Non-contact wire diameter identification method and system

Abstract

The invention relates to the technical field of measurement and automation, in particular to a non-contact wire diameter identification method and system, and the method comprises the steps: enabling a CCD camera and a laser range finder to aim at a remote cable identification target, and obtaining a cable image and the spatial position information of the cable image; inputting the cable image into a trained cable type identification model for feature extraction and classification processing to obtain a cable type, and determining an edge detection preset parameter corresponding to the cable type; edge extraction and contour segmentation are carried out based on the cable image and edge detection preset parameters, and the pixel length of the cable diameter is obtained according to geometric characteristic estimation of the segmented contour; based on the spatial position information of the cable and the pixel length of the diameter of the cable, performing scale conversion from an image to a physical world to obtain the actual length of the diameter of the cable; and based on the actual length of the cable diameter and the cable type, searching a matching item in a preset cable model database to obtain a cable model identification result.

Description

Technical Field

[0001] The present invention relates to the field of measurement technology and automation technology, and in particular to a non-contact wire diameter recognition method and system. Background Art

[0002] Currently, numerous problems and difficulties arise in determining wire diameters during on-site surveys for 10kV distribution network design. This determination often requires the involvement of substation personnel or by consulting relevant line records. However, factors such as missing records, inaccurate and inaccurate records, and replacement of substation personnel create significant uncertainty. Therefore, on-site surveys are necessary to measure wire diameters and determine their sizes and types.

[0003] Currently, existing wire diameter measurement methods are divided into contact and non-contact methods. Contact measurement is further categorized into manual inspection and measurement using robotic inspection systems. Manual inspection requires the use of a handheld caliper, requiring workers to perform on-line inspections and measurements at height, which is extremely dangerous and results in very low measurement efficiency. While robotic inspection systems can achieve continuous measurement of specific areas, they require manual installation, present safety risks, and are unable to quickly and conveniently measure wire diameters. Non-contact measurement can be further categorized into aerial inspection and measurement using stereo vision. Aerial inspection primarily utilizes drones or helicopters flying low over the wire, using infrared cameras to measure and record conditions along the line. While this method can improve measurement efficiency to a certain extent, the aerial equipment is affected by its own flight speed, making detection accuracy unreliable. Furthermore, the aerial equipment suffers from poor endurance and high operating costs, and is significantly affected by factors such as forest topography and air turbulence. Visual inspection measurement methods currently mainly use the proportional relationship between laser ranging results and image pixels captured by CCD cameras to calculate the size of the target. However, existing wire diameter measurement technology can only achieve close-range measurement in a small range, and the machine is heavy and inconvenient to move, making it impossible to achieve fast and accurate long-distance measurement of wire diameter.

[0004] Therefore, considering the current development needs and the many problems encountered in actual operation, it is necessary to develop a convenient non-contact measurement method and system that can achieve long-distance and high-precision measurement to solve the problems of insufficient line diameter accuracy and inability to measure at long distances. Summary of the Invention

[0005] In order to solve the problems of insufficient accuracy of distribution line diameter and inability to measure at long distances, the present invention aims to provide a non-contact wire diameter identification method and system. The technical solutions adopted are as follows:

[0006] In a first aspect, the present application discloses a non-contact wire diameter identification method, the method comprising:

[0007] S1. Use a CCD camera and a laser rangefinder to identify the target cable at a distance and obtain the cable image and its spatial position information;

[0008] S2. Inputting the cable image into a trained cable type recognition model for feature extraction and classification processing to obtain the cable type and determine the edge detection preset parameters corresponding to the cable type;

[0009] S3, performing edge extraction and contour segmentation based on the cable image and the preset edge detection parameters, and estimating the pixel length of the cable diameter according to the geometric characteristics of the segmented contour;

[0010] S4, based on the spatial position information of the cable and the pixel length of the cable diameter, performing a scale conversion from the image to the physical world to obtain the actual length of the cable diameter;

[0011] S5. Based on the actual length of the cable diameter and the cable type, searching for a matching item in a preset cable model database to obtain a cable model identification result.

[0012] Furthermore, in step S1, when it is determined that there are multiple cables in the environment, in the process of using a CCD camera to aim at a long-distance cable to identify a target, the method further includes:

[0013] S11, obtaining an image captured by a camera, segmenting each cable in the image, and determining the image area where each cable is located based on the segmentation result;

[0014] S12, transmitting the image captured by the camera to a display page of the user terminal, and obtaining clicked area information when it is determined that the user clicks a corresponding area on the display page for a target cable to be identified;

[0015] S13, associating and matching the click area information with the image area where each cable is located to locate the target cable;

[0016] S14. Acquire the three-dimensional coordinate information of the target cable, and adjust the rotation angle of the camera platform based on the three-dimensional coordinate information so that the target cable is in the visual center of the monitoring screen.

[0017] Furthermore, in the process of using the CCD camera to aim at the long-distance cable to identify the target, the method further includes:

[0018] S15, based on the perspective projection principle, converting the three-dimensional coordinates of the target cable relative to the camera into image coordinates;

[0019] S16, obtaining the horizontal deviation and vertical deviation of the target cable relative to the image center during its movement, and adjusting the rotation angle of the camera pan / tilt platform based on the horizontal deviation and vertical deviation through a feedback control principle so that the target cable always remains in the visual center of the monitoring image;

[0020] S17. Fine-tune the focus based on the actual distance between the target cable and the camera so that the target cable is displayed more clearly in the center of the image.

[0021] Furthermore, in step S3, edge extraction and contour segmentation are performed based on the cable image and the edge detection preset parameters, including:

[0022] S31, performing edge extraction based on the cable image and the edge detection preset parameters to determine the cable contour;

[0023] S32 , performing contour segmentation according to shape features of the contour curve and interactive information with image edges to determine a long side sub-contour and a short side sub-contour of the cable.

[0024] Furthermore, in step S31, for edge extraction in a complex environment, the method includes:

[0025] S311, preprocessing the cable image to obtain a preprocessed image;

[0026] S312, performing adaptive histogram equalization processing on the pre-processed image, performing histogram equalization on a local area of the image to enhance local contrast, thereby obtaining a contrast-enhanced image;

[0027] S313 : Based on the contrast-enhanced image, call the edge detection preset parameters to perform Canny edge detection to extract clear, continuous and complete edges.

[0028] Furthermore, in step S3, estimating the pixel length of the cable diameter according to the geometric characteristics of the segmented contour includes:

[0029] S33, sorting the two long-side sub-contours to determine a first sub-contour that is the longest and a second sub-contour that is the second longest;

[0030] S34, dividing the second sub-contour into equal parts to obtain a plurality of equal division points;

[0031] S35. For each equally divided point, calculate the normal direction and the fitting of the normal line segment according to the slope of the fitted line at the equally divided point;

[0032] S36. For each equally divided point, determine the intersection point between the corresponding normal line segment and the first sub-contour, and calculate the Euclidean distance between the equally divided point and the intersection point to obtain a distance set;

[0033] S37. For each equally divided point, calculate its minimum Euclidean distance to the first sub-contour, and compare the minimum Euclidean distance with each Euclidean distance obtained from the normal line segment. Distance items whose difference exceeds a preset threshold are removed from the distance set.

[0034] S38. Combining the distance values in the distance set, obtain the pixel length of the cable diameter.

[0035] Furthermore, in step S4, the image is converted to the physical world based on the spatial position information of the cable and the pixel length of the cable diameter to obtain the actual length of the cable diameter, including:

[0036] S41, determining a shooting distance between the cable and a CCD camera based on the spatial position information;

[0037] S42, obtaining the focal length of the CCD camera, and determining the magnification factor based on the ratio between the focal length and the shooting distance;

[0038] S43. Obtain the actual physical size corresponding to each pixel in the image, and perform scale conversion from the image to the physical world based on the pixel length of the cable diameter, the actual physical size, and the magnification factor to obtain the actual length of the cable diameter.

[0039] In a second aspect, the present application discloses a non-contact wire diameter identification system, the system comprising an image acquisition and positioning module, a cable type identification module, a cable diameter pixel length estimation module, a cable diameter actual length estimation module, and a cable model matching module, wherein:

[0040] The image acquisition and positioning module is used to use a CCD camera and a laser rangefinder to align the long-distance cable identification target to obtain the cable image and its spatial position information;

[0041] The cable type recognition module is configured to input the cable image into a trained cable type recognition model for feature extraction and classification processing to obtain the cable type and determine the edge detection preset parameters corresponding to the cable type;

[0042] The cable diameter pixel length estimation module is used to perform edge extraction and contour segmentation based on the cable image and the preset edge detection parameters, and estimate the pixel length of the cable diameter according to the geometric characteristics of the segmented contour;

[0043] The cable diameter actual length estimation module is used to perform image-to-physical scale conversion based on the spatial position information of the cable and the pixel length of the cable diameter to obtain the actual length of the cable diameter;

[0044] The cable model matching module is configured to search for matching items in a preset cable model database based on the actual length of the cable diameter and the cable type, and obtain a cable model identification result.

[0045] The present invention has the following beneficial effects:

[0046] 1) By combining a CCD camera and a laser rangefinder, it is possible to accurately capture images of cables and their spatial position information at long distances. The laser rangefinder provides precise spatial positioning information, while the CCD camera captures high-quality cable images. The combination of the two enables high-precision measurement at long distances.

[0047] 2) By extracting and segmenting the cable image's edges and estimating the geometric properties of the segmented contours, the cable diameter can be accurately calculated as pixel length in the image. Subsequently, combining the cable's spatial position information with laser ranging data enables an accurate conversion from image pixels to the physical world scale, thereby determining the actual cable diameter.

[0048] 3) Based on the actual cable diameter and cable type, a matching item is searched in the preset cable model database to obtain the cable model identification result. This method ensures the accuracy of cable model identification and provides reliable data support for subsequent distribution network design. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A flow chart of a non-contact wire diameter recognition method provided by one embodiment of the present invention;

[0050] Figure 2 This is a system structure diagram of a non-contact wire diameter recognition system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0051] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0052] like Figure 1 , which shows a flow chart of a non-contact wire diameter recognition method provided by one embodiment of the present invention, the method comprising:

[0053] Step S1: Use a CCD camera and a laser rangefinder to aim at a long-distance cable to identify a target, and obtain a cable image and its spatial position information.

[0054] Specifically, a CCD camera captures a high-resolution image of the cable, which includes information such as its shape and size. A laser rangefinder emits a laser beam at the cable and measures the time it takes for the laser beam to reflect back, thereby calculating the cable's spatial position relative to the laser rangefinder.

[0055] Step S2: Input the cable image into a trained cable type recognition model for feature extraction and classification processing to obtain the cable type and determine the edge detection preset parameters corresponding to the cable type.

[0056] Specifically, this application pre-trained the cable type recognition model using a large number of cable images. To ensure the diversity of the data set, this application collected images of physical cables from multiple angles (such as the front, side, and oblique directions). These images can more comprehensively display the shape, texture, and detailed features of the cables, helping the model learn a richer set of cable type features. In addition, to reduce the impact of changes in light intensity on cable characteristics, this application also collected cable image datasets taken in sunny and cloudy weather to increase the diversity of the dataset, enabling the model to adapt to cable images under different lighting conditions and avoid misjudgments due to cable shadows.

[0057] Step S3: performing edge extraction and contour segmentation based on the cable image and the preset edge detection parameters, and estimating the pixel length of the cable diameter according to the geometric characteristics of the segmented contour.

[0058] Specifically, this application uses edge detection algorithms from image processing technology to process cable images to extract the cable edges. Then, based on the extracted edges and the shape characteristics of the contour, contour segmentation is performed to obtain two short and two long sub-contours. The two long sub-contours correspond to the longitudinal edges of the cable, and the two short sub-contours correspond to the transverse edges of the cable.

[0059] Specifically, for the two extracted long-side sub-contours and two short-side sub-contours, this application will first sort the two long-side sub-contours to determine the primary and secondary, and then divide the second sub-contour based on the secondary length into equal parts and calculate the normal direction and normal segment fitting of each equal-division point. Then, find the intersection of each normal segment and the first sub-contour of the primary length, and calculate the Euclidean distance between the equal-division point and the intersection to form a distance set. Finally, combine these distance values to obtain the pixel length of the cable diameter.

[0060] In one embodiment, before calculating the pixel length, the method further includes: calculating, for each equally divided point A, the minimum Euclidean distance B from the equally divided point A to the first sub-contour; comparing the minimum Euclidean distance B with each Euclidean distance C calculated by the normal line segment, wherein when it is determined that the difference in distance exceeds a preset difference threshold, the corresponding Euclidean distance C is removed from the distance set.

[0061] Step S4: Based on the spatial position information of the cable and the pixel length of the cable diameter, scale conversion is performed from the image to the physical world to obtain the actual length of the cable diameter.

[0062] Specifically, this application will determine the shooting distance between the cable and the CCD camera through spatial position information, and then calculate the magnification factor through the camera focal length. Combined with the actual physical size of each pixel in the image, the pixel length, actual physical size and magnification factor of the cable diameter are used for scale conversion to finally obtain the actual length of the cable diameter.

[0063] Step S5: Based on the actual length of the cable diameter and the cable type, a matching item is searched in a preset cable model database to obtain a cable model identification result.

[0064] Specifically, this application sets query conditions based on the actual cable diameter and cable type, and constructs a corresponding query request. This query request is then sent to a pre-set cable model database, which uses a built-in matching algorithm to search for items matching the query conditions within the stored cable model data. Once a matching cable model is found, the database returns the corresponding cable model identification result.

[0065] As can be seen from the above, the present application discloses a non-contact wire diameter identification method, which, by combining a CCD camera and a laser rangefinder, enables accurate capture of cable images and their spatial position information at a long distance. Among them, the laser rangefinder can provide accurate spatial positioning information, while the CCD camera captures high-quality cable images. The combination of the two enables high-precision measurement at a long distance; through edge extraction and contour segmentation of the cable image, as well as estimation of geometric characteristics based on the segmented contour, the pixel length of the cable diameter in the image can be accurately calculated. Subsequently, by combining the spatial position information of the cable and the laser ranging data, an accurate conversion from image pixels to the scale of the physical world is achieved, thereby obtaining the actual length of the cable diameter; based on the actual length of the cable diameter and the cable type, matching items are searched in the preset cable model database, and finally the cable model identification result is obtained. This method ensures the accuracy of cable model identification and provides reliable data support for subsequent distribution network design.

[0066] In one embodiment, in step S1, when it is determined that there are multiple cables in the environment, in the process of using a CCD camera to aim at a long-distance cable to identify a target, the method further includes:

[0067] Step S11: acquiring an image captured by a camera, segmenting the cables in the image, and determining the image area where the cables are located based on the segmentation results.

[0068] Specifically, this application segments cables within an image by analyzing color, texture, and edge features. After segmentation, connected region analysis is performed on each connected region to determine the image region where each cable resides. This connected region analysis identifies interconnected sets of pixels within the image to determine the specific location and range of each cable.

[0069] Step S12: transmitting the camera-photographed image to a display page of the user terminal, and obtaining clicked area information when it is determined that the user has clicked a corresponding area on the display page for a target cable to be identified.

[0070] Specifically, this application displays the camera image in real time on the user's display page. Subsequently, the application monitors the user's clicks on the display page. When the user clearly indicates the target cable to be identified and clicks the corresponding area on the image, the application immediately captures and records this click event. Subsequently, through the interactive interface feedback mechanism, the specific location information of the user's click (i.e., the click area information) is analyzed based on the click event.

[0071] Step S13: Associating and matching the click area information with the image area where each cable is located to locate the target cable.

[0072] Specifically, the present application will search for the best matching area in each cable area previously segmented according to the click area information provided by the user, thereby determining the location of the target cable.

[0073] Step S14: acquiring the three-dimensional coordinate information of the target cable, and adjusting the rotation angle of the camera platform based on the three-dimensional coordinate information so that the target cable is in the visual center of the monitoring screen.

[0074] Specifically, this application will adjust the pitch angle θ of the camera gimbal based on the following formula: tilt and yaw angle θ pan :

[0075]

[0076] Among them, X, Y, Z are the three-dimensional coordinates of the target cable relative to the camera, and the pitch angle θ tilt and yaw angle θpan Indicates the vertical and horizontal rotation angles of the camera pan / tilt, used to adjust the camera's viewing angle so that the target cable is in the visual center of the monitoring image.

[0077] In one embodiment, in the process of using a CCD camera to identify a target at a long distance cable, the method further includes:

[0078] Step S15: Based on the perspective projection principle, the three-dimensional coordinates of the target cable relative to the camera are converted into image coordinates.

[0079] Specifically, for the three-dimensional coordinates (X, Y, Z) of the target cable relative to the camera, this application converts them into image coordinates (u, v) through the perspective projection principle. For details, please refer to the following formula:

[0080]

[0081] Where f is the focal length of the camera, which determines the size of the point projected onto the image plane; u0 and v0 are the coordinates of the center of the image, which are equal to half the width and height of the image.

[0082] It should be noted that the X and Y components of the 3D coordinates are divided by the Z component (i.e., depth information), and then scaled and translated to obtain image coordinates. This process takes into account the focal length of the camera and the position of the image center, thus accurately converting the 3D coordinates to image coordinates.

[0083] Step S16, obtaining the horizontal deviation and vertical deviation of the target cable relative to the image center during its movement, and adjusting the rotation angle of the camera pan / tilt based on the horizontal deviation and vertical deviation through the feedback control principle, so that the target cable always remains in the visual center of the monitoring screen.

[0084] Specifically, feedback control is a method of continuously detecting deviations and adjusting the system state accordingly to achieve a predetermined goal. Specifically, during implementation, this application adjusts the rotation angle of the camera gimbal by calculating the horizontal deviation u1 and the vertical deviation v1 using the following formula:

[0085] θ pan_new =θ pan +K pan *u1;

[0086] θ tilt_new =θ tilt +K tilt *v1;

[0087] Among them, K pan Indicates the adjustment gain of the yaw angle, K tilt Indicates the adjustment gain of the pitch angle, θ panIndicates the horizontal rotation angle of the camera gimbal before adjustment, θ tilt This indicates the vertical rotation angle of the camera gimbal before adjustment. It should be noted that the gain adjustment controls the amount of gimbal rotation adjustment and can be adjusted based on actual conditions for more accurate target tracking. Currently, by controlling the gimbal rotation angle, the camera's viewing angle is further altered, thereby adjusting the target cable's position in the image.

[0088] Step S17: fine-tuning the focus based on the actual distance between the target cable and the camera so that the target cable is displayed more clearly in the center of the image.

[0089] Specifically, zoom control is mainly based on the actual distance between the target cable and the camera to dynamically adjust the focal length. The change in focal length can be controlled by the following formula:

[0090]

[0091] Among them, f new represents the new focal length after fine-tuning, f0 represents the initial focal length of the camera, D represents the actual distance between the target cable and the camera, which can be obtained through the depth sensor, and Z represents the initial Z coordinate of the target in the camera coordinate system, that is, the distance between the target and the camera before the gimbal adjusts the target cable to the visual center of the image, which is also obtained through the depth sensor.

[0092] In one embodiment, in step S3, performing edge extraction and contour segmentation based on the cable image and the edge detection preset parameters includes:

[0093] Step S31 : performing edge extraction based on the cable image and the preset edge detection parameters to determine the cable outline.

[0094] Specifically, this application will perform preprocessing operations such as grayscale, filtering, and binarization on the acquired cable image; then, for the obtained preprocessed image, this application uses the FindContours function in opencv to perform contour edge extraction to determine the cable contour, where the cable contour consists of a set of sequentially arranged points.

[0095] Step S32 : performing contour segmentation based on the shape features of the contour curve and the interactive information with the image edge to determine the long side sub-contour and the short side sub-contour of the cable.

[0096] Specifically, the application calculates the curvature of each point in the contour curve and determines the shape characteristics of the contour based on the distribution and variation of the obtained curvature values. In practice, for each point in the contour curve, the application uses a polynomial to perform curve fitting (the choice of polynomial depends on the complexity of the contour curve); then, through analytical differentiation, that is, based on the second-order derivative of the fitted curve divided by the square root of the first-order derivative plus 1 cubed, the curvature of each point is calculated.

[0097] Furthermore, when performing contour segmentation, the present application first determines whether the starting and ending points of the contour are located at the edge of the image. If so, it is considered a closed contour. Then, for closed contours, based on the shape characteristics of the contour curve and the interactive information with the image edge, a geometric feature-based segmentation algorithm is used to perform contour segmentation when it is determined that there are two obvious long-side regions and two short-side regions on the contour. The segmentation algorithm will identify these regions on the contour and split the contour into two long-side sub-contours and two short-side sub-contours at the intersection of these regions.

[0098] In one embodiment, in step S31, for edge extraction in a complex environment, the method includes:

[0099] Step S311: pre-process the cable image to obtain a pre-processed image.

[0100] Specifically, during the preprocessing process, this application performs denoising on the cable image (usually using Gaussian filtering or median filtering) to remove noise in the image, laying a good data foundation for subsequent contrast enhancement and edge detection. Furthermore, this application also grayscales the denoised image to obtain a preprocessed grayscale image.

[0101] Step S312 : performing adaptive histogram equalization processing on the pre-processed image, and performing histogram equalization on a local area of the image to enhance the local contrast, thereby obtaining a contrast-enhanced image.

[0102] Specifically, this application performs adaptive histogram equalization on the pre-processed grayscale image, setting an appropriate block size (e.g., 8x8 or 16x16) and clipping limits to enhance the local contrast in the image. It should be noted that the application of adaptive histogram equalization can make the image have higher contrast in areas with uneven lighting, thereby improving the effect of subsequent Canny edge detection.

[0103] Step S313 : Based on the contrast-enhanced image, the preset edge detection parameters are called to perform Canny edge detection to extract clear, continuous and complete edges.

[0104] In one embodiment, in step S3, estimating the pixel length of the cable diameter according to the geometric characteristics of the segmented contour includes:

[0105] Step S33 , sorting the two long-side sub-contours to determine a first sub-contour that is the longest and a second sub-contour that is the second longest.

[0106] Step S34: divide the second sub-contour into equal parts to obtain a plurality of equal division points.

[0107] Specifically, bisection refers to evenly dividing the second subcontour into several segments, with the endpoints of each segment serving as bisection points. This can be achieved by calculating the total length of the second subcontour and dividing it by the desired number of bisections. Then, starting from the start or end point of the second subcontour, these bisection points are gradually found along the curve of the subcontour, based on the proportional position of each bisection point on the line segment.

[0108] Step S35 : For each equally divided point, the normal direction and the fitting of the normal line segment are calculated according to the slope of the fitted line at the equally divided point.

[0109] Specifically, for each equally divided point, this application will fit a local straight line (specifically, this can be achieved through linear regression) and calculate the slope of this fitted line. Considering that the normal direction is the perpendicular direction of this fitted line, the slope of the normal can be determined by calculating the negative reciprocal of the slope of the fitted line, and based on this, a direction vector perpendicular to the fitted line is constructed as the normal direction.

[0110] Furthermore, in the process of fitting the normal line segment, it is first necessary to select a starting point, which is usually the bisection point itself. Then, the end point is determined based on the normal direction and the preset initial line segment length.

[0111] In one embodiment, when fitting a normal line segment based on the start and end coordinates, a scaling factor can be set to adjust the initial line segment length based on the scaling factor to ensure an appropriate line segment length. The scaling factor directly takes the larger value of the image pixel length and width.

[0112] Step S36: for each equally divided point, determine the intersection point between the corresponding normal line segment and the first sub-contour, and calculate the Euclidean distance between the equally divided point and the intersection point to obtain a distance set.

[0113] Specifically, for each equally divided point, the present application uses a function in a mathematical software package to find and record the coordinates of the intersection between the corresponding normal line segment and the first sub-contour, and then calculates the Euclidean distance based on the Euclidean distance formula (i.e., for two points P1(x1, y1) and P2(x2, y2), the Euclidean distance between the two points is: Calculate the Euclidean distance between each equally divided point and the corresponding intersection point. Finally, store these calculated distance values into a set to obtain a distance set.

[0114] Step S37: for each equally divided point, calculate its minimum Euclidean distance to the first sub-contour, and compare it with the Euclidean distances obtained according to the normal line segments, wherein the distance items whose difference exceeds the preset threshold will be removed from the distance set.

[0115] Step S38: synthesize the distance values in the distance set to obtain the pixel length of the cable diameter.

[0116] Specifically, when integrating the distance values in the distance set, the average value, median value, weighted average value, etc. can be considered. The specific calculation method can be comprehensively considered based on the nature of the data, the accuracy and reliability requirements of the cable diameter estimation, and the actual application scenario.

[0117] In one embodiment, in step S4, performing scale conversion from the image to the physical world based on the spatial position information of the cable and the pixel length of the cable diameter to obtain the actual length of the cable diameter includes:

[0118] Step S41: determining a shooting distance between the cable and the CCD camera based on the spatial position information.

[0119] Specifically, the present application considers placing a laser rangefinder near the CCD camera, emitting laser light through the laser rangefinder and measuring the time it takes for the laser light to be reflected back, so as to obtain the shooting distance between the cable and the CCD camera.

[0120] Step S42: Obtain the focal length of the CCD camera, and determine the magnification factor based on the ratio between the focal length and the shooting distance.

[0121] Specifically, the magnification factor is equal to the focal length divided by the shooting distance, which represents the proportional relationship between the size of the object in the image and the actual size of the object.

[0122] Step S43: Obtain the actual physical size corresponding to each pixel in the image, and perform scale conversion from the image to the physical world based on the pixel length, actual physical size, and magnification factor of the cable diameter to obtain the actual length of the cable diameter.

[0123] Specifically, this application will multiply the pixel length of the cable diameter by the pixel size (i.e., the actual physical size) to obtain a preliminary physical estimate of the cable diameter in the image. Then, it will be multiplied by the magnification factor to correct the image magnification or reduction effect caused by the shooting distance and focal length, thereby obtaining the actual length of the cable diameter.

[0124] Please refer to Figure 2 The present application discloses a non-contact wire diameter recognition system, which includes an image acquisition and positioning module, a cable type recognition module, a cable diameter pixel length estimation module, a cable diameter actual length estimation module, and a cable model matching module, wherein:

[0125] The image acquisition and positioning module is used to use a CCD camera and a laser rangefinder to align with a long-distance cable identification target to obtain a cable image and its spatial position information.

[0126] The cable type recognition module is used to input the cable image into a trained cable type recognition model to perform feature extraction and classification processing to obtain the cable type and determine the edge detection preset parameters corresponding to the cable type.

[0127] The cable diameter pixel length estimation module is used to perform edge extraction and contour segmentation based on the cable image and the edge detection preset parameters, and estimate the pixel length of the cable diameter according to the geometric characteristics of the segmented contour.

[0128] The cable diameter actual length estimation module is used to perform scale conversion from the image to the physical world based on the spatial position information of the cable and the pixel length of the cable diameter to obtain the actual length of the cable diameter.

[0129] The cable model matching module is configured to search for matching items in a preset cable model database based on the actual length of the cable diameter and the cable type, and obtain a cable model identification result.

[0130] In one embodiment, the above modules are also used to implement the non-contact wire diameter identification method as described in any one of the above method embodiments, which is not limited in this application.

[0131] As can be seen from the above, the non-contact wire diameter identification system disclosed in the present application, by combining a CCD camera and a laser rangefinder, can accurately capture the image of the cable and its spatial position information at a long distance. Among them, the laser rangefinder can provide accurate spatial positioning information, while the CCD camera captures high-quality cable images. The combination of the two realizes high-precision measurement at a long distance; through the edge extraction and contour segmentation of the cable image, as well as the estimation of the geometric characteristics based on the segmented contour, the pixel length of the cable diameter in the image can be accurately calculated. Subsequently, by combining the spatial position information of the cable and the laser ranging data, an accurate conversion from image pixels to the scale of the physical world is achieved, thereby obtaining the actual length of the cable diameter; based on the actual length of the cable diameter and the cable type, matching items are searched in the preset cable model database, and finally the cable model identification result is obtained. This method ensures the accuracy of cable model recognition and provides reliable data support for subsequent distribution network design.

[0132] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0133] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A non-contact wire diameter recognition method, characterized in that: The method comprises: S1. Use a CCD camera and a laser rangefinder to identify the target cable at a distance and obtain the cable image and its spatial position information; S2. Inputting the cable image into a trained cable type recognition model for feature extraction and classification processing to obtain the cable type and determine the edge detection preset parameters corresponding to the cable type; S3, performing edge extraction and contour segmentation based on the cable image and the preset edge detection parameters, and estimating the pixel length of the cable diameter according to the geometric characteristics of the segmented contour; S4, based on the spatial position information of the cable and the pixel length of the cable diameter, performing a scale conversion from the image to the physical world to obtain the actual length of the cable diameter; S5. Based on the actual length of the cable diameter and the cable type, searching for a matching item in a preset cable model database to obtain a cable model identification result.

2. The method according to claim 1, characterized in that In step S1, when it is determined that there are multiple cables in the environment, in the process of using a CCD camera to aim at a long-distance cable to identify a target, the method further includes: S11, obtaining an image captured by a camera, segmenting each cable in the image, and determining the image area where each cable is located based on the segmentation result; S12, transmitting the image captured by the camera to a display page of the user terminal, and obtaining clicked area information when it is determined that the user clicks a corresponding area on the display page for a target cable to be identified; S13, associating and matching the click area information with the image area where each cable is located to locate the target cable; S14. Acquire the three-dimensional coordinate information of the target cable, and adjust the rotation angle of the camera platform based on the three-dimensional coordinate information so that the target cable is in the visual center of the monitoring screen.

3. The method according to claim 2, characterized in that In the process of using a CCD camera to aim at a long-distance cable to identify a target, the method further includes: S15, based on the perspective projection principle, converting the three-dimensional coordinates of the target cable relative to the camera into image coordinates; S16, obtaining the horizontal deviation and vertical deviation of the target cable relative to the image center during its movement, and adjusting the rotation angle of the camera pan / tilt platform based on the horizontal deviation and vertical deviation through a feedback control principle so that the target cable always remains in the visual center of the monitoring image; S17. Fine-tune the focus based on the actual distance between the target cable and the camera so that the target cable is displayed more clearly in the center of the image.

4. The method according to claim 1, wherein In step S3, edge extraction and contour segmentation are performed based on the cable image and the edge detection preset parameters, including: S31, performing edge extraction based on the cable image and the edge detection preset parameters to determine the cable contour; S32 , performing contour segmentation according to shape features of the contour curve and interactive information with image edges to determine a long side sub-contour and a short side sub-contour of the cable.

5. The method according to claim 4, characterized in that In step S31, for edge extraction in a complex environment, the method includes: S311, preprocessing the cable image to obtain a preprocessed image; S312, performing adaptive histogram equalization processing on the pre-processed image, performing histogram equalization on a local area of the image to enhance local contrast, thereby obtaining a contrast-enhanced image; S313 : Based on the contrast-enhanced image, call the edge detection preset parameters to perform Canny edge detection to extract clear, continuous and complete edges.

6. The method according to claim 4, characterized in that In step S3, the pixel length of the cable diameter is obtained by estimating according to the geometric characteristics of the segmented contour, including: S33, sorting the two long-side sub-contours to determine a first sub-contour that is the longest and a second sub-contour that is the second longest; S34, dividing the second sub-contour into equal parts to obtain a plurality of equal division points; S35. For each equally divided point, calculate the normal direction and the fitting of the normal line segment according to the slope of the fitted line at the equally divided point; S36. For each equally divided point, determine the intersection point between the corresponding normal line segment and the first sub-contour, and calculate the Euclidean distance between the equally divided point and the intersection point to obtain a distance set; S37. For each equally divided point, calculate its minimum Euclidean distance to the first sub-contour, and compare the minimum Euclidean distance with each Euclidean distance obtained from the normal line segment. Distance items whose difference exceeds a preset threshold are removed from the distance set. S38. Combining the distance values in the distance set, obtain the pixel length of the cable diameter.

7. The method according to claim 1, characterized in that In step S4, the scale conversion from the image to the physical world is performed based on the spatial position information of the cable and the pixel length of the cable diameter to obtain the actual length of the cable diameter, including: S41, determining a shooting distance between the cable and a CCD camera based on the spatial position information; S42, obtaining the focal length of the CCD camera, and determining the magnification factor based on the ratio between the focal length and the shooting distance; S43. Obtain the actual physical size corresponding to each pixel in the image, and perform scale conversion from the image to the physical world based on the pixel length of the cable diameter, the actual physical size, and the magnification factor to obtain the actual length of the cable diameter.

8. A non-contact wire diameter recognition system, characterized in that: The system includes an image acquisition and positioning module, a cable type identification module, a cable diameter pixel length estimation module, a cable diameter actual length estimation module, and a cable model matching module, wherein: The image acquisition and positioning module is used to use a CCD camera and a laser rangefinder to align the long-distance cable identification target to obtain the cable image and its spatial position information; The cable type recognition module is configured to input the cable image into a trained cable type recognition model for feature extraction and classification processing to obtain the cable type and determine the edge detection preset parameters corresponding to the cable type; The cable diameter pixel length estimation module is used to perform edge extraction and contour segmentation based on the cable image and the preset edge detection parameters, and estimate the pixel length of the cable diameter according to the geometric characteristics of the segmented contour; The cable diameter actual length estimation module is used to perform image-to-physical scale conversion based on the spatial position information of the cable and the pixel length of the cable diameter to obtain the actual length of the cable diameter; The cable model matching module is configured to search for matching items in a preset cable model database based on the actual length of the cable diameter and the cable type, and obtain a cable model identification result.

Images