contour areas, the greater the proportion of the image to be detected, indicating that more water droplets have seeped out of the unsprayed surface 7, the worse the waterproof performance, and the smaller the waterproof performance parameter; in order to analyze the humidity change of the unsprayed surface 7 caused by water seepage, it is necessary to keep the observation room 9 airtight during the experiment. At this time, water seepage will only cause the humidity value of the observation room 9 to rise, so it is directly used Indicates the change in humidity value, and uses The function adjusts the range to The way of reflecting the change of humidity value is The larger it is, the greater the increase in humidity, the greater the water seepage, the weaker the waterproof performance, and the smaller the waterproof performance parameter; the waterproof performance parameter is used to quantitatively detect the waterproof performance of the material 5 to be tested.

[0086] In another embodiment of the present invention, the implementer may also Function sets the sensitivity coefficient, such as ;in is a positive number, The bigger, The more sensitive the function is to the independent variable, The closer the function is to 1, the easier it is to adapt to the humidity units of different humidity sensors, and the implementer can set it by himself.

[0087] In other embodiments of the present invention, considering that the water droplets on the unsprayed surface 7 may drip, resulting in the water droplets remaining on the unsprayed surface 7 not fully representing the water seepage of the material to be detected 5, a pressure sensor is installed on the light-emitting bottom plate 4 to obtain the pressure sensor readings before and after spraying, and the absolute value of the difference between the pressure sensor readings is multiplied by the humidity change parameter and the water seepage parameter, and then the negative correlation mapping function is used as follows: Negative correlation mapping is performed and normalized, and the mapping results are used as waterproof performance parameters.

[0088] After obtaining the waterproof performance parameters, a test report can be generated together with the spray parameters in the spray unit such as spray intensity parameters, spray duration parameters and preparation data of the test material such as composition data to provide data support for the improvement of waterproof materials, avoid the application of unqualified waterproof materials in actual projects, or promptly formulate corresponding maintenance plans for projects with waterproof materials that do not meet performance standards to ensure waterproof performance.

[0089] In summary, in view of the technical problem that the existing waterproof material detection is mostly qualitative detection and is limited by the influence of the material's own texture, the present invention proposes a building material waterproof detection device and method. The present invention first obtains the connected domain in the image to be detected, and extracts the skeleton pixel points of each connected domain; obtains the humidity value of the unsprayed surface; further, according to the distribution characteristics of the skeleton pixel points of each connected domain, combined with the similar characteristics of the overall extension characteristics of the skeleton pixel points between the connected domains, obtains the possibility of the corrected texture of each connected domain; further, according to the possibility of the corrected texture of the connected domain, obtains the corrected gradient value of each pixel point; further, based on the corrected gradient value of each pixel point, obtains the water droplet contour, and combines the change of the humidity value before and after spraying to obtain the waterproof performance parameters of the material to be detected. The present invention analyzes the distribution of the skeleton pixel points in the image and the similar characteristics of the overall extension, evaluates the possibility that each connected domain belongs to the material texture, corrects the gradient value of the pixel point, accurately identifies the water droplet contour, and accurately and quantitatively evaluates the waterproof performance of the material in combination with the humidity dimension.

[0090] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0091] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for detecting waterproofness of building materials, characterized in that: The method comprises: Obtain an image of the unsprayed surface of the material to be inspected after spraying; obtain the humidity value of the unsprayed surface before and after spraying; obtain a connected domain in the image to be inspected, and extract skeleton pixel points of each connected domain; According to the distribution characteristics of the skeleton pixel points of each connected domain, the texture possibility of each connected domain is obtained; according to the overall extension characteristics of the skeleton pixel points of each connected domain, the skeleton extension angle of each connected domain is obtained; according to the similar characteristics of the skeleton extension angles of each connected domain and other connected domains, combined with the texture possibility, the corrected texture possibility of each connected domain is obtained; according to the gradient value of each pixel point in the image to be detected, combined with the corrected texture possibility of the connected domain to which it belongs, the corrected gradient value of each pixel point is obtained; Based on the corrected gradient value of each pixel point, the water droplet contour of the image to be detected is obtained; according to the water droplet contour and the change of the humidity value of the non-sprayed surface before and after spraying, the waterproof performance parameter of the material to be detected is obtained; The method for obtaining the texture possibility includes: Obtain the number of skeleton pixel points in each of the connected domains; construct a two-dimensional coordinate system with any pixel point in the image to be detected as the origin; in any of the connected domains, take two adjacent skeleton pixel points as a group of skeleton pixel points, and obtain the angle between the connecting line of adjacent skeleton pixel points in all skeleton pixel point groups and the positive direction of the horizontal axis of the two-dimensional coordinate system; According to the similarity features between the angles corresponding to all adjacent skeleton pixel groups in the connected domain, combined with the number of skeleton pixel points, the texture possibility corresponding to the connected domain is obtained; the similarity features between the angles corresponding to all adjacent skeleton pixel groups and the number of skeleton pixel points are both positively correlated with the texture possibility; The formula for calculating texture likelihood includes: ； in, is the serial number of the connected domain; Indicates Texture possibility of connected domains; represents the linear normalization function; Indicates The number of skeleton pixels in a connected domain; Indicates The number of adjacent skeleton pixel groups in a connected domain; Indicates the serial number of adjacent skeleton pixel groups in the connected domain; Indicates For adjacent skeleton pixel groups, the angle corresponding to the first skeleton pixel group; Indicates For adjacent skeleton pixel groups, the angle corresponding to the second skeleton pixel group; Indicates taking the absolute value; Indicates preset division by zero positive parameter; The extension feature of the skeleton pixel point is represented by the angle between the connecting line between adjacent pixels and the positive direction of the horizontal axis. The overall extension feature of the skeleton pixel point of the connected domain is represented by the average value. The average value of the angles corresponding to all skeleton pixel point groups in the connected domain is taken as the skeleton extension angle of the corresponding connected domain. The calculation formula for the modified texture probability includes: ; in, is the serial number of the connected domain; Indicates Texture possibility of connected domains; represents the linear normalization function; Indicates Corrected texture likelihood of connected domains; Represents the number of all connected domains in the image to be detected; Indicates When a connected domain is used as the target connected domain, the corresponding serial number of the comparison connected domain; Indicates The connected domain corresponds to The texture likelihood of the contrasting connected domains; Indicates The extension angle of the connected domain skeleton; Indicates The connected domain corresponds to The skeleton extension angle of the contrast connected domain; Indicates The connected domain and Texture modifier coefficients between contrasting connected domains; Indicates Texture correction coefficient of connected domains; Indicates taking the absolute value; Indicates the preset division by zero positive parameter, .

2. A method for detecting waterproofness of building materials according to claim 1, characterized in that: The method for obtaining the modified gradient value comprises: The modified texture possibility is negatively correlated and normalized as a gradient correction coefficient; the product of the gradient value of each pixel in the image to be detected and the gradient correction coefficient of the connected domain to which it belongs is used as the modified gradient value of each pixel.

3. A method for detecting waterproofness of building materials according to claim 1, characterized in that: The method for obtaining the water droplet contour comprises: Non-maximum suppression is performed based on the corrected gradient value and gradient direction of each pixel point, and then a preset double threshold detection method is used to obtain the water droplet contour of the image to be detected.

4. A method for detecting waterproofness of building materials according to claim 1, characterized in that: The method for obtaining the waterproof performance parameter includes: The change in humidity of the unsprayed surface before and after spraying is used as a humidity change parameter; the water seepage parameter is obtained by taking the ratio of the area of ​​the water droplet contour region to the area of ​​the image to be detected; The waterproof performance parameter of the material to be detected is obtained according to the humidity change parameter and the water seepage parameter corresponding to the material to be detected; the humidity change parameter and the water seepage parameter are both negatively correlated with the waterproof performance parameter.

5. A method for detecting waterproofness of building materials according to claim 3, characterized in that: The gradient value and gradient direction of the pixel point are obtained through the Sobel operator.

6. A method for detecting waterproofness of building materials according to claim 1, characterized in that: The connected domain in the image to be detected is obtained by using a K-means clustering algorithm.

7. A building material waterproof detection device, comprising a device body, characterized in that: It also includes a spray unit, a humidity acquisition unit, an image acquisition unit and a data processing unit; wherein the spray unit is used to spray the spray surface of the material to be detected; the image acquisition unit is used to obtain the image to be detected of the unsprayed surface of the material to be detected after spraying; the humidity acquisition unit is used to obtain the humidity value of the unsprayed surface before and after spraying; the data processing unit is used to receive the image to be detected and the humidity value, and implement the steps of a building material waterproof detection method as described in any one of claims 1 to 6.

Images