A method of evaluating concrete formwork for civil engineering
By acquiring multiple images of the concrete surface at different times and using thresholds for directional differences and morphological irregularities for screening and judgment, the problem of low accuracy in brush-textured concrete identification in existing technologies has been solved, and accurate identification and quality assessment of cracks have been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING ZIHUAI TECHNOLOGY CO LTD
- Filing Date
- 2026-04-20
- Publication Date
- 2026-07-14
AI Technical Summary
In existing technologies, automatic detection algorithms often misidentify normal brush marks as cracks on brush-textured concrete surfaces, resulting in low accuracy in crack identification.
By acquiring multiple images of the brush-textured concrete surface at different times, the directional differences and morphological irregularities of each contour region are determined. Combined with the differences in time variation, cracked areas and suspected cracked areas are divided. The directional differences and morphological irregularity thresholds are used for screening and judgment to accurately identify cracked areas.
It significantly improves the accuracy of crack identification in complex backgrounds, can accurately assess the forming quality of brush-textured concrete, overcomes the interference of brush texture, and improves the accuracy of identification and the reliability of assessment.
Smart Images
Figure CN122391152A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of image processing technology, specifically to a method for evaluating the molding of concrete in civil engineering projects. Background Technology
[0002] In the field of civil engineering construction and quality control, accurate assessment of the forming quality of concrete structures is a crucial step in ensuring project safety and durability. This is especially true for brushed concrete surfaces, which are prone to surface cracks due to shrinkage and other factors during the hardening process. If these minute defects are not detected and addressed promptly, they may pose long-term risks to structural performance.
[0003] In related technologies, automatic detection algorithms such as edge detection and threshold segmentation are commonly used to achieve automatic identification and evaluation of cracks.
[0004] However, the texture inherent in brushed concrete is quite similar to that of cracks, creating strong background interference. Based on the aforementioned automatic detection algorithm, normal brush marks may be incorrectly identified as cracks, resulting in low accuracy in crack identification. Summary of the Invention
[0005] To address the technical problem that automatic detection algorithms may incorrectly identify normal brush marks as cracks, resulting in low accuracy in crack identification, this application aims to provide a method for evaluating the forming of concrete in civil engineering projects. The specific technical solution adopted is as follows: This application provides a method for evaluating the forming quality of concrete for civil engineering projects. The method includes: acquiring multiple images of the brush-textured concrete surface at different times; determining the directional differences of each contour region in each image, and dividing the contour regions of the multiple images into crack regions and suspected crack regions based on the directional differences, wherein the directional differences are used to characterize the difference between the direction of the contour region and the direction of other contour regions in the corresponding image; determining the variation differences and morphological irregularities of each suspected crack region, wherein the variation differences are used to characterize the degree of change of the contour region at different times; dividing the suspected crack regions into crack regions and non-crack regions based on the morphological irregularities and variation differences of each suspected crack region; and evaluating the forming quality of the brush-textured concrete based on the number of crack regions in the multiple images.
[0006] Optionally, determining the directional difference of each contour region in each image includes: determining the directional parameter of each contour region in a first image, the directional parameter being used to characterize the dominant direction of the contour region in its respective image, the first image being any one of the plurality of images; determining the directional parameter difference of each other contour region in the first image based on the directional parameter of each contour region, the directional difference of one other contour region being the difference between the directional parameter of the other contour region and the directional parameter of the first contour region, the first contour region being any contour region in the first image; and determining the directional difference of the first contour region based on the directional parameter difference of each other contour region.
[0007] Optionally, determining the directional difference of the first contour region based on the directional parameter difference of each other contour region includes: determining the number of reference regions of the first contour region based on the directional parameter difference of each other contour region; determining the reference weight of each other contour region based on the difference between the number of reference regions of each other contour region and the number of reference regions of the first contour region; and determining the directional difference of the first contour region based on the reference weight of each other contour region and the directional parameter difference of each other contour region.
[0008] Optionally, determining the morphological irregularity of each suspected crack region includes: determining the width non-uniformity, edge roughness, and contour irregularity of a first suspected crack region, wherein the first suspected crack region is any suspected crack region; and performing a weighted summation of the width non-uniformity, edge roughness, and contour irregularity of the first suspected crack region to obtain the morphological irregularity of the first suspected crack region.
[0009] Optionally, the method further includes: obtaining a width sequence of a first suspected crack region, the width sequence including multiple width values; and determining the width non-uniformity of the first suspected crack region based on the difference between each width value in the width sequence of the first suspected crack region and the average width value.
[0010] Optionally, the method further includes: obtaining the gradient magnitude of the edge pixels of the first suspected crack region; and determining the edge roughness of the first suspected crack region based on the variance of the gradient magnitude of the edge pixels of the first suspected crack region.
[0011] Optionally, the method further includes: obtaining a contour chain code sequence of a first suspected crack region, the contour chain code sequence including multiple chain code values; determining an absolute value sequence of the first-order difference of the contour chain code sequence; and determining the contour irregularity of the first suspected crack region based on the average value of the absolute values in the absolute value sequence.
[0012] Optionally, determining the variation difference of each suspected crack region includes: aligning the multiple images to obtain the correspondence between suspected crack regions in different images; determining the corresponding suspected crack region of the first suspected crack region based on the correspondence; and determining the variation difference of the first suspected crack region based on the area change of the first suspected crack region in different images.
[0013] Optionally, the method of dividing suspected crack regions into crack regions and non-crack regions based on the morphological irregularity and variation differences of each suspected crack region includes: determining the crack region probability of the first suspected crack region by multiplying the morphological irregularity and variation differences of the first suspected crack region; determining the first suspected crack region as a crack region if the crack region probability of the first suspected crack region is greater than or equal to the crack region probability threshold; and determining the first suspected crack region as a non-crack region if the crack region probability of the first suspected crack region is less than the crack region probability threshold.
[0014] Optionally, based on the directional difference, the contour regions of the multiple images are divided into crack regions and suspected crack regions, including: if the directional difference of the first contour region is greater than or equal to the directional difference threshold, the first contour region is determined as a crack region; if the directional difference of the first contour region is less than the directional difference threshold, the first contour region is determined as a suspected crack region.
[0015] This application has the following beneficial effects: In this embodiment, the crack regions are quickly identified by the difference between the direction of the contour region and the direction of other contour regions in the image. Then, the suspected crack regions that are difficult to distinguish are accurately determined by combining the irregular shape of the suspected crack regions and the degree of change over time. The crack regions in the suspected crack regions are further analyzed. Based on the characteristics of cracks, the crack regions can be accurately identified, effectively overcoming the interference of brush texture and significantly improving the accuracy of crack identification in complex backgrounds. Based on the number of identified crack regions, the forming quality of brush-textured concrete can be accurately evaluated. Attached Figure Description
[0016] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0017] Figure 1A flowchart of a method for evaluating the molding of concrete for civil engineering, provided as an embodiment of this application; Figure 2 A flowchart of another method for evaluating concrete forming in civil engineering provided in one embodiment of this application; Figure 3 A flowchart of another method for evaluating concrete forming in civil engineering provided in one embodiment of this application; Figure 4 This is a flowchart illustrating another method for evaluating the molding of concrete for civil engineering projects, provided as an embodiment of this application. Detailed Implementation
[0018] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a concrete forming evaluation method for civil engineering projects proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0019] Unless otherwise defined, 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 application pertains.
[0020] The following description, in conjunction with the accompanying drawings, details a specific scheme for the concrete forming evaluation method for civil engineering provided in this application.
[0021] Please see Figure 1 The diagram illustrates a flowchart of a method for evaluating the molding of concrete for civil engineering, provided in one embodiment of this application.
[0022] like Figure 1 As shown, the concrete forming evaluation method for civil engineering includes S101-S105.
[0023] S101. Obtain multiple images of the brush-textured concrete surface at different times.
[0024] In one alternative implementation, a high-resolution camera can be used to photograph the brushed concrete every 10 minutes to obtain the multiple images.
[0025] Optionally, the multiple images can be images taken from the same location and from the same angle.
[0026] It should be noted that these multiple images refer to at least two images.
[0027] Optionally, the multiple images can be subjected to mean filtering to eliminate noise in the multiple images.
[0028] Optionally, the grayscale value of each pixel in each image can also be obtained.
[0029] S102. Determine the directional differences of each contour region in each image, and divide the contour regions of the multiple images into crack regions and suspected crack regions based on the directional differences.
[0030] Among them, orientation difference is used to characterize the difference between the orientation of the contour region and the orientation of other contour regions in the image.
[0031] It should be understood that a contour region is a closed area in an image consisting of consecutive pixels with the same or similar pixel values (such as grayscale values or color values).
[0032] For example, a contour area can be a crack, a brush mark, or a portion of a crack or a brush mark.
[0033] In one alternative implementation, the Sobel edge detection algorithm can be used to detect edges in each image based on the grayscale values of the pixels in each image, thereby obtaining the edge pixels in each image. Then, based on the edge pixels in each image, a contour extraction algorithm can be used to extract the contours in each image, thereby obtaining the contour regions in each image.
[0034] It should be noted that there are usually multiple contour regions in an image. When fewer than two contour regions are extracted from an image, it can be directly determined that there are no significant textures or cracks in the image.
[0035] It should be understood that the direction of the brush marks on the brush-textured concrete surface is basically consistent, while the direction of the cracks is irregular. Therefore, based on the directional differences of each contour area, the contour area can be initially divided into crack areas and suspected crack areas.
[0036] It is understandable that the number of brush marks in an image should be much greater than the number of cracks. Therefore, the dominant direction in the image can represent the direction of the brush marks. The difference between the direction of a contour region and the direction of other contour regions is actually the difference between the direction of the contour region and the dominant direction of the image.
[0037] In one alternative implementation, the geometric direction of each contour region can be determined, the angular range of the directions of most contour regions can be statistically analyzed, the angular range can be determined as the dominant direction of the image, and the angle between the geometric direction of a contour region and the dominant direction can be determined as the directional difference of the contour region.
[0038] In one implementation of this application, the contour region can be divided into crack region and suspected crack region based on the directional difference threshold.
[0039] Specifically, taking the first contour region as an example, the first contour region is any contour region in the first image. If the directional difference of the first contour region is greater than or equal to the directional difference threshold, the first contour region is identified as a crack region; if the directional difference of the first contour region is less than the directional difference threshold, the first contour region is identified as a suspected crack region.
[0040] It should be understood that the first image is any one of the plurality of images.
[0041] It is understandable that if the directional difference of the first contour region is greater than or equal to the directional difference threshold, it indicates that the direction of the first contour region is significantly different from the brush direction of the first image. In this case, it can be determined that the first contour region contains cracks. The cracks cause the direction of the first contour region to be significantly different from the direction of other contour regions. Therefore, the first contour region can be identified as a crack region.
[0042] If the directional difference of the first contour region is less than the directional difference threshold, it means that the difference between the direction of the first contour region and the direction of the brush pattern in the first image is small. There may be cracks in the first contour region, or it may be a change in the direction of the brush pattern caused by human intervention. Therefore, it is necessary to further determine whether the first contour region is a crack region. At this time, the first contour region can be identified as a suspected crack region.
[0043] For example, the directional difference threshold can be 0.6.
[0044] Understandably, identifying a portion of the crack region based on directional differences, and then analyzing only the suspected crack regions, can reduce computational load and improve image analysis efficiency while ensuring the accuracy of crack region identification.
[0045] S103. Determine the morphological irregularities and variations of each suspected crack region.
[0046] Among them, the variation difference is used to characterize the degree of change of the contour region at different times.
[0047] It should be understood that cracks are usually irregular in shape, while brush marks are usually regular in shape. Therefore, one can determine whether a suspected crack area is a crack area by assessing whether the morphology of the suspected crack area is regular.
[0048] In one alternative implementation, the eccentricity of the suspected crack region can be determined as morphological irregularity, or the contrast obtained based on the gray-level co-occurrence matrix can be determined as morphological irregularity.
[0049] It is understandable that during the hardening and molding process of brush-textured concrete, the cracks will gradually extend and enlarge outward. Therefore, based on the changes in the suspected crack area in images at different times, such as whether it enlarges or deforms, it can be determined whether the suspected crack area includes a crack, that is, whether it is a crack area.
[0050] In one alternative implementation, the contour similarity of a suspected crack region in different images can be used to determine the variation of the suspected crack region.
[0051] Alternatively, the contour similarity between two contour regions can be determined by calculating the distance between their Hu moment vectors; the smaller the distance, the more similar the contour shapes.
[0052] S104. Based on the irregularity and variation of the morphology of each suspected crack region, the suspected crack region is divided into crack region and non-crack region.
[0053] In one alternative implementation, a morphological irregularity threshold and a variation difference threshold can be set separately. Then, suspected crack regions with morphological irregularity greater than or equal to the morphological irregularity threshold and variation difference greater than or equal to the variation difference threshold are identified as crack regions. Suspected crack regions with morphological irregularity less than the morphological irregularity threshold and / or variation difference less than the variation difference threshold are identified as non-crack regions.
[0054] For example, when normalizing the morphological irregularity and the variation difference respectively, the morphological irregularity threshold can be set to 0.7 and the variation difference threshold can be set to 0.7.
[0055] In one implementation of this application, the product of the irregularity and variation of the morphology of the first suspected crack region can be used to determine the crack region probability of the first suspected crack region. If the crack region probability of the first suspected crack region is greater than or equal to the crack region probability threshold, the first suspected crack region is determined as a crack region; if the crack region probability of the first suspected crack region is less than the crack region probability threshold, the first suspected crack region is determined as a non-crack region.
[0056] The first suspected crack region is any one of the suspected crack regions.
[0057] Optionally, the product of the irregularity and variation of the morphology of multiple suspected crack regions can be normalized.
[0058] For example, the normalization process can be maximum and minimum value normalization.
[0059] For example, after normalizing the product of the irregularity and variation of the morphology of multiple suspected crack regions, the probability threshold of the crack region can be adjusted according to the actual engineering requirements for crack sensitivity. For example, it can be set to 0.7 in cases with high requirements and 0.5 in cases with low requirements.
[0060] It should be understood that by fusing the evidence of irregularity in form and variability in change—two separate evidences from static and dynamic dimensions—through a product approach, the resulting probability of a crack region can comprehensively consider both the immediate morphology and historical evolution of the region, ensuring the comprehensiveness and high reliability of the final determination.
[0061] S105. Evaluate the forming quality of brush-textured concrete based on the number of cracked regions in multiple images.
[0062] It should be understood that the more cracked areas there are in the multiple images, the more cracks the brush-textured concrete actually produces, and the worse the molding quality of the brush-textured concrete.
[0063] In one alternative implementation, the forming quality of the brush-textured concrete can be determined to be substandard if cracked areas are present in multiple images.
[0064] In another alternative implementation, the average number of cracked regions in each image can be determined, and then the forming quality of the brush-textured concrete can be determined based on the correspondence between quantity and quality.
[0065] For example, the quantity-quality correspondence can include three correspondences: when the quantity is 0, the corresponding quality is excellent; when the quantity is greater than 0 and less than or equal to 1, the corresponding quality is acceptable; and when the quantity is greater than 1, the corresponding quality is poor.
[0066] In this embodiment, the crack regions are quickly identified by the difference between the direction of the contour region and the direction of other contour regions in the image. Then, the suspected crack regions that are difficult to distinguish are accurately determined by combining the irregular shape of the suspected crack regions and the degree of change over time. The crack regions in the suspected crack regions are further analyzed. Based on the characteristics of cracks, the crack regions can be accurately identified, effectively overcoming the interference of brush texture and significantly improving the accuracy of crack identification in complex backgrounds. Based on the number of identified crack regions, the forming quality of brush-textured concrete can be accurately evaluated.
[0067] Combination Figure 1 ,like Figure 2As shown, in one implementation of this application embodiment, the determination of the directional differences of each contour region in each image can be specifically implemented through S201-S203.
[0068] S201. Determine the orientation parameters of each contour region in the first image.
[0069] The orientation parameter is used to characterize the dominant orientation of the contour region in the image to which it belongs, and the first image is any one of the plurality of images.
[0070] In one implementation, the minimum bounding rectangle of each contour region can be obtained based on the minimum bounding rectangle algorithm. Then, the midpoints of the two minimum sides of the minimum bounding rectangle of each contour region are connected to obtain a straight line of each contour region, which represents the dominant direction of the contour region.
[0071] Understandably, after determining the straight line of the contour region, since the straight line does not have angular characteristics, a reference direction can be selected, and the dominant direction can be quantified based on the reference direction.
[0072] In one alternative implementation, the average of the minimum angles between the straight lines of each contour region and the straight lines of all other contour regions can be calculated. The straight line of the contour region with the smallest average angle is determined as the main direction line of the image. The main direction line is used to characterize the dominant direction of most contour regions in the image, which can also be understood as the brush pattern direction in the image. The dominant direction of the image is determined as the reference direction.
[0073] For example, for a straight line in contour region 1 in image 1, calculate the angle between it and the straight lines in each other contour region in image 1, and take the smallest angle value. Repeat this process for all contour regions in image 1 to obtain a set of minimum angle values, that is, the minimum angle value of each contour region. The arithmetic mean of this set of minimum angle values is determined as the average angle of the straight lines in contour region 1.
[0074] Next, using the main direction line as a reference, a rectangular coordinate system is set up, with the y-axis of the rectangular coordinate system set to be parallel to the main direction line, and all pixels in the image are placed in the first quadrant of the rectangular coordinate system. In this rectangular coordinate system, the direction parameters of each contour region are calculated.
[0075] It should be noted that the origin of this rectangular coordinate system is not subject to specific constraints and can be determined based on the content of the image.
[0076] Understandably, in brush-textured concrete images, since most brush marks are in the same direction, keeping the y-axis of the Cartesian coordinate system parallel to the direction of most brush marks makes it easier to calculate the directional differences of each contour region, making the measurement of directional differences clearer.
[0077] Alternatively, the orientation parameters of a contour region can be determined by the following formula: in, Image of brushed concrete The Middle The orientation parameters of each contour region Image of brushed concrete The Middle The ordinate of the first reference point on the straight line of the contour region. Image of brushed concrete The Middle The ordinate of the second reference point on the straight line of the contour region. Image of brushed concrete The Middle The x-coordinate of the first reference point on the straight line of the contour region. Image of brushed concrete The Middle The x-coordinate of the second reference point on the straight line of the contour region.
[0078] In the above formula, the first reference point and the second reference point are the... Any two points on a straight line within a contour region Indicates the first The slope of the straight line in each contour region represents the angle.
[0079] S202. Determine the difference in orientation parameters for each other contour region in the first image based on the orientation parameters of each contour region.
[0080] Wherein, the direction difference of another contour region is the difference between the direction parameter of the other contour region and the direction parameter of the first contour region, and the first contour region is any contour region in the first image.
[0081] It should be understood that the difference in orientation parameters for each other contour region is the difference in orientation parameters for the first contour region when analyzing the first contour region.
[0082] Optionally, the orientation parameters of each other contour region can be compared with the orientation parameters of the first contour region, and the absolute value of the difference can be determined as the orientation parameter difference of each other contour region.
[0083] S203. Based on the difference in orientation parameters of each other contour region, determine the orientation difference of the first contour region.
[0084] In one alternative implementation, the directional difference of the first contour region can be determined by the arithmetic mean of the differences in the directional parameters of the plurality of other contour regions.
[0085] It should be understood that, during the brush pattern manufacturing process, some brush patterns may have directions inconsistent with the majority. Directly using the difference in direction parameters between each other contour region to calculate the directionality of each contour region is inaccurate. Therefore, the reference value of the difference in direction parameters between each contour region should be considered. Specifically, when calculating the difference in direction parameters between two contour regions, the number of contour regions with similar direction parameters is first calculated. The smaller the difference in the number of similar contour regions, the less reference value is used in calculating the difference in direction parameters. Based on these characteristics, the directionality of each contour region in each brush-patterned concrete image is calculated.
[0086] In another alternative implementation, the number of reference regions of the first contour region can be determined based on the difference in orientation parameters of each other contour region; the reference weight of each other contour region can be determined based on the difference between the number of reference regions of each other contour region and the number of reference regions of the first contour region; and the orientation difference of the first contour region can be determined based on the difference between the reference weight of each other contour region and the difference in orientation parameters of each other contour region.
[0087] It should be understood that the reference area for the first contour region is the contour region with a smaller difference in directional parameters compared to the first contour region.
[0088] Optionally, a direction parameter difference threshold can be set, and other contour regions with direction parameter differences less than or equal to the direction parameter difference threshold can be determined as reference regions of the first contour region.
[0089] Optionally, the number of reference regions for each contour region in the first image can be determined based on this method.
[0090] For example, by combining the mathematical characteristics of the 8-direction chain code system, the threshold for the difference of the direction parameter can be set based on experience, such as 3, which can both accommodate normal changes in the brush pattern and identify true direction anomalies.
[0091] It should be understood that the more reference regions there are, the more regions are similar to the directional parameters of the first contour region, and the more prevalent the dominant direction of the first contour region is. The fewer reference regions there are, the less prevalent the dominant direction of the first contour region is.
[0092] Optionally, the absolute value of the difference between the number of reference regions in each other contour region and the number of reference regions in the first contour region can be calculated, and then the absolute value of the difference can be linearly normalized to obtain the reference weight of each other contour region.
[0093] It should be understood that most brush marks in brushed concrete images tend to be elongated, but some brush marks may be curved. Representing the direction of brush marks based on slope is inaccurate. Therefore, the aspect ratio (i.e., the degree of elongation) of the minimum bounding rectangle of each contour region can be used as a weighting coefficient for the direction of each contour region. The larger the aspect ratio, the greater the reference significance of the slope of the contour region, and vice versa.
[0094] Optionally, the directional difference of a contour region satisfies the following formula: in, Image of brushed concrete The Middle The directional differences of each contour region Image of brushed concrete The number of contour regions in the image. Image of brushed concrete The Middle The orientation parameters of each contour region Image of brushed concrete The Middle The orientation parameters of each contour region Image of brushed concrete The Middle The length of the longest side of the smallest bounding rectangle of each contour region. Image of brushed concrete The Middle The length of the shortest side of the smallest bounding rectangle of the contour region. Image of brushed concrete The Middle The number of reference regions for each contour region Image of brushed concrete The Middle The number of reference regions for each contour region Indicates to Take the absolute value. Indicates to Take the absolute value. This represents a linear normalization function, such as minimax normalization, used to map numerical values to [0,1].
[0095] Optionally, during the normalization process based on minimization, when the difference between the maximum and minimum values is less than a preset small amount, the normalization result is directly set to 0 or 1 to avoid division by zero errors.
[0096] In this formula, Characterizing the first The contour region and the first The degree of directional difference between two contour regions can be understood as a weighting factor for scale differences. The larger the value, the more significant the directional difference between the two contour regions. In brushed concrete, most brush marks have a consistent direction, while crack directions are often irregular. Therefore, areas with a large directional difference from most contour regions are more likely to be cracks. Characterizing the degree of difference in directional consistency between two contour regions can be understood as Interpreted as a weighting factor for size differences, when The smaller the value, the smaller the weight, and the contribution of the directional parameter difference is reduced; when The larger the value, the greater the weight, and the contribution of the difference in orientation parameters is amplified. Based on this, it can be accurately characterized that "when two contour regions are of similar size, the difference in orientation parameters between them has high reference value"; Indicates the first The aspect ratio of the smallest bounding rectangle of a contour region, after normalization, indicates that the region is more elongated and narrower as it approaches 1, and more square as it approaches 0. In this formula, it serves as a weighting coefficient: the absolute value of the slope contributes most when the region is elongated and narrow; the contribution of the absolute value of the slope is suppressed when the region is not elongated and narrow. This ensures that only the directional information of elongated and narrow regions is fully trusted.
[0097] The directional difference derived from this formula is a dimensionless numerical index used to quantify the degree to which the orientation of the first contour region deviates from the mainstream orientation in the image. This directional difference comprehensively considers the directional differences between the first contour region and all other contour regions, and through similar group weight adjustment, makes the calculation results more reflective of true directional anomalies.
[0098] By determining the number of reference regions, the popularity or niche appeal of a contour region's orientation can be effectively assessed. When both contour regions have few reference regions (i.e., both belong to "niche" orientations), the importance of the difference in their orientation parameters decreases; conversely, when a "popular" orientation region is compared with a "niche" orientation region, the difference in their orientation parameters will be given higher importance.
[0099] The methods provided in S201-S203 above quantify the dominant direction of the contour region by calculating direction parameters, providing a reliable and calculable basis for subsequent comparison of directional differences, thus freeing the judgment of directional differences from subjectivity and achieving automation.
[0100] In one implementation of the embodiments of this application, combined with Figure 1 ,like Figure 3 As shown, the irregularity of the morphology of each suspected crack region can be determined through S301-S302.
[0101] S301. Determine the width non-uniformity, edge roughness, and contour irregularity of the first suspected crack region.
[0102] The first suspected crack area is any suspected crack area.
[0103] It should be understood that cracks on brushed concrete are usually irregular, specifically manifested as: large variations in crack width, relatively rough edges, and irregular shapes. Based on these characteristics, the width non-uniformity, edge roughness, and contour irregularity of the first suspected crack area can be determined respectively.
[0104] In one alternative implementation, a width sequence of the first suspected crack region can be obtained, and the width non-uniformity of the first suspected crack region can be determined based on the difference between each width value in the width sequence and the average width.
[0105] It should be understood that this width sequence includes multiple width values.
[0106] Optionally, a skeleton extraction algorithm can be used to obtain the skeleton line of the first suspected crack area. Then, multiple sampling points are set along the skeleton line, and a perpendicular line to the skeleton line is drawn at each sampling point. The distance between the two intersection points of each perpendicular line and the edge of the contour area is calculated, and all distance values are arranged in order to form the width sequence.
[0107] Optionally, the skeleton line of the first suspected crack region can be obtained based on the Zhang-Suen algorithm.
[0108] Optionally, the arithmetic mean of all width values in the width sequence can be determined as the average width of the first suspected crack region. Then, the difference between each width value and the average width is calculated separately.
[0109] Optionally, the width non-uniformity of a suspected crack region satisfies the following formula: in, Image of brushed concrete The Middle Uneven width of suspected crack area Indicates the first The number of width values in the width sequence of a suspected crack region Indicates the first The width sequence of the first suspected crack region One width value, Indicates the first The average width of a sequence of suspected crack regions.
[0110] In this formula, after summing the absolute deviations of the width sequence, the result is divided by the sequence length. This eliminates the influence of region length on the calculation results of non-uniformity, making It only characterizes the dispersion of the width value. Thus, a uniform long region will have a smaller... The value, while a short, uneven region will have a larger value. value.
[0111] Understandably, by constructing a width sequence and calculating its dispersion to quantify width non-uniformity, the intuitive perception of width variation by the human eye can be transformed into a precise mathematical indicator, making the judgment of this feature objective and measurable.
[0112] In one optional implementation, the gradient magnitude of the edge pixels of the first suspected crack region is obtained; based on the variance of the gradient magnitude of the edge pixels of the first suspected crack region, the edge roughness of the first suspected crack region is determined.
[0113] Optionally, assuming the first suspected crack region is a suspected crack region in the first image, the above-mentioned acquisition of the gradient magnitude of the edge pixels of the first suspected crack region can specifically be: extracting the gradient magnitude of each pixel in the first image based on the gradient operator, and extracting the gradient magnitude corresponding to the edge pixels of the first suspected crack region.
[0114] Optionally, the variance of the gradient magnitude of the edge pixels of the first suspected crack region can be determined as the edge roughness of the first suspected crack region.
[0115] It should be understood that the larger the variance, the rougher the edge of the first suspected crack region.
[0116] Understandably, when the gradient change at the crack edge is drastic and irregular, the variance of the gradient amplitude is large; while the brush edge is smooth, the variance of the gradient amplitude is small. Using the variance of the gradient amplitude of edge pixels to define edge roughness can effectively capture the jaggedness and roughness of the edge.
[0117] In one optional implementation, a contour chain code sequence of a first suspected crack region is obtained, the contour chain code sequence including multiple chain code values; the absolute value sequence of the first-order difference of the contour chain code sequence is determined; and the contour irregularity of the first suspected crack region is determined based on the average value of the absolute values in the absolute value sequence.
[0118] Optionally, starting from any edge pixel in the first suspected crack area, the entire contour can be traversed according to the 8-direction chain code rule, and the direction code value of each step can be recorded to obtain the contour chain code sequence.
[0119] It should be noted that when obtaining the contour chain code sequence, at least three step-size direction code values should be recorded so that the number of elements in the contour chain code sequence is greater than or equal to 3.
[0120] Optionally, the absolute values of the first-order differences between adjacent elements in the contour chain code sequence are calculated to form an absolute value sequence.
[0121] Optionally, the irregularity of the profile of a suspected crack region satisfies the following formula: in, Image of brushed concrete The Middle The outline of the suspected crack area is irregular. Indicates the first In the absolute value sequence of the first-order difference chain code of the suspected crack region, the first... The value of each element, Indicates the first The number of elements in the absolute value sequence of the first-order differential chain code of a suspected crack region.
[0122] The arithmetic mean of all elements in the absolute value sequence of a first-order differential chain code is represented by the first-order differential chain code. The average directional change per unit length or unit step of the profile of a suspected crack region. This value quantifies the... The degree of tortuosity of the overall outline of a suspected crack region; the larger the value, the more frequent and drastic the changes in the outline direction, and the more curved and complex the outline. The greater the irregularity in the shape of a suspected crack area, the better.
[0123] It should be understood that when the profile is a smooth curve, the changes in adjacent chain code values are relatively small. The smaller the value, the more drastic the changes in adjacent chain code values when the outline is irregular. Relatively large.
[0124] Understandably, quantifying contour irregularities by averaging the first-order difference absolute value sequence of the chain code transforms the tortuosity of the contour into a numerical feature, which can accurately capture the complexity and randomness of crack contours, forming a stark contrast with the smooth contours of regular brush marks.
[0125] S302. The width non-uniformity, edge roughness and contour irregularity of the first suspected crack region are weighted and summed to obtain the morphological irregularity of the first suspected crack region.
[0126] It should be noted that, due to the different dimensions and numerical ranges of width non-uniformity, edge roughness, and contour irregularity, in order to eliminate the influence of dimensions and make the product results comparable, they need to be normalized separately before multiplying.
[0127] For example, the normalization method can be minimization normalization.
[0128] Optionally, the weights of these three factors can be determined by actual needs. In the absence of special requirements, the weights of these three factors can be 1 / 3 respectively.
[0129] Based on the methods described in S301-S302 above, the morphological characteristics of the contour region can be comprehensively described from three dimensions: width variation, edge characteristics, and contour shape. By fusing them through multiplicative relationships, it is ensured that only those regions that exhibit irregular characteristics in multiple dimensions are judged as cracks, thus improving the accuracy of static discrimination.
[0130] Combination Figure 1 ,like Figure 4 As shown, in one implementation of this application embodiment, the determination of the variation difference of each suspected crack region can be specifically implemented through S401-S403.
[0131] S401. Align multiple images to obtain the correspondence between suspected crack areas in different images.
[0132] In one alternative implementation, the image at the last moment among multiple images can be determined as the current image, and the remaining images can be determined as historical images. Feature points in multiple images are detected based on a feature detection algorithm. Feature descriptor matching is used to establish the correspondence between feature points in two adjacent time-step images. Based on the correspondence between feature points, a robust estimation algorithm is used to calculate the spatial transformation matrix between images. The historical images are then resampled using the spatial transformation matrix to align them with the current image in the spatial coordinate system.
[0133] Alternatively, the feature detection algorithm can be the SIFT algorithm.
[0134] Next, for each suspected crack region in the current image (e.g., the first suspected crack region), calculate its overlap with all suspected crack regions in each historical image. Select the suspected crack region in the historical image with the highest overlap as the corresponding suspected crack region of the first suspected crack region. Repeat this operation to determine the corresponding suspected crack regions in historical images for each suspected crack region in the current image, thus obtaining the correspondence between suspected crack regions in different images.
[0135] S402. Determine the corresponding suspected crack region based on the correspondence relationship.
[0136] It should be understood that the correspondence includes the corresponding suspected crack region for each suspected crack region in each image. Therefore, based on the correspondence, the corresponding suspected crack region for the first suspected crack region can be determined.
[0137] S403. Based on the area change of the first suspected crack region in different images, determine the variation of the first suspected crack region.
[0138] In this embodiment, the variation of the first suspected crack region can be a variation in area.
[0139] It should be understood that during the hardening process, the edge expansion of a real crack may be unstable and abrupt due to the release of internal stress, resulting in significant changes in area between adjacent images, while the area of static brush marks or noise remains stable.
[0140] In one alternative implementation, an area sequence of the first suspected crack region in multiple images can be obtained, and then the difference in variation can be determined based on the difference between adjacent area values in the area sequence.
[0141] Optionally, the variation of a suspected crack region satisfies the following formula: in, Image of brushed concrete The Middle Differences in the changes of suspected crack areas Indicates the number of multiple images. Indicates the first The first image The area of the suspected crack region. Indicates the first The first image The area of the suspected crack region. Indicates to Take the absolute value.
[0142] In this formula, Indicates the first image between adjacent images The change in area of each suspected crack region quantifies the drastic change in the crack at adjacent time points. Therefore, the larger the value, the more pronounced the crack change. The suspected crack area is an active, expanding crack; This represents the average area fluctuation of a suspected crack region per unit time, characterizing the instability or activity of a suspected crack region over a period of time. Determining this value as the variation difference can accurately reflect the degree of change of crack characteristics over time.
[0143] It should be noted that the above formula is based on the premise of the number of images acquired. Greater than or equal to 2. When At that time, since the differences in change cannot be calculated, it can be... The default value is 0 or this step is not performed.
[0144] Based on the methods provided by S401-S403, the same region can be tracked at different time points through image alignment and region matching. The differences in change are then determined by calculating the area changes. By utilizing the key physical characteristic that cracks can dynamically expand, dynamic and decisive evidence is provided for crack identification.
[0145] In summary, the embodiments of this application utilize the characteristics of cracked areas to determine whether an area is a cracked area, avoiding interference from the original texture of brushed concrete on the segmentation of cracked areas, and improving the accuracy of cracked area screening.
[0146] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0147] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
Claims
1. A method for evaluating the molding of concrete in civil engineering projects, characterized in that, The method includes: Acquire multiple images of the brush-textured concrete surface at different times; The directional difference of each contour region in each image is determined, and the contour regions of the multiple images are divided into crack regions and suspected crack regions based on the directional difference. The directional difference is used to characterize the difference between the direction of the contour region and the direction of other contour regions in the image to which it belongs. The variation and morphological irregularity of each suspected crack region were determined, and the variation was used to characterize the degree of change of the contour region at different times. Based on the irregularity and variation of the morphology of each suspected crack region, the suspected crack regions are divided into crack regions and non-crack regions. The forming quality of the brush-textured concrete is evaluated based on the number of cracked areas in the multiple images.
2. The method for evaluating the molding of concrete for civil engineering projects according to claim 1, characterized in that, Determine the directional differences of each contour region in each image, including: Determine the orientation parameters of each contour region in the first image. The orientation parameters are used to characterize the dominant orientation of the contour region in its respective image. The first image is any one of the plurality of images. Based on the direction parameters of each contour region, the direction parameter difference of each other contour region in the first image is determined. The direction difference of another contour region is the difference between the direction parameters of the other contour region and the direction parameters of the first contour region. The first contour region is any contour region in the first image. The directional differences of the first contour region are determined based on the difference in directional parameters of each of the other contour regions.
3. The method for evaluating the molding of concrete for civil engineering projects according to claim 2, characterized in that, Determining the directional difference of the first contour region based on the directional parameter difference of each of the other contour regions includes: Based on the difference in orientation parameters for each of the other contour regions, the number of reference regions for the first contour region is determined. The reference weight of each other contour region is determined based on the difference between the number of reference regions of each other contour region and the number of reference regions of the first contour region. The directional difference of the first contour region is determined based on the difference between the reference weight of each other contour region and the directional parameter of each other contour.
4. The method for evaluating the molding of concrete for civil engineering projects according to claim 1, characterized in that, Determine the morphological irregularities of each suspected crack region, including: The width non-uniformity, edge roughness, and contour irregularity of the first suspected crack region are determined, and the first suspected crack region is any suspected crack region. The morphological irregularity of the first suspected crack region is obtained by weighted summation of the width non-uniformity, edge roughness, and contour irregularity.
5. The method for evaluating the molding of concrete for civil engineering projects according to claim 4, characterized in that, The method further includes: Obtain the width sequence of the first suspected crack region, wherein the width sequence includes multiple width values; The width non-uniformity of the first suspected crack region is determined based on the difference between each width value and the average width in the width sequence of the first suspected crack region.
6. The method for evaluating the molding of concrete for civil engineering projects according to claim 4, characterized in that, The method further includes: Obtain the gradient magnitude of the edge pixels of the first suspected crack region; The edge roughness of the first suspected crack region is determined based on the variance of the gradient magnitude of the edge pixels of the first suspected crack region.
7. The method for evaluating the molding of concrete for civil engineering projects according to claim 4, characterized in that, The method further includes: Obtain the contour chain code sequence of the first suspected crack region, wherein the contour chain code sequence includes multiple chain code values; Determine the absolute value sequence of the first-order difference of the contour chain code sequence; The contour irregularity of the first suspected crack region is determined based on the average of the absolute values in the absolute value sequence.
8. The method for evaluating the molding of concrete for civil engineering projects according to claim 1, characterized in that, Determine the variation in each suspected crack region, including: The multiple images are aligned to obtain the correspondence between suspected crack areas in different images; Based on the correspondence, the corresponding suspected crack region of the first suspected crack region is determined; Based on the area variation of the first suspected crack region in different images, the variation difference of the first suspected crack region is determined.
9. The method for evaluating the molding of concrete for civil engineering projects according to claim 1, characterized in that, Based on the morphological irregularities and variations of each suspected crack region, the suspected crack regions are divided into crack regions and non-crack regions, including: The product of the irregularity and variation of the morphology of the first suspected crack region is determined as the probability of the first suspected crack region being a crack region. If the probability of a cracked area in the first suspected cracked area is greater than or equal to the cracked area probability threshold, the first suspected cracked area is determined to be a cracked area. If the probability of a cracked area in the first suspected cracked area is less than the cracked area probability threshold, the first suspected cracked area is determined to be a non-cracked area.
10. The method for evaluating the molding of concrete for civil engineering projects according to claim 1, characterized in that, Based on the directional differences, the contour regions of the multiple images are divided into crack regions and suspected crack regions, including: If the directional difference of the first contour region is greater than or equal to the directional difference threshold, the first contour region is determined as a crack region. If the directional difference of the first contour region is less than the directional difference threshold, the first contour region is identified as a suspected crack region.