Cable conductor sectional area detection method based on improved Canny operator

Abstract

The invention relates to a cable conductor sectional area detection method based on an improved Canny operator, and belongs to the field of power cables. According to the method, the improved Canny operator is adopted for edge extraction, the edge information of the cable conductor is effectively extracted through methods of guided filtering noise reduction, multi-direction gradient calculation, local adaptive threshold and the like, and the accuracy of edge detection is improved. Meanwhile, an improved CPDA global angle calculation algorithm is adopted, circular ring sampling and a least square method are used for fitting included angle double sides, and the angle information of the cable conductor is accurately calculated. And finally, in combination with an improved Canny operator and a CPDA global angle calculation algorithm, non-contact detection of the sectional area of the cable conductor in the cable trench is realized, and the conductor area of the overlapped cable part is effectively identified. The method is suitable for cable conductor sectional area detection in the fields of power transmission, communication networks, traffic signal systems and the like, and has important practical value.

Description

Technical Field

[0001] The invention belongs to the field of power cables and relates to a cable conductor cross-sectional area detection method based on an improved Canny operator. Background Art

[0002] Cables play a vital role in modern industry and infrastructure, and are widely used in power transmission, communication networks, traffic signal systems and other fields. The performance and reliability of cables are directly related to the stability and efficiency of the entire system, and the cross-sectional area of ​​the cable conductor, as a key parameter, determines the current carrying capacity, electrical characteristics and mechanical strength of the cable. In the process of power transmission, accurate measurement of the cross-sectional area of ​​the conductor is particularly important to ensure the safe operation of the power system. If the cross-sectional area of ​​the conductor is insufficient, the cable may cause the insulation layer to be damaged due to overheating, leading to serious accidents such as short circuits or fires. In addition, in communication networks, changes in the cross-sectional area of ​​the conductor will affect the quality and stability of signal transmission, thereby affecting the normal operation of the communication system.

[0003] Cables are usually installed in a closed or semi-closed environment in a cable trench. The environment is complex and difficult to access and operate directly. The cross-sectional area calculation and cable identification of the cables in the cable trench require staff to enter the cable trench to perform corresponding operations. For example, the traditional cable conductor cross-sectional area measurement method mainly relies on physical contact, using a caliper or vernier caliper to measure the outer diameter of the cable, and then calculating the cross-sectional area of ​​the conductor through a formula. Although these methods can achieve high accuracy under laboratory conditions, they have obvious limitations in practical applications. First, in a closed environment such as a cable trench, it is often very difficult or even impossible to directly contact the cable for measurement. Secondly, the traditional method can only measure the outer diameter of the cable, and cannot directly reflect the actual cross-sectional area of ​​the internal conductor, especially when the thickness of the cable's outer coating material is uneven or the cable is aging and deformed.

[0004] In order to overcome these difficulties, researchers and engineers began to explore the use of advanced technical means to detect the cross-sectional area of ​​cable conductors. In recent years, with the rapid development of computer vision and image processing technology, cable conductor cross-sectional area detection technology based on image processing has gradually become a research hotspot. This method uses a camera to obtain images of the cable, and extracts the edge and shape information of the cable conductor through a series of algorithm processing to calculate the cross-sectional area of ​​the conductor. Image processing technology has the advantages of non-contact, high precision, and strong real-time performance, and is particularly suitable for measurements in complex environments such as cable trenches. Cable images obtained by high-resolution camera equipment can significantly improve the accuracy of subsequent processing after pre-processing steps such as grayscale, denoising, and image enhancement. The Canny edge detection algorithm can be used to extract the edges of cable conductors.

[0005] In addition to image processing technology, thermal imaging technology is also an effective detection method. The cable will generate heat during operation, and the temperature distribution map of the cable surface can be obtained through infrared imaging equipment. According to the principle of heat conduction and the material properties of the cable, the cross-sectional area of ​​the cable conductor can be calculated. Thermal imaging technology has the advantages of non-contact and fast speed, especially when the cable load changes, it can monitor the temperature change of the cable in real time. However, the accuracy of thermal imaging may be affected by factors such as ambient temperature and surface contamination, so multiple measurements and data corrections are required to improve accuracy.

[0006] In the process of testing cable conductors, the cross-sectional area of ​​the conductor is an important parameter for evaluating the quality of the cable. So far, the main testing methods used include wire diameter measurement method, weighing method and DC resistance method. The wire diameter measurement method is mainly applicable to single, untwisted cable conductors. The weighing method calculates the cross-sectional area of ​​the conductor by cutting several meters of cable and weighing and measuring the conductor length. This method is a destructive test. The required cable cutting length is large, and the cut cable cannot be used anymore. The weighing process is also relatively complicated. Therefore, it is not suitable for measuring the cross-sectional area of ​​cable conductors in cable trenches. The DC resistance method calculates the cross-sectional area of ​​the conductor by measuring the DC resistance and conductor length of the conductor. This method has high requirements on the cable length. If the cable length changes proportionally with the conductor cross-sectional area, this method cannot effectively determine the actual cross-sectional area of ​​the conductor.

[0007] Compared with the existing cross-sectional area detection method of cable conductors, the present invention adopts improved Canny operator edge extraction and CPDA global angle calculation algorithm, which is a non-contact cable cross-sectional area measurement method, avoiding physical damage to the cable; the noise reduction method adopted can improve the accuracy of edge detection, making the cross-sectional area detection result more accurate; the image processing technology adopted can greatly reduce the detection time and improve the detection efficiency compared with the traditional cross-sectional area measurement method, which is suitable for application in practical engineering.

[0008] In view of the problems in the prior art that when detecting the cross-sectional area of ​​cable conductors in cable trenches, manual entry into the cable trenches is required to contact the cable conductors, and the cable overlap leads to insufficient detection accuracy and detection speed of the cable conductor cross-sectional area. This patented invention will use an improved Canny edge detection algorithm and a CPDA global angle calculation algorithm to realize non-contact detection of the cross-sectional area of ​​cable conductors in cable trenches, thereby improving the detection accuracy and detection speed of the cable cross-sectional area, reducing the input of manpower and material resources, and ensuring the safe and stable operation of the power grid. Summary of the invention

[0009] In view of this, an object of the present invention is to provide a cable conductor cross-sectional area detection method based on an improved Canny operator.

[0010] In order to achieve the above object, the present invention provides the following technical solutions:

[0011] A cable conductor cross-sectional area detection method based on an improved Canny operator, the method comprising the following steps:

[0012] Improved Canny operator edge extraction: Guided filtering is used instead of Gaussian filtering for edge-preserving smoothing and noise reduction; based on the Canny operator's horizontal and vertical gradient calculation, two more direction detections of 45° and 135° are added; a local adaptive threshold is used instead of the global fixed method of the dual threshold of the non-maximum suppression technology to complete the edge detection of the cable;

[0013] Improved CPDA global angle calculation algorithm: ring sampling to extract global angle related pixels; least square method to fit the angle bilaterally; calculate the angle size based on the fitted bilaterally;

[0014] Cable cross-sectional area detection: Combined with the improved Canny operator edge extraction and CPDA global angle calculation algorithm, the area of ​​numerous cables in the cable trench is detected.

[0015] Further, the principle of the guided filtering is shown in the following formula:

[0016]

[0017] Where i and j are the subscripts of the pixel points, I is the guide image; P j is the input cable image, Q i is the cable image output after filtering; W ij (I) Determine the weights used in the weighted average operation for the guide image I.

[0018] Further, the cable edge detection processing method of the local adaptive threshold is as follows:

[0019] The 3×3 matrix is ​​used to move smoothly in the cable image, and the grayscale average value of the pixels in the 3×3 smoothing box is used as the adaptive threshold;

[0020] Compare the adaptive threshold within the 3×3 smooth box with the gradient magnitude of the center pixel:

[0021] If the gradient amplitude of the center point is greater than the adaptive threshold within the smoothing box, the point is an edge point and the output value is 1;

[0022] If the gradient amplitude of the center point is less than or equal to 1 / 2 of the threshold, the point is not an edge point and the output value is 0;

[0023] If the gradient amplitude of the center point is between 1 / 2 of the adaptive threshold and the entire adaptive threshold, its output value is the same as the output value of the previous center pixel;

[0024] If the previous center pixel is an edge point, the output value is 1, otherwise the output value is 0.

[0025] Furthermore, the gradient amplitude calculation formula is:

[0026]

[0027] The gradient direction θ(x,y) is:

[0028]

[0029] Where: (x, y) is the pixel point in the cable image; g x (x,y),g y (x,y)g 45o (x,y),g 135o (x, y) are the gradient components in the X, Y, 45° and 135° directions at the pixel point (x, y).

[0030] Furthermore, the ring sampling radius is adjusted according to actual conditions.

[0031] Further, the algorithm flow of the least squares fitting is as follows:

[0032] Determine the form of the fitted curve;

[0033] Determine the dataset of the variables;

[0034] Calculate the elements of the coefficient matrix and tail matrix as follows;

[0035]

[0036] Solve the system of equations to obtain the values ​​of the unknown coefficients.

[0037] The beneficial effects of the present invention are as follows: improving a series of complex operation problems such as the low accuracy of the detection of the cross-sectional area of ​​the cable conductor in the existing cable trench, the low detection speed and the need for contact area measurement, applying image processing technology, and realizing non-contact detection of the cross-section of the cable conductor in the cable trench by adopting an improved Canny edge detection algorithm and a CPDA global angle calculation algorithm, effectively identifying the cable conductor area of ​​the overlapping cable part, improving the detection accuracy and detection speed of the cable cross-sectional area, effectively reducing the investment of manpower and material resources, and reducing the cable maintenance time.

[0038] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0040] Figure 1 This is a schematic diagram of cable area detection;

[0041] Figure 2 It is an improved Canny operator;

[0042] Figure 3 This is the cable area detection flow chart. DETAILED DESCRIPTION

[0043] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0044] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0045] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0046] 1. Improve Canny operator edge extraction

[0047] The traditional Canny operator does not work well when detecting the edge of cable images. This is because the traditional Canny operator uses Gaussian filtering to smooth the cable image and cannot effectively remove local noise. When the image intensity changes slowly, false edges may be detected or local edges may be lost. In addition, the traditional Canny operator only detects cable images from the horizontal and vertical directions, making it difficult to fully extract the edge features of the cable. The non-maximum suppression technology used by the traditional Canny operator is the key to extracting accurate edge information. However, the Canny operator uses a globally fixed dual threshold method when performing non-maximum suppression, making it difficult to determine the optimal value of the threshold.

[0048] Based on this, the improved Canny operator is adopted. The main improvements of this method are as follows:

[0049] 1) Use guided filtering instead of Gaussian filtering for edge-preserving (smoothing) noise reduction;

[0050] 2) Add 45° and 135° direction detection based on the Canny operator's horizontal and vertical gradient calculations;

[0051] 3) A local adaptive threshold is used instead of a global fixed method of dual thresholds of non-maximum suppression technology to complete the edge detection of the cable.

[0052] The improved Canny operator is shown as follows Figure 1 shown.

[0053] 1.1 Guided Filtering

[0054] Guided filtering filters the cable image through the guide map, making the output image similar to the input cable image and its texture similar to the guide map. Guided filtering can not only smooth the edge of the cable image, but also accurately detect the cable edge information. Its principle is shown in formula (1):

[0055]

[0056] Where i and j are the subscripts of the pixel points, I is the guide image; P j is the input cable image, Q i is the cable image output after filtering; W ij (I) Determine the weights used in the weighted average operation for the guide image I.

[0057]

[0058] In the formula, ω k is a sliding window centered on pixel k; |ω| is the sliding window ω k The number of pixels in μ k is the mean value of the pixels in the window; I i and Ij is the value of two adjacent pixels; δ k is the variance of the pixels in the sliding window; ε is the penalty value.

[0059] According to formula (2), I i and I j On both sides of the cable edge, (I i -μ k ) and (I j -μ k ) are of different signs, and are of the same sign when they are on the same side of the cable edge. When the weights of the different signs are much smaller than the weights of the same sign, the pixels on both sides of the cable edge are given smaller weights, and the smoothing effect is weaker, while the pixels in the flat area of ​​the cable are given larger weights, and the smoothing effect is more obvious, which can maintain the contour edge of the cable image.

[0060] 1.2 Local Adaptive Threshold

[0061] The non-maximum suppression technology of the Canny operator adopts a global fixed method of dual thresholds. Whether the optimal threshold can be accurately determined and the selection of the optimal value have a great influence on the contour edge feature information of the cable. In this regard, the present invention adopts a local adaptive threshold method to replace the global fixed method of dual thresholds of the non-maximum suppression technology in the traditional Canny operator to realize the edge detection of the cable. The cable edge detection processing method of the local adaptive threshold is as follows.

[0062] The 3×3 matrix is ​​used to smoothly move within the cable image, and the grayscale average of the pixels within the 3×3 smoothing box is used as the adaptive threshold T. d , which is calculated as shown in formula (3).

[0063]

[0064] Where: M (x,y) is the gray value at (x, y) in the 3×3 smooth box. Since the 3×3 matrix smooth box is constantly changing, the image pixel gray value and threshold in the matrix smooth box will also change accordingly. Then, the adaptive threshold T in the 3×3 smooth box is d Compare with the gradient amplitude of the center pixel: if the gradient amplitude of the center point is greater than the adaptive threshold in the smooth box, then the point is an edge point and the output value is 1; if the gradient amplitude of the center point is less than or equal to 1 / 2 of the threshold, then the point is not an edge point and the output value is 0; if the gradient amplitude of the center point is between the adaptive threshold T d If the value is between 1 / 2 of the threshold and the entire adaptive threshold, the output value is the same as the output value of the previous center pixel. If the previous center pixel is an edge point, the output value is 1, otherwise the output value is 0, see formula (4).

[0065]

[0066] Where: M(x,y) is the output value of the edge point position, is the gradient amplitude at the point (x, y); Z is the output value, indicating that the output value is the same as the output value of the previous central pixel.

[0067] 1.3 Gradient Amplitude

[0068] The present invention adopts the improved Canny operator to detect the cable image from the horizontal and vertical directions while retaining the Sobel operator. It adds the gradient calculation in the two diagonal directions, taking the 45° and 135° diagonal directions in the counterclockwise direction. The cable is processed for the gradient amplitude. The convolution kernel in the 8-neighborhood of the center pixel can be taken as:

[0069]

[0070]

[0071] Where: S x S y S 45° S 135° The convolution kernels in the X, Y, 45° and 135° directions are used to calculate the cable image segmented by the convolution kernel. The gradient components of the image in the X, Y, 45° and 135° directions are g respectively. x , g y , g 45° , g 135° , and finally use the gradient components in each direction to calculate the gradient component amplitude and gradient direction.

[0072] The gradient component amplitude G(x,y) is:

[0073]

[0074] The gradient direction (angle) θ(x,y) is:

[0075]

[0076] Where: (x, y) is the pixel point in the cable image; g x (x,y),g y (x,y)g 45° (x,y),g 135° (x, y) are the gradient components in the X, Y, 45° and 135° directions at the pixel point (x, y).

[0077] 2. Improved CPDA global angle calculation algorithm

[0078] The research on corner point extraction algorithms related to CPDA is already very mature. Based on the research results of previous scholars, the present invention adopts an algorithm suitable for extracting corner points of cable conductor contours. The specific process is as follows:

[0079] (1) Ring sampling extracts global angle-related pixels. The global angle calculation must consider both the candidate corner point and the surrounding adjacent pixels. Therefore, it is necessary to sample and process the vicinity of the candidate corner point separately. The sampling method is ring sampling, and the collected pixels are the pixels adjacent to the point to be measured in the ring.

[0080] (2) Least squares method to fit the two sides of the angle. After sampling, the pixels can be divided into two parts according to the adjacency relationship. The two parts of pixels are fitted in the form of straight lines, and the two fitted straight lines are used as the two sides of the angle.

[0081] (3) Calculate the angle based on the fitted two sides. The slopes of the straight lines where the two sides are located after fitting are n and 1 and n 2 , the angle calculation formula is

[0082] θ=|arctann 1 -arctann 2 | (7)

[0083] 2.1 Region extraction based on ring sampling

[0084] The global angle needs to be calculated in combination with the corner point and the adjacent pixels, and the scale of the adjacent pixels needs to be limited to a certain range. Based on the above idea, the present invention selects a circular ring sampling method with adjustable radius to extract the corner point area. The circular ring sampling takes a certain point as the center, selects the pixel points connected to the sampling center inside the circular ring path of a given radius, and maps the selected pixel points to the rectangular coordinate system of the pixel distance space.

[0085] The ring sampling radius can be adjusted according to the actual situation. If the radius is too large and exceeds the feature size, it will affect the detection result. If the radius is too small, it will lead to insufficient information and inability to judge.

[0086] 2.2 Bilateral fitting based on least squares method

[0087] To calculate the angle, it is necessary to first determine the two sides of the angle. In a digital image, pixels are discrete points, and the two sides of the angle are also discrete points. Therefore, to calculate the angle, it is necessary to first fit a straight line that can accurately represent the two sides. The fitting method adopted in the present invention is the least squares method.

[0088] The least squares method of curve fitting is to find an approximate function through given discrete data points to minimize the sum of squares of the errors between the observed value and the fitted value. i ,y i)(i=1,2,…,n), in the linearly independent function group Find a function in Let the error sum of squares be:

[0089]

[0090] To reach the minimum,

[0091]

[0092] In order to distinguish the proportion of different data and make the theory more general, the least squares error is considered as a weighted sum of squares

[0093]

[0094] The above formula is composed of (m+1) variables a 0 ,a 1 ,···,a m The quadratic function

[0095]

[0096] According to the necessary conditions for the extreme value of a multivariate function, we have

[0097]

[0098] Introducing the inner product

[0099]

[0100] Formula (12) can be expressed as

[0101]

[0102] Rewrite it into matrix form

[0103] GA=d (16)

[0104] in

[0105] A=(a 0 ,a 1 ,…,a m ) T (17)

[0106] d=(d 0 ,d 1 ,…,d m ) T (18)

[0107]

[0108] because Linearly independent, so equation (16) has a unique solution

[0109]

[0110] So we get the function

[0111]

[0112] If i =1, yes

[0113]

[0114] Take ω i =1, and {1,x,x 2 ,…,x m} is used as the basis function for polynomial fitting, and the corresponding set of equations is

[0115]

[0116] Solve a 0 ,a 1 ,a 2 ,…,a m , we can get the fitting polynomial

[0117]

[0118] The least squares fitting in practical applications is polynomial fitting. By analyzing the above theoretical process, the algorithm flow of the least squares fitting in polynomial form can be summarized as follows:

[0119] (1) Determine the form of the fitting curve.

[0120] (2) Determine the variable data set (x i ,y i )(i=1,2,···n).

[0121] (3) Calculate each element of the coefficient matrix G and the tail matrix d according to formula (24).

[0122] (4) Solve the equation system GA=d and obtain the values ​​of the unknown coefficients.

[0123] Based on the above-mentioned cable conductor cross-sectional area detection method based on the improved Canny operator, the advanced image processing technology is used to perform non-contact measurement of the cable conductor cross-sectional area. The specific detection process is as follows:

[0124] (1) Improved Canny operator edge extraction: Use guided filtering instead of Gaussian filtering for edge-preserving smoothing and noise reduction. Gradient calculations are performed in four directions: horizontal, vertical, 45°, and 135°.

[0125] (2) Improve the CPDA global angle calculation algorithm: Use ring sampling to extract global angle related pixels. Use the least squares method to fit the two sides of the angle and calculate the angle size;

[0126] (3) Cable cross-sectional area detection: The improved Canny operator edge extraction and CPDA global angle calculation algorithm are combined to detect the areas of numerous cables in the cable trench.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A cable conductor cross-sectional area detection method based on an improved Canny operator, characterized in that: The method comprises the following steps: Improved Canny operator edge extraction: Guided filtering is used instead of Gaussian filtering for edge-preserving smoothing and noise reduction; based on the Canny operator's horizontal and vertical gradient calculation, two more direction detections of 45° and 135° are added; a local adaptive threshold is used instead of the global fixed method of the dual threshold of the non-maximum suppression technology to complete the edge detection of the cable; Improved CPDA global angle calculation algorithm: ring sampling to extract global angle related pixels; least square method to fit the angle bilaterally; calculate the angle size based on the fitted bilaterally; Cable cross-sectional area detection: Combined with the improved Canny operator edge extraction and CPDA global angle calculation algorithm, the area of ​​numerous cables in the cable trench is detected.

2. The cable conductor cross-sectional area detection method based on the improved Canny operator according to claim 1, characterized in that: The principle of guided filtering is shown in the following formula: Where i and j are the subscripts of the pixel points, I is the guide image; P j is the input cable image, Q i is the cable image output after filtering; W ij (I) Determine the weights used in the weighted average operation for the guide image I.

3. The cable conductor cross-sectional area detection method based on the improved Canny operator according to claim 1, characterized in that: The cable edge detection processing method of the local adaptive threshold is as follows: The 3×3 matrix is ​​used to move smoothly in the cable image, and the grayscale average value of the pixels in the 3×3 smoothing box is used as the adaptive threshold; Compare the adaptive threshold within the 3×3 smooth box with the gradient magnitude of the center pixel: If the gradient amplitude of the center point is greater than the adaptive threshold within the smoothing box, the point is an edge point and the output value is 1; If the gradient amplitude of the center point is less than or equal to 1 / 2 of the threshold, the point is not an edge point and the output value is 0; If the gradient amplitude of the center point is between 1 / 2 of the adaptive threshold and the entire adaptive threshold, its output value is the same as the output value of the previous center pixel; If the previous center pixel is an edge point, the output value is 1, otherwise the output value is 0.

4. The cable conductor cross-sectional area detection method based on the improved Canny operator according to claim 1, characterized in that: The gradient amplitude calculation formula is: The gradient direction θ(x,y) is: Where: (x, y) is the pixel point in the cable image; g x (x,y),g y (x,y)g 45o (x,y),g 135o (x, y) are the gradient components in the X, Y, 45° and 135° directions at the pixel point (x, y).

5. The cable conductor cross-sectional area detection method based on the improved Canny operator according to claim 1, characterized in that: The ring sampling radius is adjusted according to actual conditions.

6. The cable conductor cross-sectional area detection method based on the improved Canny operator according to claim 1, characterized in that: The algorithm flow of the least squares fitting is as follows: Determine the form of the fitted curve; Determine the dataset of the variables; Calculate the elements of the coefficient matrix and tail matrix as follows; Solve the system of equations to obtain the values ​​of the unknown coefficients.

Images