A method for detecting the quality of a rubber tube for construction

Abstract

The application relates to the technical field of image processing, and discloses a rubber pipe quality detection method for building construction, which comprises the following steps: acquiring a rubber pipe surface gray image; obtaining a growth starting point according to a gray distribution, obtaining a direction guide line according to a defect feature of the rubber pipe surface, obtaining a gradient direction closeness degree of each direction guide line according to a gradient direction quantization value of the direction guide line and the growth starting point, acquiring a guide line rich highlighting coefficient of a center pixel point and a regional gully offset degree of a growth region, obtaining a guide line growth support coefficient of each direction guide line according to the gradient direction closeness degree of each direction guide line, the guide line rich highlighting coefficient of the center pixel point and the regional gully offset degree of the growth region, and combining the guide line growth support coefficient and a region growing algorithm to realize rubber pipe surface defect detection. The application aims to improve the accuracy of rubber pipe surface defect detection and realize defect detection and discrimination of the rubber pipe surface.

Description

Technical Field

[0001] This invention relates to the field of image processing, and more specifically to a method for quality inspection of rubber hoses used in building construction. Background Technology

[0002] In the construction industry, rubber hoses are crucial tools for transporting various materials, and their quality directly impacts construction efficiency and safety. Substandard hoses can lead to material leaks and transport obstructions, affecting not only construction progress but also potentially causing safety accidents. Without regular maintenance, poor-quality hoses with surface defects will have a significantly reduced lifespan. Therefore, selecting high-quality hoses requires careful consideration of surface defects.

[0003] Traditional methods for detecting surface defects in rubber hoses often rely on simple contact measurements. These methods are not only inefficient and susceptible to subjective factors and lighting conditions, but they are also often unable to detect subtle, hidden surface defects (such as microcracks, material inhomogeneity, and embedded micro-foreign objects).

[0004] As the performance requirements for rubber hoses in construction projects continue to increase, higher demands are being placed on the accuracy and comprehensiveness of hose quality testing. Therefore, developing an efficient, accurate, and comprehensive quality testing method for rubber hoses used in construction is of paramount importance. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for quality inspection of rubber hoses used in building construction, thereby resolving the existing issues.

[0006] The present invention provides a method for quality inspection of rubber hoses used in construction, which adopts the following technical solution:

[0007] One embodiment of the present invention provides a method for quality inspection of rubber hoses used in construction, the method comprising the following steps:

[0008] Acquire images of the hose surface;

[0009] The growth starting point is obtained based on the grayscale distribution of the hose surface image. The region where the growth starting point is located is defined as the growth region, and the directional guidelines in the neighborhood of the growth starting point are obtained. The gradient direction of each pixel in the hose surface image is quantized and statistically analyzed to obtain a gradient direction histogram. The gradient direction quantization value of the directional guidelines and the growth starting point is obtained based on the gradient direction histogram. The gradient direction similarity of each directional guideline is obtained based on the gradient direction quantization value of the directional guideline and the growth starting point. The guideline richness and highlighting coefficient of the center pixel is obtained based on the gradient direction similarity of each directional guideline and the gradient magnitude distribution. The regional groove offset of the growth region is obtained. The gradient magnitude flatness coefficient of each directional guideline is obtained based on the gradient magnitude distribution of each pixel. The guideline growth support coefficient of each directional guideline is obtained based on the gradient direction similarity of each directional guideline, the difference in gradient magnitude flatness coefficient, the guideline richness and highlighting coefficient of the center pixel, and the regional groove offset of the growth region.

[0010] The growth region inclusion conditions are obtained based on the growth support coefficients of the directrixes in each direction; the growth region inclusion conditions are used as growth criteria and combined with the region growth algorithm to segment the hose surface image; the hose surface defect detection and discrimination are realized based on the segmentation results.

[0011] Preferably, obtaining the directional directrixes within the neighborhood of the growth starting point includes:

[0012] The growth starting point is used as the center pixel to construct a neighborhood. The straight line containing the center pixel in the neighborhood is rotated at 45° intervals with the center pixel as the center to obtain the guide lines in each direction.

[0013] Preferably, the step of obtaining the degree of similarity of the gradient directions of each directional guideline based on the gradient direction quantization value between the directional guideline and the growth starting point includes:

[0014] The Euclidean distance between the quantized gradient direction value of each directrix and the quantized gradient direction value of the growth starting point is used as the degree of closeness of the gradient direction of each directrix.

[0015] Preferably, the expression for obtaining the line enrichment and highlighting coefficient of the center pixel based on the proximity of the gradient directions and the distribution of gradient magnitudes of the line guides is:

[0016]

[0017] In the formula, Center pixel The richness of the directrix within the neighborhood highlights the coefficient. The number of direction lines. For preset coefficients, The first pixel in the neighborhood of the center pixel The gradient magnitude of the directrix in each direction. The gradient magnitude of the center pixel. For the first The gradient direction of the directrix. Center pixel The degree to which the gradient directions are similar For normalization function, It is the arctangent function. It is a natural exponential function.

[0018] Preferably, the expression for obtaining the regional gully offset of the growth area is:

[0019]

[0020] In the formula, For growth area Regional gully offset, For growth area The number of pixels, Center pixel gradient magnitude, For growth area The gradient magnitude of the j-th pixel. Center pixel With growth area The Euclidean distance between the j-th pixels.

[0021] Preferably, obtaining the gradient magnitude flatness coefficient of the directrix in each direction based on the gradient magnitude distribution of each pixel includes:

[0022] The variance of the gradient magnitude of all pixels on each direction directrix is ​​used as the gradient magnitude flatness coefficient of each direction directrix.

[0023] Preferably, the guideline growth support coefficient for each guideline is obtained based on the similarity of gradient directions and the difference in gradient magnitude flatness coefficients, the guideline richness and prominence coefficient of the center pixel, and the regional groove offset of the growth region. The expression is as follows:

[0024]

[0025] In the formula, For the first The support coefficient for the growth of a directrix in a given direction. For the first The proximity of the gradient directions of the directrixes For the first The gradient magnitude flatness coefficient of the directrix in each direction. For growth area The gradient magnitude flatness coefficient, For growth area Regional gully offset, For growth area With each direction of the directrix The groove offset of the region of the union of the sets of pixels involved. Center pixel The richness of the directrix within the neighborhood highlights the coefficient. This is the normalization function.

[0026] Preferably, the step of obtaining the growth region addition conditions based on the directrix growth support coefficient of each direction directrix includes:

[0027] The growth region is added when the growth support coefficient of the directional directrix is ​​greater than the average growth support coefficient of the directrix of each directional directrix.

[0028] Preferably, the step of segmenting the tubing surface image by incorporating growth region inclusion conditions as growth criteria and combining them with a region growing algorithm includes the following steps:

[0029] Add the directional guidelines that meet the growth region addition conditions to the region where the growth start point is located, and then take the pixels that meet the growth region addition conditions as the new growth start points; if none of the directional guidelines meet the growth region addition conditions, randomly select any pixel in the neighborhood of the growth start point as the new growth start point for region growth, and so on, to complete the tube surface image segmentation.

[0030] Preferably, the step of detecting and judging surface defects of the hose based on the segmentation results includes:

[0031] Calculate the absolute value of the difference between the mean gray value of each region after segmentation of the hose surface image and the mean gray value of the standard hose image, and define the regions where the absolute value of the difference exceeds a set threshold as defect regions.

[0032] The present invention has at least the following beneficial effects:

[0033] This invention primarily obtains guidelines in various directions by analyzing the defect features of the rubber hose surface, analyzes the gradient direction similarity between pixels on the guidelines and the growth starting point, and combines this with a region growing algorithm to detect and discriminate defects on the rubber hose surface, thereby improving the accuracy of defect detection and discrimination. This invention also distinguishes defects on the rubber hose surface by comparing the grayscale differences between defect images and standard images, improving the efficiency and accuracy of defect detection.

[0034] Furthermore, this invention constructs the gradient direction proximity and gradient magnitude flatness coefficient of the guidelines in each direction, obtains the guideline richness and prominence coefficient of the center pixel and the regional groove offset of the growth region, calculates the guideline growth support coefficient of each guideline, and thus obtains the growth region addition conditions for each guideline. Combined with the region growth algorithm, a segmented image of the hose surface is obtained. Based on the segmentation results, defect detection and discrimination of the hose surface are realized, solving the problem of large defect detection errors caused by unclear defect edges due to the influence of illumination on the hose surface. This invention has the beneficial effects of high stability, high accuracy, and high efficiency in defect monitoring. Attached Figure Description

[0035] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 A flowchart illustrating the steps of a method for quality inspection of rubber hoses used in building construction, as provided in one embodiment of the present invention;

[0037] Figure 2 This is a schematic diagram of the direction alignment;

[0038] Figure 3 This is a gradient direction histogram. Detailed Implementation

[0039] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method for quality inspection of construction hoses proposed according to the present invention. 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.

[0040] 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 invention pertains.

[0041] The following description, in conjunction with the accompanying drawings, details a specific scheme for a method of quality inspection of rubber hoses used in building construction provided by the present invention.

[0042] Please see Figure 1 The diagram illustrates a flowchart of a method for quality inspection of construction hoses according to an embodiment of the present invention. The method includes the following steps:

[0043] Step S001: Acquire an image of the tube surface using an image acquisition device and perform preprocessing.

[0044] Construction sites present complex environments where hoses may be subjected to harsh conditions such as high pressure, impact, abrasion, temperature variations, and chemical corrosion. For example, high-pressure hoses used for pumping concrete may suddenly burst under high pressure if they have minute cracks, air bubbles, or structural damage that are difficult to detect. This could not only cause material to splash and injure people, but also damage surrounding equipment and even lead to more serious safety accidents. The method of this invention, through precise image analysis, can capture early defects that are difficult to detect with the human eye, such as microcracks, uneven material distribution, and damage to the inner lining layer, eliminating potential safety hazards in their infancy. This provides the most direct protection for the lives of personnel on construction sites.

[0045] Specifically, in this embodiment, the surface RGB image of the hose will be obtained by using a CMOS camera under uniform lighting conditions from above. This image will serve as the data source for detecting and judging surface defects of the hose. It should be noted that there are many methods for obtaining the surface RGB image of the hose. The specific image acquisition method can be implemented by existing technology and is not within the scope of protection of this embodiment. Therefore, it will not be described in detail.

[0046] Then, the RGB image of the hose surface is converted into a grayscale image using the averaging method. Next, the grayscale image of the hose surface is denoised using a bilateral filtering denoising algorithm to remove noise interference. Since the averaging method and the bilateral filtering denoising algorithm are both existing known technologies, they will not be described in detail here.

[0047] Thus, the denoised grayscale image of the hose surface can be obtained according to the method described in this embodiment, which can serve as the data basis for subsequent detection and discrimination of defects on the hose surface.

[0048] Step S002: Based on the defect features of the hose surface, obtain the directional guidelines in the neighborhood around the growth starting point, obtain the growth region addition conditions based on the gradient direction similarity between the directional guidelines and the region where the growth starting point is located, and use the region growing algorithm to achieve hose surface image segmentation.

[0049] Specifically, this embodiment obtains the directional guidelines within the neighborhood of the growth starting point based on the defect characteristics of the hose surface. Simultaneously, it obtains a gradient direction histogram by quantifying the gradient direction of each pixel in the hose surface image. Based on the gradient direction histogram, it calculates the proximity of the directional guidelines to the gradient direction of the growth starting point, thereby obtaining the guideline richness and prominence coefficient of the center pixel and the regional groove offset of the growth region. Combining the gradient direction proximity, the guideline richness and prominence coefficient, the regional groove offset, and the gradient amplitude flatness coefficient, it obtains the guideline growth support coefficient for each directional guideline, thus obtaining the growth region addition conditions. Using a region growth algorithm, the accuracy of hose surface defect detection is improved, achieving the detection and discrimination of hose surface defects. The specific process for constructing the growth region addition conditions for each directional guideline in the hose surface image is as follows:

[0050] First, the gradient of the grayscale image of the hose surface is calculated using the Sobel operator to obtain the gradient magnitude and gradient direction of each pixel in the hose surface image. The specific expressions for the gradient magnitude and gradient direction are as follows:

[0051]

[0052]

[0053] In the formula, The gradient magnitude of each pixel. The gradient direction for each pixel. This represents the gradient magnitude in the horizontal direction. This represents the gradient magnitude in the vertical direction.

[0054] Rubber hoses typically exhibit three types of defects: scratches, bumps, and white spots. Because of the cylindrical shape of the hose, a distinct boundary line appears when light shines from an angle. If the Canny algorithm is used for defect contour segmentation, the extra lines can easily cause errors in the segmentation of the defect area. Therefore, to eliminate this influence and better identify defects on the hose surface, this embodiment employs a region growing method to construct a region from pixels with similar characteristics, thereby achieving defect detection on the hose surface. For the grayscale image of the hose surface, a randomly selected pixel is used as the growth starting point, and the area where the growth starting point is located is denoted as the growth region, represented as... .

[0055] Taking the growth starting point as the center pixel, denoted as... Take the center pixel The surrounding neighborhood, in this embodiment, is taken as... of The implementation can adjust the neighborhood as needed; this embodiment does not impose any restrictions on this. of The pixels in the neighborhood are denoted as follows from left to right and from top to bottom: , No. The gradient magnitude of each pixel is denoted as . , No. The gradient direction of each pixel is denoted as... .

[0056] For the area around the center pixel In the neighborhood, four directional guidelines are set as the locations where defect areas are most likely to exist. The similarity between the center pixel and the neighborhood containing the four directional guidelines is used as the criterion for adding the directional guidelines to the growth region. A schematic diagram of the locations of the four directional guidelines is shown below. Figure 2 As shown.

[0057] The gradient magnitudes of the four directional directrixes are obtained by summing and averaging the gradients of all pixels along each directrix. and gradient direction .

[0058] To obtain pixels Dot around The approximate distribution of gradient magnitudes along the four directional guidelines within the neighborhood is shown in this embodiment. The gradient direction histogram is used to statistically analyze the gradient information of each pixel in the image.

[0059] In this embodiment, the horizontal axis of the gradient direction histogram is divided into 6 ranges, namely... , … The gradient directions are denoted as 0, 1, ..., 5, a total of six degree ranges. The number of pixels whose gradient direction falls within each of these six degree ranges is then counted. This count is used as the ordinate to obtain the gradient direction histogram of the hose surface image. The gradient direction histogram is shown below. Figure 3 As shown.

[0060] Calculate the angles of the four directional lines separately. The number of pixels within the specified degree range, i.e., the ordinate. The value of is denoted as the ordinate value of the directrix in different directions in the gradient direction histogram. The ordinate value of the center pixel in the gradient direction histogram is denoted as... This yields the degree of proximity between the directrix in each direction and the gradient direction of the center pixel. The specific expression for this degree of proximity is as follows:

[0061]

[0062] In the formula, For the first The proximity of the gradient directions of the directrixes For normalization function, For the first The ordinate values ​​of each direction directrix in the gradient back-side histogram. This represents the ordinate value of the center pixel in the gradient direction histogram. The larger the value, the closer the gradient direction of the directional guideline is to the gradient direction of the center pixel.

[0063] To characterize the complexity of the four directional guidelines within the neighborhood of the center pixel after quantization to a six-degree range—that is, the more irregular the gradient magnitudes of the four directional guidelines after quantization to a six-degree range, the richer the texture within the neighborhood of the center pixel, indirectly characterizing the neighborhood features of that pixel and thus enabling more accurate growth during region growing—the average of the similarity between the calculated directional guidelines and the gradient direction of the center pixel is calculated and denoted as the center pixel's gradient direction. The degree of similarity in gradient directions is expressed as... Therefore, the profilometry richness and highlighting factor of the center pixel is constructed, and the specific expression of the profilometry richness and highlighting factor is as follows:

[0064]

[0065] In the formula, Center pixel The richness of the directrix within the neighborhood highlights the coefficient. The number of direction lines. As a preset coefficient, in this embodiment The implementer can set it up themselves. The first pixel in the neighborhood of the center pixel The gradient magnitude of the directrix in each direction. The gradient magnitude of the center pixel. For the first The gradient direction of the directrix. Center pixel The degree to which the gradient directions are similar For normalization function, It is the arctangent function. It is a natural exponential function.

[0066] in, For the center pixel relative to the first The degree of prominence of the area along the directional guideline; a larger value indicates that the center pixel is more prominent than the first directional guideline. The larger the gradient amplitude of a directional guideline, the greater the gradient amplitude of the defective area on the hose surface compared to the normal hose surface. Therefore, when there is a decrease in gradient amplitude, it indicates that the pixel on that directional guideline is more likely to be the location of the center pixel. Simultaneously, when the gradient direction of that directional guideline is relatively close to the gradient direction of the center pixel, i.e. If the value is smaller, it means that the pixels included in that directional guideline are more likely to be the area where the center pixel is located.

[0067] Then, the region where the central pixel X is located, i.e., the growth region, is analyzed, and the regional groove offset of the growth region is constructed. The specific expression of the regional groove offset is as follows:

[0068]

[0069] In the formula, For growth area Regional gully offset, For growth area The number of pixels, Center pixel gradient magnitude, For growth area The gradient magnitude of the j-th pixel. Center pixel With growth area The Euclidean distance between the j-th pixels.

[0070] Among them, if most pixels in the growth region A value greater than 0 indicates that pixel X is located at the center of the growth region. When the value can be both positive and negative, with a higher proportion of negative values, it indicates that there are still many pixels in the growth region that have not yet grown, i.e., the region groove offset. The larger the value, the greater the likelihood of growing more pixels during the region growth process.

[0071] The gradient magnitude flatness coefficient of each direction directrix is ​​obtained by analyzing the distribution of gradient magnitudes of pixels within the directrix of each direction in the image of the hose surface. The specific expression for the gradient magnitude flatness coefficient is as follows:

[0072]

[0073] In the formula, Image of the surface of the tubing The gradient magnitude flatness coefficient of the nth directrix, where n is the nth directrix. The number of pixels contained in each directional guideline. For the first The first direction on the alignment line Gradient magnitude of each pixel For the first The average gradient magnitude of n pixels on a directional directrix, if The smaller the value, the flatter the gradient magnitude of the pixels along that direction. It should be noted that the gradient magnitude flatness coefficient for each direction is the variance of the gradient magnitudes of all pixels along that direction.

[0074] The magnitude of a pixel's gradient indicates the presence of edges with significant changes. This embodiment calculates the gradient magnitude and direction of four directional guidelines within the X-neighborhood of the center pixel, aiming to correlate the gradient magnitude and direction of these four guidelines with the gradient magnitude of the center pixel. The gradient magnitude and direction are compared. Among the four directional guidelines, if there exists a directional guideline whose gradient direction is related to the center pixel... When the gradient directions are relatively close and the difference between the gradient magnitude flatness coefficient of the pixels contained in the direction guideline and the gradient magnitude flatness coefficient of the pixels in the growth region is small, the current direction guideline is added to the growth region.

[0075] Based on this, the directrix growth support coefficients for each direction are constructed, and the specific expression for the directrix growth support coefficients is as follows:

[0076]

[0077] In the formula, For the first The support coefficient for the growth of a directrix in a given direction. For the first The proximity of the gradient directions of the directrixes For the first The gradient magnitude flatness coefficient of the directrix in each direction. For growth area The gradient magnitude flatness coefficient, This represents the difference between the gradient magnitude flatness coefficient of the directrix and the gradient magnitude flatness coefficient of the growth region. The smaller the value, the more similar the gradient magnitude of the corresponding directrix is ​​to the gradient magnitude of the pixels within the growth region. For growth area Regional gully offset, For growth area With each direction of the directrix The groove offset of the region of the union of the sets of pixels involved. Center pixel The richness of the directrix within the neighborhood highlights the coefficient.

[0078] Among them, when If it is greater than 0, it means that the center pixel is... The regional groove offset increases when pixels within the neighborhood directional guideline are added to the growth region, i.e., the center pixel... The closer the pixel is to the center, the more the grooves in the area after adding the pixel on the directional guideline tend to be similar to the defect area on the tube surface. This indicates that the pixel on the directional guideline is more likely to support the addition of a growth area. At the same time, the larger the richness and prominence coefficient of the guideline in the X neighborhood of the center pixel, the more prominent the directional guideline is and the more complex the guidelines are. This is consistent with the characteristic that the texture of the defect area on the tube surface is more complex and extended than that of the normal tube. The pixel in this neighborhood is more likely to be added to the growth area.

[0079] This yields the guideline growth support coefficients of the guidelines in each direction within the neighborhood of the center pixel X. The growth region inclusion conditions are obtained based on the directrix growth support coefficients of each direction directrix. The specific expression for the growth region inclusion conditions is as follows:

[0080]

[0081] In the formula, The number of directional guidelines in the neighborhood of the center pixel is defined as follows: if the gradient direction similarity of each directional guideline is greater than the average gradient direction similarity of all directional guidelines in the neighborhood of the center pixel, then the directional guideline is considered a pixel of the same type as the center pixel and added to the region where the center pixel is located. After addition, each added pixel is used as a new growth starting point. The gradient direction distribution of pixels in the surrounding neighborhood is analyzed with the growth starting point as the center. It should be noted that if the neighborhood of the growth starting point contains pixels that are already in the growth region, then those pixels are not analyzed. If none of the directional guidelines in the neighborhood of the growth starting point meet the growth region addition condition, then any pixel in the neighborhood that is not in the growth region is randomly selected as a new growth starting point, and growth starts again until there are no isolated pixels in the entire tube surface image, at which point growth stops, and the tube surface image segmentation is completed.

[0082] Step S003: Combine the segmentation results of the hose surface image with the grayscale distribution of the standard hose surface image to obtain the defect discrimination accuracy of each region, thereby realizing the detection and discrimination of hose surface defects.

[0083] The defect discrimination accuracy of the hose surface image is obtained by comparing the grayscale distribution of the standard hose surface image with the grayscale distribution of each region segmented from the captured hose surface image. The specific expression for the defect discrimination accuracy is as follows:

[0084]

[0085] In the formula, After segmenting the image of the hose surface, the first... Defect detection accuracy in each region For normalization function, The grayscale mean of the standard rubber tube surface image. After segmenting the image of the tubing surface, the first The average gray level of each region, with a set threshold. ,like This proves that the area is a defective area. This proves that the region is a non-defect region. In this embodiment... The implementer may make adjustments according to the actual situation, and this embodiment does not impose any restrictions on this.

[0086] Thus, defect detection and identification on the surface of the hose were achieved.

[0087] For rubber hoses used in construction, quality inspection can adhere to stricter standards; any defects that could lead to leaks, ruptures, or affect delivery efficiency are considered acceptable. Any area that fails to meet the quality standards is directly deemed unqualified. Quality can also be graded based on the accuracy of the hose defect assessment. For example, it can be categorized as "qualified," "conditionally qualified" (requires use under specific conditions or with a marked usage range), and "unqualified," etc. Details will not be elaborated here.

[0088] In summary, the embodiments of the present invention solve the problem of large defect detection errors caused by unclear defect edges due to the influence of light on the surface of the rubber hose. By combining the gradient direction of the pixel points of the rubber hose surface image with the region growing algorithm, the accuracy and efficiency of defect detection on the rubber hose surface are improved, and defect detection and discrimination on the rubber hose surface are realized.

[0089] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0090] 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.

[0091] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for quality inspection of rubber hoses used in construction, characterized in that, The method includes the following steps: Acquire images of the hose surface; The growth starting point is obtained based on the grayscale distribution of the hose surface image. The region where the growth starting point is located is defined as the growth region, and the directional guidelines in the neighborhood of the growth starting point are obtained. The gradient direction of each pixel in the hose surface image is quantized and statistically analyzed to obtain a gradient direction histogram. The gradient direction quantization value of the directional guidelines and the growth starting point is obtained based on the gradient direction histogram. The gradient direction similarity of each directional guideline is obtained based on the gradient direction quantization value of the directional guideline and the growth starting point. The guideline richness and highlighting coefficient of the center pixel is obtained based on the gradient direction similarity of each directional guideline and the gradient magnitude distribution. The regional groove offset of the growth region is obtained. The gradient magnitude flatness coefficient of each directional guideline is obtained based on the gradient magnitude distribution of each pixel. The guideline growth support coefficient of each directional guideline is obtained based on the gradient direction similarity of each directional guideline, the difference in gradient magnitude flatness coefficient, the guideline richness and highlighting coefficient of the center pixel, and the regional groove offset of the growth region. The growth region inclusion conditions are obtained based on the growth support coefficients of the directrixes in each direction; the growth region inclusion conditions are used as growth criteria and combined with the region growth algorithm to segment the hose surface image; the hose surface defect detection and discrimination are realized based on the segmentation results; The expression for the profilometry coefficient of the center pixel, obtained based on the proximity of the gradient directions and the distribution of gradient magnitudes of the profilometry, is as follows: In the formula, Center pixel The richness of the directrix within the neighborhood highlights the coefficient. The number of direction lines. For preset coefficients, The first pixel in the neighborhood of the center pixel The gradient magnitude of the directrix in each direction. The gradient magnitude of the center pixel. For the first The gradient direction of the directrix. Center pixel The degree to which the gradient directions are similar For normalization function, It is the arctangent function. It is a natural exponential function; The expression for obtaining the regional gully offset of the growth region is: In the formula, For growth area Regional gully offset, For growth area The number of pixels, Center pixel gradient magnitude, For growth area The gradient magnitude of the j-th pixel. Center pixel With growth area The Euclidean distance between the j-th pixels; The guideline growth support coefficient for each direction guideline is obtained based on the similarity of gradient directions and the difference in gradient magnitude flatness coefficients, the guideline richness and prominence coefficient of the center pixel, and the regional groove offset of the growth region. The expression is as follows: In the formula, For the first The support coefficient for the growth of a directrix in a given direction. For the first The proximity of the gradient directions of the directrixes For the first The gradient magnitude flatness coefficient of the directrix in each direction. For growth area The gradient magnitude flatness coefficient, For growth area Regional gully offset, For growth area With each direction of the directrix The groove offset of the region of the union of the sets of pixels involved. Center pixel The richness of the directrix within the neighborhood highlights the coefficient. This is the normalization function.

2. The method for quality inspection of rubber hoses used in construction according to claim 1, characterized in that, The specific method for obtaining the directrixes in each direction within the neighborhood of the growth starting point is as follows: The growth starting point is used as the center pixel to construct a neighborhood. The straight line containing the center pixel in the neighborhood is rotated at 45° intervals with the center pixel as the center to obtain the guide lines in each direction.

3. The method for quality inspection of rubber hoses used in construction according to claim 1, characterized in that, The degree of similarity between the gradient directions of each directional guideline and the gradient direction quantization value of the growth starting point is obtained, including: The Euclidean distance between the quantized gradient direction value of each directrix and the quantized gradient direction value of the growth starting point is used as the degree of closeness of the gradient direction of each directrix.

4. The method for quality inspection of rubber hoses used in construction according to claim 1, characterized in that, The step of obtaining the gradient magnitude flatness coefficient of the directrix in each direction based on the gradient magnitude distribution of each pixel includes: The variance of the gradient magnitude of all pixels on each direction directrix is ​​used as the gradient magnitude flatness coefficient of each direction directrix.

5. The method for quality inspection of rubber hoses used in construction according to claim 1, characterized in that, The specific method for obtaining the growth region addition conditions based on the directrix growth support coefficient of each direction directrix includes: The growth region is added when the growth support coefficient of the directional directrix is ​​greater than the average growth support coefficient of the directrix of each directional directrix.

6. The method for quality inspection of rubber hoses used in construction according to claim 1, characterized in that, The step of segmenting the hose surface image by combining the addition conditions of the growth region as a growth criterion with a region growing algorithm includes: Add the directional guidelines that meet the growth region addition conditions to the region where the growth start point is located, and then take the pixels that meet the growth region addition conditions as the new growth start points; if none of the directional guidelines meet the growth region addition conditions, randomly select any pixel in the neighborhood of the growth start point as the new growth start point for region growth, and so on, to complete the tube surface image segmentation.

7. The method for quality inspection of rubber hoses used in building construction according to claim 1, characterized in that, The specific method for detecting and identifying surface defects in the hose based on the segmentation results is as follows: Calculate the absolute value of the difference between the mean gray value of each region after segmentation of the hose surface image and the mean gray value of the standard hose image, and define the regions where the absolute value of the difference exceeds a set threshold as defect regions.

Images