Cable fire resistance testing method and system based on machine vision

Abstract

The present invention relates to the field of image processing technology, and in particular to a cable fire resistance test method and system based on machine vision, the method comprising: collecting multiple cable fire resistance test images and preprocessing to generate an HSV image set; taking any image in the image set as a target image, and any pixel point in the target image as a target pixel point, calculating the surface eigenvalue of the target pixel point, and the surface eigenvalue is positively correlated with the HSV change value of the target pixel point; calculating the flame influence degree of the target pixel point; marking the target pixel point whose flame influence degree is greater than a preset threshold as a flame point, and then detecting the position and shape of the flame in the image set. The present invention calculates the flame influence degree, marks the pixel point whose flame influence degree is greater than the preset threshold as a flame point, and realizes accurate identification of the flame position, thereby improving the accuracy and reliability of the cable test results.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a cable fire resistance testing method and system based on machine vision. Background Art

[0002] Cables are wires used to transmit electrical energy or signals. They can be a single or a collection of multiple conductors, usually surrounded by insulating materials to protect the conductors and prevent current leakage. Cables are widely used in power transmission, communications, industrial automation, home and commercial wiring, and other fields.

[0003] In order to verify the ability of cables to maintain normal operation for a certain period of time under fire conditions, it is necessary to conduct fire resistance tests on cables. By using machine vision systems or other sensors, the response of cables under flame conditions is monitored in real time, including parameters such as flame size and position. Modern fire resistance tests not only rely on traditional flame temperature and action time evaluation, but also introduce more advanced monitoring technologies to ensure the accuracy of test results and the reliability of cable performance.

[0004] At present, the patent application document with the publication number CN116678887A discloses a detection device, method and equipment for mineral insulated cables. The detection method of the mineral insulated cable is through a sample information entry module and a test parameter determination module. It ensures that each mineral insulated cable sample can be tested under standardized conditions, and the performance of the cable sample under extreme conditions (such as flames and high temperatures) is evaluated through the test result determination module and comprehensive consideration of the appearance image data, infrared image data and operation data. The detection method of the mineral insulated cable tests the safety of the cable in an emergency by simulating the test under actual fire conditions. However, during the test, there is interference from fire light, and the flame edge is unclear due to the interference from fire light, which reduces the accuracy of the measurement of the flame size and position, thereby reducing the accuracy and reliability of the cable test results. Summary of the invention

[0005] In order to solve the technical problem of reduced accuracy and reliability of the above-mentioned cable test results, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a cable fire resistance testing method based on machine vision comprises:

[0007] Collect multiple cable fire resistance test images and pre-process them to generate an HSV image set; use any image in the image set as a target image, and use any pixel point in the target image as a target pixel point;

[0008] Calculate the surface eigenvalue of the target pixel, which is positively correlated with the HSV change value of the target pixel;

[0009] Calculate the flame influence degree of the target pixel point. The flame influence degree is:

[0010] ; In the formula, The target pixel In the target image The degree of flame influence in The target pixel In the image The surface feature change value and target pixel point The difference in surface feature change value in the initial state, For the initial state and the target image In all images between The difference in surface feature change values ​​between two consecutive images, The target pixel In the image The HSV value and the target pixel in The sum of the deviations of the HSV values ​​corresponding to the initial state, represents normalization processing, is an exponential function with base e;

[0011] The target pixel points whose flame influence degree is greater than a preset threshold are marked as flame points to obtain the position and shape of the flame in the image set.

[0012] By adopting the above technical solution, the collected image is converted into the HSV color space to highlight the flame area, the entire image set is traversed and the surface feature value of each pixel is calculated. By comprehensively considering factors such as the surface feature change value, continuity change, and the total color deviation, the calculated flame influence degree can more accurately reflect the actual impact of the flame on the cable. The pixel points with a flame influence degree greater than the preset threshold are marked as flame points, and the flame position can be accurately identified, thereby improving the accuracy and reliability of the cable test results.

[0013] Preferably, calculating the surface feature value of the target pixel point includes:

[0014] Calculating a first difference sum and a first variance sum respectively, and taking the product of the first difference sum and the first variance sum as the surface feature value;

[0015] The first difference sum is the sum of the differences between the HSV of the target pixel in the target image and the HSV of the target pixel in the previous image of the target image;

[0016] The first variance sum is the sum of the variances of the HSV of the target pixel in all images before the target image.

[0017] By taking the product of the first difference sum and the first variance sum as the surface feature value, the information of color change and color stability is integrated, which helps to more comprehensively evaluate the color behavior of the pixel point and facilitates computer processing and analysis.

[0018] Preferably, after calculating the surface eigenvalue of the target pixel, the surface eigenvalue is normalized, and the normalized surface eigenvalue satisfies the relationship:

[0019] ; In the formula, is the normalized surface eigenvalue, For the target image Target pixel The surface characteristic values ​​of For the target image Target pixel The minimum value of the surface eigenvalue of For the target image Target pixel The maximum value of the surface eigenvalue.

[0020] Normalization can reduce overflow and underflow problems in numerical calculations and improve the numerical stability of the algorithm.

[0021] Preferably, the target pixel point In the image The process of obtaining the surface feature change value in includes:

[0022] Get the target pixel In the target image and target image The surface feature values ​​of all previous images are used to construct a first surface feature group;

[0023] The target pixel The preset window is the center, and any other pixel in the window is obtained in the target image and target image The surface feature values ​​of all previous images are used to construct a second surface feature group;

[0024] The target pixel In the image The surface feature change value in satisfies the relationship:

[0025] ; In the formula, The target pixel In the image The surface feature change value in is the pixel in the window The surface characteristic values ​​of is the similarity between the first surface feature group and the second surface feature group, For all pixels in the window, Indicates normalization processing.

[0026] By comparing the color changes of pixels at the same position in different images, the changes in image content can be captured more accurately. By analyzing the color changes, the influence of fire on image analysis can be eliminated to a certain extent.

[0027] Preferably, the similarity between the first surface feature group and the second surface feature group is calculated by a DTW algorithm.

[0028] Preferably, calculating the surface feature value of the target pixel point includes:

[0029] Calculate the average value of the sum of the gradient changes of the H value, S value and V value of the target pixel in the target image;

[0030] The average value is used as the surface feature value of the target pixel.

[0031] Preferably, calculating the surface feature value of the target pixel point includes:

[0032] Calculate the gradient changes of H value, S value and V value of the target pixel in the target image respectively;

[0033] The surface feature value of the target pixel is obtained by weighted averaging the gradient changes of H value, S value and V value.

[0034] By taking a weighted average of the gradient changes of H, S, and V, different color channels can be given different weights. For example, if hue changes are more important for a specific image analysis task, a higher weight can be given to the gradient changes of H.

[0035] In a second aspect, a cable fire resistance testing system based on machine vision includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned cable fire resistance testing method based on machine vision is implemented.

[0036] The beneficial effects of the present invention are as follows:

[0037] The present invention generates an HSV image set by preprocessing the cable fire resistance test image and calculates the surface eigenvalues ​​of the pixels in the image, so as to more accurately analyze the color change of the cable material at high temperature. By calculating the degree of flame influence and setting a threshold, the position and shape of the flame can be accurately detected, thereby enhancing the accuracy and reliability of the test results. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0039] Figure 1 It is a method flow chart of steps S1 to S4 in the cable fire resistance testing method based on machine vision in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0041] The embodiment of the present invention discloses a cable fire resistance testing method based on machine vision, referring to Figure 1 , including steps S1 to S4, which are specifically as follows:

[0042] S1: Collect multiple cable fire resistance test images and pre-process them to generate an HSV image set; use any image in the image set as a target image, and any pixel point in the target image as a target pixel point.

[0043] In the embodiment of the present invention, representative and uniform cable samples are selected to ensure that the length, diameter and material of the samples meet the requirements of the test standard. Before the test, all test equipment is calibrated, including thermometers, timers, image acquisition equipment, etc., to ensure the accuracy of the test data. In addition to temperature and time conditions, environmental factors such as humidity and airflow in the test room also need to be controlled.

[0044] Specifically, the cable sample is placed in the cable fire resistance test equipment according to the standard requirements, and a common camera or a high-speed camera is used to continuously capture the cable fire resistance test images once per second. In the embodiment of the present invention, the acquisition time is 30 minutes.

[0045] Convert the images captured by the camera into a format suitable for processing (such as TIFF or JPEG) to minimize the quality loss of the image during the conversion process, and then perform necessary image enhancement processing, including contrast adjustment, brightness adjustment, etc., to more clearly display features such as flames.

[0046] Image processing software is used to convert RGB images into HSV color space images, and then all processed HSV images are arranged in chronological order to form a time series image set.

[0047] It should be noted that HSV color space is an existing technology and will not be described in detail here. Converting images from RGB color space to HSV color space provides greater convenience and accuracy when performing image processing tasks such as color detection and segmentation. At the same time, due to its characteristics that are closer to human vision, HSV color space is widely used in the field of image analysis in computer vision.

[0048] In order to facilitate subsequent analysis, any image in the above image set is taken as the target image, and any pixel point in the target image is taken as the target pixel point.

[0049] S2: Calculate the surface eigenvalue of the target pixel. The surface eigenvalue is positively correlated with the HSV change value of the target pixel.

[0050] After the image set is obtained in the above step S1, the dynamics of color changes in the image set are analyzed.

[0051] Specifically, the first difference sum and the first variance sum are calculated respectively, and the product of the first difference sum and the first variance sum is used as the surface feature value; the first difference sum is the sum of the differences between the HSV of the target pixel in the target image and the HSV of the target pixel in the previous image of the target image; the first variance sum is the sum of the variances of the HSV of the target pixel in all images before the target image.

[0052] For example, in the image set The image is the analysis object and is marked as the target image. , select the target image Any pixel on , marking it as the target pixel .

[0053] For the target image and the target image Calculate the target pixel at the same position in the previous or next image in the image set The difference between the three channels of Hue (hue), Saturation (saturation), and Value (brightness) is summed to get the target pixel point The first difference and.

[0054] For the target pixel , collect target pixels In the target image The data on the three channels of Hue, Saturation, and Value in all previous images are used to calculate the target pixel point The variances of the three channels of Hue (hue), Saturation (saturation), and Value (brightness) in the historical image are then summed to obtain the first variance sum.

[0055] The first difference sum calculated above is multiplied by the first variance sum, and the result is used as the target pixel point The surface characteristic values ​​of .

[0056] For example, the target pixel The surface eigenvalues ​​of satisfy the following relationship:

[0057]

[0058] In the formula, For the target image Target pixel The surface characteristic values ​​of is the first difference and, is the first sum of variances.

[0059] By calculating the first difference sum, the color change of the target pixel between two consecutive frames is measured; by calculating the first variance sum, the color stability of the target pixel in all images before the target image is measured; combining color change and color stability, a comprehensive index is obtained to analyze the surface characteristics of the pixel.

[0060] Higher This indicates that the target pixel has experienced significant color changes in the image set, or the color changes in the historical images are unstable.

[0061] In addition, when calculating the target pixel After the surface eigenvalues ​​are normalized, the surface eigenvalues ​​after normalization satisfy the relationship:

[0062]

[0063] In the formula, is the normalized surface eigenvalue, For the target image Target pixel The surface characteristic values ​​of For the target image Target pixel The minimum value of the surface eigenvalue of For the target image Target pixel The maximum value of the surface eigenvalue.

[0064] In other embodiments, by calculating the target pixel In the target image The average value of the sum of the gradient changes of the H (Hue) value, S (Saturation) value, and V (Value) value is used as the target pixel point The surface characteristic values ​​of .

[0065] In other embodiments, the target pixel points can also be calculated separately In the target image The gradient changes of H value, S value and V value are weighted averaged to get the target pixel point. The surface characteristic values ​​of .

[0066] According to the calculation of target pixel The surface eigenvalues ​​of all other pixels in the image can be calculated in the same way.

[0067] S3: Calculate the flame influence degree of the target pixel.

[0068] Before the cable fire test, the cable surface is stable and no visible changes occur, so the characteristic parameters such as color and brightness of each pixel in the image are consistent. As the test progresses, the cable begins to be affected by the high temperature, which can cause phenomena such as scorching, shedding, flames and smoke on the cable surface. Due to the changes in the cable surface, the color, brightness and other characteristics of the corresponding pixels in the image will also change. When the cable surface changes, the changes in a single pixel may not be obvious enough, especially when the changes are small or the camera resolution is not high enough. By analyzing the changes in a single pixel and its surrounding pixels, the changes in surface characteristics can be estimated more accurately.

[0069] Specifically, get the target pixel In the target image and target image The surface feature values ​​of all previous images are used to construct a first surface feature group;

[0070] The target pixel Set a 5×5 window as the center and obtain the pixel value of any other pixel in the window in the target image. and target image Surface feature values ​​in all previous images, constructing a second surface feature group;

[0071] The target pixel In the image The surface feature change value in satisfies the relationship:

[0072]

[0073] In the formula, The target pixel In the image The surface feature change value in is the pixel in the window The surface characteristic values ​​of is the similarity between the first surface feature group and the second surface feature group, For all pixels in the window, represents the normalization process, where .

[0074] The similarity between the first surface feature group and the second surface feature group is calculated by a DTW (Dynamic Time Warping) algorithm.

[0075] A high value for surface feature change indicates that the target pixel and its surrounding area have undergone significant changes in the image set.

[0076] Then calculate the target pixel The flame influence degree satisfies the following expression:

[0077]

[0078] In the formula, The target pixel In the target image The degree of flame influence in The target pixel In the image The surface feature change value and target pixel point The difference in surface feature change value in the initial state, For the initial state and the target image In all images between The difference in surface feature change values ​​between two consecutive images, The target pixel In the image The HSV value and the target pixel in The sum of the deviations of the HSV values ​​corresponding to the initial state, represents normalization processing, is an exponential function with base e.

[0079] Larger This means a stronger fire effect, indicating that the target pixel The changes related to fire are more significant.

[0080] According to the above target pixel The flame influence degree of all other pixels can be calculated by the same method.

[0081] S4: Mark the target pixel points whose flame influence degree is greater than a preset threshold as flame points, so as to obtain the position and shape of the flame in the image set.

[0082] During the cable burning time, the flame influence degree of each pixel point is normalized, and the pixel points whose flame influence degree is greater than a preset threshold (in the embodiment of the present invention, the threshold is set to 0.8) are marked as flame points. The marked flame points are analyzed to determine the characteristics of the flame, such as the shape and position.

[0083] In addition, the two images with the largest difference in the surface feature change value of the pixel point are found. The time points corresponding to these two images, i.e., the start and end time of the flame, are determined as the actual burning time of the cable during the test. At the same time, in the time period other than the burning time, by observing the surface feature change value of the pixel point, the change range of the cable surface during the experiment can be intuitively understood.

[0084] An embodiment of the present invention also discloses a cable fire resistance testing system based on machine vision, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a cable fire resistance testing method based on machine vision according to the present invention is implemented.

[0085] The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0086] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0087] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0088] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A cable fire resistance test method based on machine vision, characterized in that: include: Collect multiple cable fire resistance test images and preprocess them to generate HSV image sets; Taking any image in the image set as a target image, and taking any pixel point in the target image as a target pixel point; Calculate the surface eigenvalue of the target pixel, which is positively correlated with the HSV change value of the target pixel; Calculate the flame influence degree of the target pixel point. The flame influence degree is: ; In the formula, The target pixel In the target image The degree of flame influence in The target pixel In the image The surface feature change value and target pixel point The difference in surface feature change value in the initial state, For the initial state and the target image In all images between The difference between the surface feature change values ​​in any two consecutive images, The target pixel In the image The HSV value and the target pixel in The sum of the deviations of the HSV values ​​corresponding to the initial state, represents normalization processing, is an exponential function with base e; The target pixel points whose flame influence degree is greater than a preset threshold are marked as flame points, so as to obtain the position and shape of the flame in the image set; The target pixel In the image The process of obtaining the surface feature change value in includes: Get the target pixel In the target image and target image The surface feature values ​​of all previous images are used to construct a first surface feature group; Target pixel The preset window is the center, and any other pixel in the window is obtained in the target image and target image The surface feature values ​​of all previous images are used to construct a second surface feature group; The target pixel In the image The surface feature change value in satisfies the relationship: ; In the formula, The target pixel In the image The surface feature change value in is the pixel in the window The surface characteristic values ​​of is the similarity between the first surface feature group and the second surface feature group, are all the pixels in the window.

2. The cable fire resistance testing method based on machine vision according to claim 1 is characterized in that: The calculating of the surface feature value of the target pixel point comprises: Calculating a first difference sum and a first variance sum respectively, and taking the product of the first difference sum and the first variance sum as the surface feature value; The first difference sum is the sum of the differences between the HSV of the target pixel in the target image and the HSV of the target pixel in the previous image of the target image; The first variance sum is the sum of the variances of the HSV of the target pixel in all images before the target image.

3. The cable fire resistance testing method based on machine vision according to claim 2 is characterized in that: After calculating the surface eigenvalue of the target pixel, the surface eigenvalue is normalized, and the normalized surface eigenvalue satisfies the relationship: ; In the formula, is the normalized surface eigenvalue, For the target image Target pixel The surface characteristic values ​​of For the target image Target pixel The minimum value of the surface eigenvalue of For the target image Target pixel The maximum value of the surface eigenvalue.

4. The cable fire resistance testing method based on machine vision according to claim 3 is characterized in that: The similarity between the first surface feature group and the second surface feature group is calculated by a DTW algorithm.

5. The cable fire resistance testing method based on machine vision according to claim 1, characterized in that: The calculating of the surface feature value of the target pixel point comprises: Calculate the average value of the sum of the gradient changes of the H value, S value and V value of the target pixel in the target image; The average value is used as the surface feature value of the target pixel.

6. The cable fire resistance testing method based on machine vision according to claim 1, characterized in that: The calculating of the surface feature value of the target pixel point comprises: Calculate the gradient changes of H value, S value and V value of the target pixel in the target image respectively; The surface feature value of the target pixel is obtained by weighted averaging the gradient changes of H value, S value and V value.

7. Cable fire resistance testing system based on machine vision, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the cable fire resistance testing method based on machine vision according to any one of claims 1 to 6 is implemented.

Images