Image recognition-based concrete crack identification method, system, medium and device

Abstract

This invention provides a method, system, medium, and device for concrete crack identification based on image recognition. The method includes: acquiring a crack image corresponding to a target concrete; converting the crack image to obtain a converted image, wherein the conversion method includes at least perspective conversion and coordinate conversion; performing image processing on the converted image to obtain a target image, wherein the image processing method includes at least crack identification and image trimming; obtaining a set of cracks within the target image using a preset set algorithm based on the target image; performing vectorization operations on each crack set to obtain a set of lines for the crack set; and inputting the line set into drawing software for drawing to complete the crack identification operation. This invention enables automated image localization of concrete cracks, vectorizing cracks into multi-segment lines, thereby freeing up heavy, mechanical manual labor, greatly improving work efficiency, and reducing labor costs.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of civil engineering and image processing technology, in particular to a concrete crack identification method, system, medium and equipment based on image recognition. BACKGROUND

[0002] With the development of civil engineering technology and people's demand for large, high, deep buildings, the application of large (bulk) concrete is increasing year by year; with the improvement of domestic living standards, the labor wage increases rapidly, and the number of employees in the civil industry and other harsh environmental conditions gradually decreases.

[0003] At present, in the process of daily maintenance of concrete, manual observation and measurement are required, such as crack investigation of structures at sea, which is affected by tide, wind and waves, and the working time is extremely limited. It may take several months to measure manually, not only the labor cost is high, but also the work efficiency is bottlenecked, which is difficult to improve. SUMMARY

[0004] The purpose of the present application is to provide a concrete crack identification method, system, medium and equipment based on image recognition, which can solve the problems of low efficiency and high cost of existing concrete crack identification.

[0005] In a first aspect, the present application provides a concrete crack identification method based on image recognition, which comprises:

[0006] Obtaining a crack image corresponding to the target concrete;

[0007] Converting the crack image to obtain a converted image, wherein the conversion method includes perspective conversion and coordinate conversion;

[0008] Image processing the converted image to obtain a target image, wherein the image processing method includes crack identification and image trimming;

[0009] Using a preset set algorithm based on the target image to obtain a crack set in the target image;

[0010] Vectorizing each crack set to obtain a line set of the crack set;

[0011] Inputting the line set into a drawing software to draw a drawing, so as to complete the crack identification work.

[0012] The cracks in the concrete are identified by using images to improve work efficiency. Specifically, perspective conversion and / or affine transformation of the images are used to convert the captured images to a planar projection (using an open source library opencv); then the images are trimmed by using a coordinate envelope map, and after completion, each crack set is summed up and vectorized, and finally the vectorization results are imported into engineering drawing software, thereby reducing the workload of on-site detection workers and engineers, and completing the heavy task work with the aid of computers. Not only can the crack identification be realized, but also the work efficiency is provided. After using the method, the work time can be compressed to several weeks or even several days by marking and taking pictures, and the time saved during the period is far greater than the equipment modification cost required by the method. Moreover, in combination with the unmanned aerial vehicle technology, remote (short distance) scanning of the structures on islands and water areas can be realized, and the quality detection of large structures such as offshore wind power, seawalls, dams and bridges has great value.

[0013] In a possible implementation of the present application, the crack image is obtained based on a preset image acquisition device, wherein the angle between the vertical plane of the shooting lens and the crack plane of the target concrete is less than a preset shooting angle.

[0014] In a possible implementation of the present application, the crack image is converted to obtain a converted image, specifically including:

[0015] The crack image is converted to a normal projection plane by using perspective conversion to obtain vertex data;

[0016] In the process of converting the crack image to the normal projection plane, the crack image is converted into a corresponding coordinate space by using coordinate conversion, and the coordinate conversion includes translation, scaling, rotation, flipping and shearing;

[0017] After the perspective conversion and the coordinate conversion are completed, the area image on the normal projection plane is taken as the converted image.

[0018] In a possible implementation of the present application, the converted image is processed to obtain a target image, specifically including:

[0019] The converted image is traversed, and crack pixels and division line pixels in the converted image are retained;

[0020] Pixels in the converted image except the crack pixels and the division line pixels are set to zero;

[0021] The vertex data is used for image trimming, and pixels other than the vertex connection lines are cut off to obtain the target image.

[0022] In a possible implementation of the present application, the coordinate range enclosed by each vertex line is identified by coordinate calculation on the vertex data, so that the image within the coordinate range is retained and the image outside the coordinate range is cut off.

[0023] In a possible implementation of the present application, the crack set in the target image is obtained by using a preset set algorithm, and the method specifically includes the following steps:

[0024] S1, extracting a point with a target value as a target point set in the target image;

[0025] S2, taking any point in the target point set as a starting origin;

[0026] S3, judging the square distance between any point in the target point set and the starting origin, extracting a first feature point in the target point set with a square distance less than or equal to a preset value from the starting origin into a list Last, and synchronously adding the first feature point into a crack Clist list, and the keyword of Clist in the crack dictionary is the number of times of taking any point;

[0027] S4, deleting the starting origin in the target point set and the point in the crack Clist list, traversing the list Last, judging the distance between any point in the list Last and any point in the target point set, and putting a second feature point less than or equal to the preset value into another list newLast, and synchronously adding the second feature point into the crack Clist list;

[0028] S5, replacing the list newLast into the list Last, and repeating steps S3-S4;

[0029] S6, waiting for the list Last to be empty, repeating steps S2-S5 until the target point set is empty, and stopping the loop;

[0030] S7, obtaining the crack set based on the crack Clist list data.

[0031] In a possible implementation of the present application, the vectorization operation is performed on each crack set to obtain a line set of the crack set, and the method specifically includes:

[0032] The vectorization operation is performed on the crack set by using a preset barycenter algorithm; wherein,

[0033] A series of barycenter points are obtained based on the barycenter algorithm;

[0034] A corresponding point family is obtained based on the barycenter point, and each point family is in one direction;

[0035] Based on the centroid and the family of points, the centerline of the crack line is obtained, and the vectorization operation of each crack set is completed to obtain the line set of the crack set.

[0036] Secondly, this application provides a concrete crack recognition system based on image recognition, the concrete crack recognition system based on image recognition includes:

[0037] The acquisition module is used to acquire crack images corresponding to the target concrete.

[0038] A conversion module is used to convert the crack image to obtain a converted image, wherein the conversion method includes at least perspective conversion and coordinate conversion;

[0039] The processing module is used to perform image processing on the converted image to obtain the target image, wherein the image processing method includes at least crack recognition and image trimming;

[0040] The set module is used to obtain a set of cracks within the target image based on the target image using a preset set algorithm;

[0041] The calculation module is used to perform vectorization operations on each of the crack sets to obtain the line set of the crack sets;

[0042] The input module is used to input the line set into the drawing software for drawing to complete the crack identification task.

[0043] Thirdly, this application provides the aforementioned computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the image recognition-based concrete crack identification method.

[0044] Fourthly, this application provides the aforementioned electronic device, which includes a processor and a memory; wherein the memory is used to store a computer program, and the processor is used to load and execute the computer program to enable the electronic device to perform the image recognition-based concrete crack identification method.

[0045] As described above, the image recognition-based concrete crack identification method, system, medium, and equipment of the present invention can realize automated image localization of concrete cracks and vectorize cracks into multi-segment lines to liberate heavy and mechanical manual labor, thereby greatly improving work efficiency and reducing labor costs. Furthermore, combined with UAV technology, it can also realize remote (short-distance) scanning of buildings (structures) in islands and waters, which is of great value for the quality inspection of large-scale structures such as offshore wind power, seawalls, dams, and bridges. Attached Figure Description

[0046] Figure 1The diagram shows a scenario application of the image recognition-based concrete crack identification method of the present invention in one embodiment;

[0047] Figure 2 The diagram shows the method steps of the image recognition-based concrete crack identification method of the present invention in one embodiment;

[0048] Figure 3 The diagram shows an actual case of the image recognition-based concrete crack identification method of the present invention in one embodiment;

[0049] Figure 4 The diagram shows the method steps of the image recognition-based concrete crack identification method of the present invention in another embodiment;

[0050] Figure 5 The diagram shows a perspective transformation structure of an embodiment of the image recognition-based concrete crack identification method of the present invention.

[0051] Figures 6A-6B The diagram shown is a perspective transformation schematic of an embodiment of the image recognition-based concrete crack identification method of the present invention;

[0052] Figure 7 The image shown is a schematic diagram of the converted image in one embodiment of the image recognition-based concrete crack identification method of the present invention;

[0053] Figure 8 The diagram shows the method steps of the image recognition-based concrete crack identification method of the present invention in another embodiment;

[0054] Figures 9A-9B The image shown is a schematic diagram of a target image in one embodiment of the image recognition-based concrete crack identification method of the present invention;

[0055] Figures 10A-10E The diagram shows a set of cracks in one embodiment of the image recognition-based concrete crack identification method of the present invention.

[0056] Figure 11 The diagram shows a crack vector diagram in one embodiment of the image recognition-based concrete crack identification method of the present invention.

[0057] Figure 12 The diagram shown is a CAD drawing schematic of an embodiment of the image recognition-based concrete crack identification method of the present invention.

[0058] Figure 13 The diagram shown is a structural schematic of an embodiment of the image recognition-based concrete crack recognition system of the present invention.

[0059] Figure 14The diagram shown is a structural schematic of an electronic device according to an embodiment of the present invention.

[0060] Component designation explanation

[0061] S202~S212 Step S402~S406 Step S802~S806 Step S1~S7 Step P1~P7 Step 130 Concrete crack identification system based on image recognition 131 Acquisition module 132 Conversion module 133 Processing module 134 Collection module 135 Operation module 136 Input module Detailed Implementation

[0062] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0063] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0064] In practical application, this application specifically includes seven steps, namely: Regional division, Image development and shooting, Perspective and coordinate transformation Crack identification and image retouching Find the set of cracks. Crack vector transformation and Vectorized results can be imported into engineering drawing software (such as AutoCAD), where, for example... Figure 1 As shown, the steps during application are... and This can be done in the field, requiring staff to take photos on-site in advance to obtain the corresponding image data. The steps are as follows: and steps ~ Can be used for office work, steps The steps involve using an image acquisition device to acquire data. ~ The process can be automated using computers, thus integrating the entire identification process into a single detection device. Specifically, the steps... and The crack detection part mainly uses the technology of the open-source image processing library OpenCV, and from the steps... Image retouching steps The original algorithm content is the core focus of this application.

[0065] Please see Figure 2 In one embodiment of the invention, the image recognition-based concrete crack identification method of the present invention includes the following steps:

[0066] Step S202: Obtain the crack image corresponding to the target concrete;

[0067] Step S204: Convert the crack image to obtain a converted image;

[0068] Step S206: Perform image processing on the converted image to obtain the target image;

[0069] Step S208: Based on the target image, obtain the set of cracks within the target image using a preset set algorithm;

[0070] Step S210: Perform vectorization operation on each set of cracks to obtain the line set of the crack set;

[0071] Step S212: Input the line set into the drawing software to draw the drawing, so as to complete the crack identification operation.

[0072] It should be noted that the acquired crack image can be input by the user terminal or acquired by a preset image acquisition device in actual application. When acquiring the image using an image acquisition device, the angle between the vertical plane of the shooting lens and the crack plane of the target concrete needs to be smaller than a preset shooting angle, which can generally be set to 10°. The smaller the angle between the vertical plane of the shooting lens and the crack plane of the target concrete, the higher the accuracy of identifying concrete cracks.

[0073] Furthermore, before the user inputs the crack image, the target concrete needs to be photographed to obtain the corresponding image. First, region division is required. Region division is mainly used for survey objects such as large volumes and large areas of concrete, which cannot be displayed in a single photograph, or where the crack is too small to be clearly represented in a single photograph. The key points of region division are that the vertex and boundary markers must be clear, and the vertex and boundary markers need to have a large color contrast with the component and the crack itself. If necessary, special developing agents can be used; developing agents can be chalk or markers, etc. See details for further information. Figure 3Introducing a real-world engineering case, the target concrete is the top surface of a wind turbine foundation, with an outer diameter of 13.6m and a tower outer diameter of 5.042m. Workers divided the foundation into 36 sections, each corresponding to a 10° angle (or 18 sections, each 20° angle, depending on actual needs). The sections were divided using a rope, with red ropes used to create a strong color contrast with the structural components. A black tape marker was placed every 40cm on the rope to indicate the distance along the rope's direction. The user then photographed the target concrete in the divided sections to obtain corresponding crack images. The photographs were required to capture a panoramic view of the area, with clearly distinguishable section lines, sufficient pixel resolution for crack recognition, and a higher perpendicularity between the camera lens and the shooting plane for a higher recognition accuracy.

[0074] Furthermore, in one embodiment of the invention, after acquiring the crack image, the crack image is converted to obtain the converted image, such as... Figure 4 As shown, the specific steps include the following:

[0075] Step S402: Use perspective transformation to transform the crack image onto the orthographic projection plane to obtain vertex data;

[0076] Step S404: During the process of converting the crack image to the orthographic projection plane, coordinate transformation is used to convert the crack image into the corresponding coordinate space. The coordinate transformation includes translation, scaling, rotation, flipping, and shearing.

[0077] Step S406: After completing the perspective transformation and the coordinate transformation, the region image on the orthographic projection plane is used as the transformed image.

[0078] It should be noted that perspective transformation essentially projects an image onto a new viewing plane, such as... Figure 5 As shown, perspective transformation is a transformation that utilizes the collinearity of the perspective center, image point, and target point, and rotates the projection plane (perspective plane) around the trace line (perspective axis) by a certain angle according to the law of perspective rotation, thereby disrupting the original projection fiber bundle while maintaining the geometric shape of the projection on the projection plane. This application utilizes perspective transformation to... Figure 3 The images captured in the image are transformed into the orthographic projection plane, and after translation, shearing, and rotation, the images within the region are transformed into the required coordinate space. The main mathematical methods used are as follows:

[0079] (1) Translation

[0080]

[0081] (2) Scaling

[0082]

[0083] (3) Rotation

[0084]

[0085] (4) Flip

[0086]

[0087] (5) Miscutting

[0088]

[0089] for Figure 3 The area image shown has the following lengths after conversion: πd×1 / 36=3.14×5.042 / 36 = 0.440m along the circumference of the tower base and 3.14×13.6 / 36=1.186m along the outer diameter of the foundation.

[0090] Accordingly, perspective transformation requires at least three vertices on the projection plane (three points determine a plane, but four points are generally used), their coordinates in the image, and their corresponding points on the projection plane (coordinates on the projection plane). Assume the transformed image has four points A, B, C, and D, with coordinates A(200, 200), B(638, 162), C(200, 4200), and D(1332, 4101). AC is the Y-axis direction, and AB is the X-axis direction. Using the string and markings, find the corresponding four points a, b, c, and d in the original image, as shown below. Figure 6A and Figure 6B As shown, during the conversion process, attention should also be paid to the size of the converted image to avoid pixels in the original image area exceeding the boundaries of the new image. The converted image is as follows: Figure 7 As shown.

[0091] Furthermore, in one embodiment of the invention, after obtaining the converted image, image processing is required to obtain the target image, such as... Figure 8 As shown, the specific steps include the following:

[0092] Step S802: Traverse the transformed image and retain the crack pixels and dividing line pixels in the transformed image;

[0093] Step S804: Set all pixels in the converted image except for the crack pixels and the dividing line pixels to zero;

[0094] Step S806: Based on the vertex data, perform image trimming and remove pixels outside the lines connecting each vertex to obtain the target image.

[0095] It should be noted that the processing methods for the converted image include at least crack identification and image retouching. Figure 7 The image is transformed to identify cracks, retaining the pixels of cracks and dividing lines, and setting other pixels to zero. The identified image is as follows: Figure 9A As shown, the specific method involves iterating through the image, dividing it into color ranges, setting the color range containing the developer and the dividing lines to "255", and setting the remaining color values ​​to "0". Based on the equations of the four points ABCD, where AC and BD are straight lines and AB and CD are arcs, pixels outside the designated areas are removed for image retouching. The resulting image is shown below. Figure 9B As shown, specifically by performing coordinate calculations on the vertex data, the coordinate range enclosed by the lines connecting each vertex is identified, thereby preserving the image within the coordinate range and cutting out the image outside the coordinate range. Coordinate calculation is an essential technical means for those skilled in the art, and will not be elaborated here.

[0096] Furthermore, in one embodiment of the invention, a preset set algorithm is used to obtain the set of cracks within the target image. The set algorithm proposed in this application is an original algorithm, which can be referred to as the "Snake Algorithm." The algorithm specifically includes the following steps:

[0097] S1, extract the points in the target image whose pixel value is the target value to obtain the target point set;

[0098] S2, randomly select a point from the set of target points as the starting origin;

[0099] S3, determine the square distance between any point in the target point set and the starting origin, extract the first feature point whose square distance between the target point set and the starting origin is less than or equal to a preset value and put it into a list Last, and simultaneously add the first feature point to the crack Clist list. The keyword of CLIST in the crack dictionary is the number of times any point is taken.

[0100] S4, delete the starting point and the points in the crack Clist list from the target point set, traverse the list Last, determine the distance from any point in the current list Last to any point in the target point set, put the second feature points less than or equal to the preset value into another list newLast, and simultaneously add the second feature points to the crack Clist list.

[0101] S5, Substitute the list newLast into the list Last, and repeat steps S3~S4;

[0102] S6. When the Last set of the list is empty, repeat steps S2 to S5 until the target point set is empty and then stop the loop.

[0103] S7, obtain the crack set based on the crack Clist list data.

[0104] It should be noted that the target value is "255", which means extracting a set of target points T with a pixel value of "255" from the target image. A point is randomly selected from the target point set T as the starting origin O. The relationship between the square distance of any remaining point X in the target point set T and the starting origin O and a predetermined value r is determined. The first feature point P with a distance less than or equal to r is placed in a list Last, and simultaneously added to the crack list Clist. C is then placed in the crack dictionary crackdict. The keyword of CLIST in the crack dictionary is the number of times any point is taken. The relationship determination formula is as follows:

[0105] ;

[0106] In this context, the coordinates of points in the target point set T are represented by (x, y), and the formula calculates the difference between the x and y coordinates of two points, not the distance between the two points. Therefore, it is called the square distance.

[0107] Delete the origin O and points from the crack list Clist in the target point set T. Traverse the list Last, determine the distance from any point in the current list Last to any point in the target point set T, and add the second feature point P' (less than or equal to r) to another list newLast. Simultaneously add the second feature point P' to the crack list Clist. Substitute the list newLast into the list Last, and repeat steps S3-S4. When the list Last is empty, repeat steps S2-S5 until the target point set is empty. After stopping the loop, obtain the crack set based on the crack list Clist data. Specifically, the idea of ​​the set algorithm proposed in this application is that if there are many pixels in an image, and the distance between them is less than or equal to r, they are considered to be points in the same crack set, thus dividing the pixels in an image into several sets to obtain the corresponding crack set. In this application, the target image was processed using an set algorithm to obtain a set of "7" cracks. Since "9" and "26-2-3" in the target image are work site numbers, they are ignored. The processed image is as follows. Figures 10A-10E As shown.

[0108] Furthermore, in one embodiment of the invention, after obtaining each set of cracks, the crack sets need to be vectorized. This vectorization involves another original algorithm, namely the centroid algorithm, with the specific steps as follows:

[0109] The crack set is vectorized using a preset centroid algorithm; wherein,

[0110] A series of centroid points are obtained based on the centroid algorithm described above;

[0111] Based on the centroid, a corresponding point family is obtained, and each point family represents a direction;

[0112] Based on the centroid and the family of points, the centerline of the crack line is obtained, and the vectorization operation of each crack set is completed to obtain the line set of the crack set.

[0113] It should be noted that crack vectorization needs to solve two problems: one is to depict the crack based on its outline, and the other is to ensure that the line depicting the crack outline is located at the center of the crack. To address this, this application proposes an original algorithm called the "centroid algorithm." The main idea of ​​the algorithm is as follows: For a set of cracks, take any point and calculate its distance to the remaining points. Points with a distance less than *r* are added to a new set, and the centroid of this set is calculated. Using the centroid as the center of a circle and *r* and *kr* as radii, the points within the circle are calculated and added to a new set. The set algorithm from the previous embodiment is then introduced to find the families of points in the new set with a distance radius of *k'•r*. Each family of points represents a direction. Points in the new set are then removed from the crack set. The search continues, starting with each family of points, until no new direction is found. Then, a new family of points is selected, and the search restarts until all family of points have been searched.

[0114] Specifically, with Figure 10A For example, the specific steps of the centroid algorithm include: P1, randomly selecting a point Y from a single set of cracks; P2, finding the set L0 of all points with Y as the center and a radius r (circle radius), calculating the average of the (x,y) coordinates of all points in L0 to obtain the centroid point C0, then finding the set L1 of all points with a radius r from C0, and finding the centroid point C1 for all points in L1, repeating the loop until the distance between two consecutive centroid points is less than or equal to a threshold d, where the threshold d is determined according to the actual situation of the image, generally 1 / 10 of the crack width pixels, recording the obtained centroid as centroid point C, putting it into the centroid point set U0, and deleting the point set L corresponding to C from the crack set.

[0115] Further, in step P3, with the centroid C as the center, calculate the set V of points within a radius from r to k·r (k is a coefficient, determined by the degree of discontinuity of the crack points; the larger k is, the faster the traversal and the coarser the vector diagram; generally, "2-3" is acceptable); in step P4, arbitrarily select a point from set V, and introduce the set algorithm from the above embodiment to calculate a dictionary crackdict with a predetermined value of k'·r (the value of k' mainly depends on the density of crack branches, generally "0.8-1.0", appropriately reduced when the branches are very dense), where each record in the dictionary represents a cluster of points, called a point family (in step P2 above, the point set L corresponding to C has been deleted, so this step involves forward probing; the number of point families corresponds to the number of crack branches ahead). Calculate the average (x, y) coordinates of all points in each point family to obtain the centroid C. i And establish the center of gravity C i and its corresponding point family L i The one-to-one correspondence is sorted according to the number of points in the point family, with the centroid C as the center point. i And point family L i Record the entries in the new dictionary, with entries containing more dots appearing first, and those containing fewer dots appearing later.

[0116] Furthermore, in step P5, extract the centroid C' and the point family L' of the point family with the most points from step P4 above, repeat steps P3 and P4, and record the remaining centroids and point families into the spare centroid set U. bc and backup point family set U bl In step P6, when no new point family is found in step P5, the series of centroids already obtained are placed into the line set, which is the first vectorized line in the crack set. Then, in the spare centroid set U... bc and backup point family set U bl Extract a centroid point and a family of points, and repeat steps P3 to P5 to obtain other vectorized lines; at P7, after processing all centroid points and family of points, the complete set of vectorized lines for the crack set is obtained. Among these, for... Figure 10A The vector diagram obtained after processing the crack set using the centroid algorithm is as follows: Figure 11 As shown.

[0117] Furthermore, in one embodiment of the invention, the line set is input into drawing software for drawing to complete the crack identification task.

[0118] It should be noted that the above embodiment completed the vectorization of the crack point set, resulting in a line set of the crack set. Based on the description in the above embodiment of "points A, B, C, and D, with coordinates A(200, 200), B(638, 162), C(200, 4200), D(1332, 4101), AC direction as the Y-axis direction, AB direction as the X-axis direction," the CAD drawing space coordinates are adjusted and rotated to the sector of the currently processed image to align the image coordinates with the CAD coordinates. Using the pline command in CAD, the line set from the above embodiment is drawn onto the CAD drawing, completing the conversion from image to CAD drawing. Figure 11 The resulting vector graphic converted into a drawing is as follows: Figure 12 As shown.

[0119] Please see Figure 13 In one embodiment, this embodiment provides a concrete crack recognition system 130 based on image recognition, the system comprising:

[0120] The acquisition module 131 is used to acquire the crack image corresponding to the target concrete;

[0121] The conversion module 132 is used to convert the crack image to obtain a converted image, wherein the conversion method includes at least perspective conversion and coordinate conversion;

[0122] Processing module 133 is used to perform image processing on the converted image to obtain a target image, wherein the image processing method includes at least crack recognition and image trimming;

[0123] The set module 134 is used to obtain a set of cracks in the target image based on the target image using a preset set algorithm;

[0124] The calculation module 135 is used to perform vectorization operations on each of the crack sets to obtain the line set of the crack sets;

[0125] Input module 136 is used to input the line set into drawing software for drawing to complete the crack identification task.

[0126] Since the specific implementation of this embodiment corresponds to the aforementioned method embodiment, the same details will not be repeated here, and those skilled in the art should also understand this. Figure 13 The division of the modules in the embodiments is only a logical functional division. In actual implementation, they can be fully or partially integrated into one or more physical entities. These modules can be fully implemented in software through processing element calls, fully implemented in hardware, or some modules can be implemented in software through processing element calls and some modules can be implemented in hardware.

[0127] In addition, this application provides a computer-readable storage medium as described above, on which a computer program is stored, which, when executed by a processor, implements the image recognition-based concrete crack identification method.

[0128] See Figure 14 This embodiment provides an electronic device, which includes at least a memory and a processor connected via a bus. The memory stores a computer program, and the processor executes the computer program stored in the memory to perform all or part of the steps in the aforementioned method embodiment.

[0129] In summary, this invention enables automated image localization of concrete cracks, vectorizing the cracks into multi-segment lines to liberate heavy, mechanical manual labor, thereby greatly improving work efficiency and reducing labor costs. Furthermore, combined with UAV technology, it can also achieve remote (short-range) scanning of buildings and structures in islands and waters, which is of great value for the quality inspection of large-scale structures such as offshore wind power, seawalls, dams, and bridges.

[0130] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0131] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0132] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0133] This application also provides a computer-readable storage medium. Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing a processor. The program can be stored in a computer-readable storage medium, which is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disk, and any combination thereof. The storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state drive (SSD)).

[0134] This application embodiment may also provide a computer program product comprising one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in this application embodiment are generated. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0135] When the computer program product is executed by a computer, the computer performs the method described in the foregoing method embodiments. The computer program product can be a software installation package; when the foregoing method is required, the computer program product can be downloaded and executed on the computer.

[0136] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0137] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for identifying concrete cracks based on image recognition, characterized in that, include: Obtain the crack image corresponding to the target concrete; The crack image is transformed to obtain a transformed image, wherein the transformation method includes at least perspective transformation and coordinate transformation; The transformed image is processed to obtain a target image. In the target image, the pixel values ​​of crack pixels and dividing line pixels are non-zero target values, and the pixel values ​​of other pixels are set to 0. The image processing method includes at least crack recognition and image trimming. The image trimming is to remove the image outside the region enclosed by the line connecting the vertices after perspective transformation. Based on the target image, a set of cracks within the target image is obtained using a preset set algorithm; The set of cracks within the target image is obtained using a pre-defined set algorithm, specifically including the following steps: S1, extract the points in the target image whose pixel value is the target value to obtain the target point set; S2, randomly select a point from the set of target points as the starting origin; S3, determine the square distance between any point in the target point set and the starting origin, extract the first feature point whose square distance between the target point set and the starting origin is less than or equal to a preset value and put it into a list Last, and simultaneously add the first feature point to the crack Clist list. The keyword of Clist in the crack dictionary is the number of times any point is taken. S4, delete the starting point and the points in the crack Clist list from the target point set, traverse the list Last, determine the distance from any point in the current list Last to any point in the target point set, put the second feature points less than or equal to the preset value into another list newLast, and simultaneously add the second feature points to the crack Clist list. S5, Substitute the list newLast into the list Last, and repeat steps S3 to S4; S6. When the Last set of the list is empty, repeat steps S2 to S5 until the target point set is empty and then stop the loop. S7, obtain the crack set based on the crack Clist list data; Perform vectorization operations on each of the aforementioned crack sets to obtain the line set of the crack sets; The step of performing vectorization operations on each of the crack sets to obtain the line set of the crack sets specifically includes: A series of centroid points were obtained based on the centroid algorithm; Based on the centroid, a corresponding point family is obtained, and each point family represents a direction; Based on the centroid and the family of points, the center line of the crack line is obtained, and the vectorization operation of each crack set is completed to obtain the line set of the crack set. The centroid algorithm includes: for a set of cracks, take any point, calculate the distance to the other points, put points smaller than the circle radius r into a set, and calculate the centroid of the set; Using the centroid of the set as the center and r and k·r as the radii, find the points inside the ring and put them into a new set. The value of k is determined according to the degree of discontinuity of the crack points. Based on the set algorithm, a family of points with a distance radius of k′·r is obtained in the new set. Each family of points represents a direction. Points in the new set are then removed from the crack set. The value of k′ is determined according to the density of crack branches. The denser the branches, the smaller k′ is. The search begins with each family of points until no new direction can be found. Then, the search starts again with a different family of points until all families of points have been searched. The set of lines is input into drawing software for drawing to complete the crack identification task.

2. The concrete crack identification method based on image recognition according to claim 1, characterized in that, The image of the crack is acquired using a preset image acquisition device, wherein the angle between the vertical plane of the shooting lens and the crack plane of the target concrete is less than a preset shooting angle.

3. The concrete crack identification method based on image recognition according to claim 1, characterized in that, The process of converting the crack image to obtain the converted image specifically includes: The crack image is transformed onto the orthographic projection plane using perspective transformation to obtain vertex data; In the process of transforming the crack image to the orthographic projection plane, coordinate transformation is used to convert the crack image into the corresponding coordinate space. The coordinate transformation includes translation, scaling, rotation, flipping, and shearing. After completing the perspective transformation and the coordinate transformation, the region image on the orthographic projection plane is used as the transformed image.

4. The concrete crack identification method based on image recognition according to claim 3, characterized in that, The step of processing the converted image to obtain the target image specifically includes: Traverse the transformed image and retain the crack pixels and dividing line pixels in the transformed image; Set all pixels in the transformed image except for the crack pixels and the dividing line pixels to zero; Image reshaping is performed based on the vertex data, and pixels outside the lines connecting each vertex are removed to obtain the target image.

5. The concrete crack identification method based on image recognition according to claim 4, characterized in that, By calculating the coordinates of the vertex data, the coordinate range enclosed by the lines connecting each vertex is identified, thereby preserving the image within the coordinate range and cutting out the image outside the coordinate range.

6. A concrete crack identification system based on image recognition, employing the concrete crack identification method based on image recognition as described in any one of claims 1-5, characterized in that, include: The acquisition module is used to acquire crack images corresponding to the target concrete. A conversion module is used to convert the crack image to obtain a converted image, wherein the conversion method includes at least perspective conversion and coordinate conversion; The processing module is used to perform image processing on the converted image to obtain the target image, wherein the image processing method includes at least crack recognition and image trimming; The set module is used to obtain a set of cracks within the target image based on the target image using a preset set algorithm; The calculation module is used to perform vectorization operations on each of the crack sets to obtain the line set of the crack sets; The input module is used to input the line set into the drawing software for drawing to complete the crack identification task.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the image recognition-based concrete crack identification method as described in any one of claims 1 to 5.

8. An electronic device, characterized in that, The electronic device includes a processor and a memory; wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to enable the electronic device to perform the image recognition-based concrete crack identification method as described in any one of claims 1 to 5.

Images