A method and device for analyzing wire icing based on image processing and a storage medium

Abstract

The application provides a kind of based on image processing's wire icing analysis method, device and storage medium, the method comprises: obtaining the longitudinal pixel cumulative histogram of the wire icing image to be analyzed;The longitudinal pixel cumulative histogram is input into the trained convolutional neural network, to obtain the distance ratio threshold set and the pixel number threshold set output by the trained convolutional neural network;According to the longitudinal pixel cumulative histogram, the distance ratio threshold set, the pixel number threshold set and the diameter of the wire, the icing class label and the icing thickness in the wire icing image to be analyzed are obtained. The scheme of the present application can automatically and accurately analyze the category and thickness of wire icing by computer equipment, providing a simple and practical analysis method for users.

Description

Technical Field

[0001] This application relates to the field of image recognition technology, and in particular to a method, apparatus, storage medium and computer equipment for analyzing wire icing based on image processing. Background Technology

[0002] In cold regions, icing of transmission lines is one of the major threats to the safe operation of power systems. Icing increases the weight of conductors and wind resistance, leading to accidents such as conductor breakage and tower collapse, causing large-scale power outages. Traditional icing detection methods mainly rely on manual inspections or sensor monitoring. However, manual inspections are inefficient, costly, and difficult to cover remote areas; sensor monitoring is easily affected by the environment, difficult to maintain, and has limited ability to distinguish between different icing types such as mixed frost, rime, and hoarfrost. Therefore, there is an urgent need for a solution that can automatically analyze the type of icing on transmission lines. Summary of the Invention

[0003] The purpose of this application is to address at least one of the aforementioned technical deficiencies, particularly the technical deficiency that the existing technology cannot automatically analyze the type of icing on conductors.

[0004] In a first aspect, embodiments of this application provide a method for analyzing conductor icing based on image processing, the method comprising:

[0005] Obtain the longitudinal pixel cumulative histogram of the icing image of the conductor to be analyzed;

[0006] The vertical pixel cumulative histogram is input into the trained convolutional neural network to obtain the distance ratio threshold set and the pixel number threshold set output by the trained convolutional neural network.

[0007] Based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, the pixel number threshold set, and the conductor diameter, an ice category label and ice thickness are obtained to reflect the ice accretion image of the conductor to be analyzed.

[0008] In one embodiment, the set of pixel number thresholds includes a first pixel number threshold and a second pixel number threshold for determining the position of the conductor and the position of the icing in the conductor icing image to be analyzed;

[0009] The step of obtaining the icing category label reflecting the icing image of the conductor to be analyzed based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, and the pixel number threshold set includes:

[0010] In the vertical pixel cumulative histogram, the x-coordinates of points where the number of pixels is greater than the first pixel number threshold are located. The point with the smallest x-coordinate and the point with the largest x-coordinate are the first endpoint and the second endpoint of the conductor, respectively. The distance between the first endpoint and the second endpoint of the conductor is a first distance. In the vertical pixel cumulative histogram, the x-coordinates of points where the number of pixels is greater than the second pixel number threshold are located. The point with the smallest x-coordinate and the point with the largest x-coordinate are the first endpoint and the second endpoint of the ice-covered conductor, respectively. The distance between the first endpoint of the ice-covered conductor and the first endpoint of the conductor is a second distance, which is less than the distance between the first endpoint of the ice-covered conductor and the second endpoint of the conductor. The distance between the second endpoint of the ice-covered conductor and the second endpoint of the conductor is a third distance, which is less than the distance between the second endpoint of the ice-covered conductor and the first endpoint of the conductor.

[0011] The icing category label is determined based on the first distance, the second distance, and the third distance.

[0012] In one embodiment, the step of determining the icing category label based on the first distance, the second distance, and the third distance includes:

[0013] The distance ratio threshold set includes a first distance ratio threshold and a second distance ratio threshold for determining the conductor position and icing position in the icing image of the conductor to be analyzed;

[0014] The ratio of the second distance to the first distance is calculated as the first ratio, the ratio of the third distance to the first distance is calculated as the second ratio, and the absolute value of the difference between the first ratio and the second ratio is calculated as the first absolute value. If the first absolute value is greater than the first distance ratio threshold, the first label reflecting the icing category of the icing image of the conductor to be analyzed as mixed rime is used as the icing category label.

[0015] If the first absolute value is less than the first distance ratio threshold and the first ratio is greater than the second distance ratio threshold, then the second label, which reflects the icing category of the icing image of the conductor to be analyzed as rime ice, will be used as the icing category label.

[0016] If the first absolute value is less than the first distance ratio threshold and the first ratio is less than the second distance ratio threshold, then the third label, which reflects the icing category of the icing image of the conductor to be analyzed as rime, will be used as the icing category label.

[0017] In one embodiment, the step of obtaining an ice accretion category label reflecting the ice accretion image of the conductor to be analyzed based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, and the pixel number threshold set further includes:

[0018] If the first end point of the icing result is the same as the first end point or the second end point of the conductor, and the second end point of the icing result is the same as the first end point or the second end point of the conductor, then the fourth label, which reflects the icing category of the icing image of the conductor to be analyzed as no icing, will be used as the icing category label.

[0019] In one embodiment, the set of pixel number thresholds includes a first pixel number threshold for determining the position of the conductor and a second pixel number threshold for determining the icing position;

[0020] The step of obtaining the icing category label reflecting the icing image of the conductor to be analyzed based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, and the pixel number threshold set includes:

[0021] The vertical pixel cumulative histogram, the first distance ratio threshold, the second distance ratio threshold, the first pixel number threshold, and the second pixel number threshold are input into the trained decision tree model to obtain the icing category label output by the trained decision tree model.

[0022] In one embodiment, the icing thickness is calculated based on the first ratio, the second ratio, and the diameter of the conductor.

[0023] In one embodiment, before the step of obtaining the longitudinal pixel cumulative histogram of the icing image of the conductor to be analyzed, the method includes: obtaining an original icing image of the conductor, wherein the original icing image of the conductor is an image recording the icing to be analyzed.

[0024] The original conductor icing image is binarized to identify the conductor region in the original conductor icing image, and the original conductor icing image is segmented based on the conductor region to obtain a segmented conductor icing image.

[0025] Based on the angle between the conductor and the vertical direction in the segmented conductor icing image, the segmented conductor icing image is corrected to obtain the conductor icing image to be analyzed.

[0026] Secondly, embodiments of this application provide an apparatus for analyzing conductor icing based on image processing, characterized in that it includes:

[0027] The histogram acquisition module is used to acquire the longitudinal pixel cumulative histogram of the icing image of the conductor to be analyzed;

[0028] The threshold acquisition module is used to input the vertical pixel cumulative histogram into the trained convolutional neural network to obtain the distance ratio threshold set and the pixel number threshold set output by the trained convolutional neural network.

[0029] The icing category label and icing thickness acquisition module is used to obtain icing category labels and icing thickness in the icing image of the conductor to be analyzed based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, the pixel number threshold set, and the conductor diameter.

[0030] Thirdly, embodiments of this application provide a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the image processing-based conductor icing analysis method described in any of the above embodiments.

[0031] Fourthly, embodiments of this application provide a computer device, which includes: one or more processors, and a memory;

[0032] The memory stores computer-readable instructions, which, when executed by the one or more processors, perform the steps of the image processing-based conductor icing analysis method described in any of the above embodiments.

[0033] In the image processing-based method, apparatus, storage medium, and computer device provided in this application embodiment for conductor icing analysis, considering that mixed rime ice may exhibit asymmetrical icing on both sides, while frost and hoarfrost have relatively symmetrical icing but different thicknesses, this application can utilize this characteristic to automatically perform conductor icing category analysis. The computer device can input the vertical pixel cumulative histogram of the conductor icing image to be analyzed into a trained convolutional neural network to obtain a distance ratio threshold set and a pixel number threshold set output by the trained convolutional neural network. The computer device can determine the icing category of the conductor icing to be analyzed based on the vertical pixel cumulative histogram and the distance ratio thresholds. Thus, the computer device can automatically and accurately analyze the icing category of conductors, providing users with a simple and practical analysis method. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is an application environment diagram of an image processing-based conductor icing analysis method in one embodiment.

[0036] Figure 2This is a flowchart illustrating an image processing-based method for analyzing conductor icing in one embodiment.

[0037] Figure 3 This is a schematic diagram illustrating the training process of a convolutional neural network and a decision tree model in one embodiment.

[0038] Figure 4 This is a schematic diagram of the structure of an image processing-based wire icing analysis method apparatus in one embodiment.

[0039] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0040] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0041] In one embodiment, the image processing-based conductor icing analysis method provided in this application can be applied to... Figure 1 In the application environment shown, the terminal device 102 can be a user-held device with image capture and network connectivity, and can be, but is not limited to, a personal computer, laptop, smartphone, tablet, IoT device, and portable wearable device. The computer device 104 can be any device with data processing capabilities, such as a single server, a server cluster consisting of multiple servers, a personal computer, laptop, smartphone, or tablet.

[0042] Users can take pictures of the icing on the conductor to be analyzed using terminal device 102 to obtain images of the icing. Terminal device 102 can send the images of the icing on the conductor to be analyzed to computer device 104 via a wireless network, so that computer device 104 can perform icing category analysis on the conductor based on the images and output icing category labels to terminal device 102 to reflect the icing on the conductor.

[0043] In one embodiment, this application provides an image processing-based method for analyzing conductor icing. The following embodiments apply this method to... Figure 1 Let's take computer equipment as an example for explanation. Figure 2 As shown, the image processing-based method for analyzing conductor icing provided in this application may include the following steps:

[0044] S202: Obtain the longitudinal pixel cumulative histogram of the icing image of the conductor to be analyzed.

[0045] The image of the conductor being analyzed refers to an image recording the icing on the conductor to be analyzed, which can be obtained by taking pictures of the icing on the conductor using a drone. The icing on the conductor to be analyzed refers to the icing on the conductor for which icing type and thickness analysis is required. The vertical pixel cumulative histogram is a histogram where the horizontal axis is the pixel coordinate of the image of the conductor being analyzed and the vertical axis is the cumulative number of pixels in the image of the conductor being analyzed; it reflects the pixel distribution of the image of the conductor being analyzed.

[0046] It is understood that in this step, the computer device can acquire the longitudinal pixel cumulative histogram of the icing image of the conductor to be analyzed in any way. For example, the computer device can receive the longitudinal pixel cumulative histogram sent by a terminal device or other computer device, or the computer device can use any brightness histogram acquisition algorithm to process the icing image of the conductor to be analyzed to obtain the longitudinal pixel cumulative histogram.

[0047] S204: Input the vertical pixel cumulative histogram into the trained convolutional neural network to obtain the distance ratio threshold set and the pixel number threshold set output by the trained convolutional neural network.

[0048] The trained convolutional neural network can be pre-trained using a first training dataset. The first training dataset may include multiple sets of first training data, each set of which may include a longitudinal pixel cumulative histogram of the training guideline icing image, a pre-determined set of distance ratio thresholds, and a pre-determined set of pixel number thresholds. It is understood that the specific network structure of the convolutional neural network can be determined based on actual factors such as the computing power of the computer equipment and the accuracy requirements of the model; this paper does not impose specific restrictions on this. In one example, the convolutional neural network used in this application can be a convolutional neural network built based on TensorFlow. Following the order from input to output, the network structure of the convolutional neural network is as follows: input layer, concatenation layer, flattening layer, fully connected layer, activation function, and output layer. The concatenation layer includes two convolutional layers and a pooling layer.

[0049] Specifically, since icing generally cannot achieve full coverage, in the image of icing on the conductor to be analyzed, the number of pixels accumulated vertically on the conductor will be greater than the number of pixels accumulated vertically on the ice. Therefore, based on this, this application can distinguish the positions of the conductor and the ice in the vertical pixel accumulation histogram. If the icing type of the conductor to be analyzed is mixed frost, then when the conductor is photographed by a drone, since the formation of mixed frost is related to wind direction, the icing thickness on both sides of the conductor will be asymmetrical. Therefore, when the icing type of the conductor to be analyzed is mixed frost, in the vertical pixel accumulation histogram of the image of the conductor to be analyzed, the lengths of the horizontal coordinate regions occupied by the ice on both sides of the conductor will be different.

[0050] If the icing type of the conductor to be analyzed is rime or hoarfrost, then when the drone captures images of the icing on the conductor, there will be no asymmetry in the icing thickness on both sides of the conductor. However, due to different formation mechanisms, rime ice is thicker than hoarfrost ice. Therefore, when the icing type of the conductor to be analyzed is rime, in the vertical pixel cumulative histogram of the conductor's icing image, the ratio of the length of the horizontal coordinate region occupied by rime ice to the length of the horizontal coordinate region occupied by the conductor in the histogram will be larger than the ratio of the proportional length of the horizontal coordinate region occupied by hoarfrost to the length of the horizontal coordinate region occupied by the conductor in the histogram.

[0051] Based on this, this application can utilize this feature to analyze the icing of the conductor under analysis from the perspective of the cumulative vertical pixel amount of the image, and then determine the icing category of the conductor. Considering that the height of the conductor under analysis varies when the icing image is taken, which affects the determination of the icing thickness, this application can introduce a trained convolutional neural network to process the vertical pixel cumulative histogram of the conductor under analysis image, so as to accurately determine the length of the region on the horizontal axis of the histogram occupied by the icing, and accurately determine the icing category of the conductor under analysis based on the length of the region on the horizontal axis of the histogram occupied by the icing.

[0052] Specifically, the computer device can input the longitudinal pixel cumulative histogram of the icing image of the conductor to be analyzed into a trained convolutional neural network, and obtain a distance ratio threshold set and a pixel number threshold set output by the trained convolutional neural network. The distance ratio threshold set is used to determine the icing thickness of the conductor to be analyzed, and it may include at least one pixel number threshold. The pixel number threshold set can be used to determine the conductor position and icing position within the icing of the conductor to be analyzed.

[0053] S206: Based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, and the pixel number threshold set, an icing category label is obtained to reflect the icing category of the conductor to be analyzed.

[0054] In this step, the computer equipment can determine the first endpoint, second endpoint, first endpoint of icing, and second endpoint of the conductor based on the vertical pixel cumulative histogram of the image of the conductor icing to be analyzed and the set of pixel number thresholds. For example, the computer equipment can locate each peak with a pixel number greater than a first pixel number threshold in the vertical pixel cumulative histogram, and take the horizontal coordinate of the peak with the smallest horizontal coordinate as the first endpoint of the conductor, and the horizontal coordinate of the peak with the largest horizontal coordinate as the second endpoint of the conductor; in the vertical pixel cumulative histogram, it can locate each peak with a pixel number greater than a second pixel number threshold, and take the horizontal coordinate of the peak with the smallest horizontal coordinate as the first endpoint of icing, and the horizontal coordinate of the peak with the largest horizontal coordinate as the second endpoint of icing.

[0055] Given that the first endpoint, second endpoint, first icing endpoint, and second icing endpoint of the conductor corresponding to the icing image to be analyzed are determined, the computer device can calculate the distance between the first endpoint and the second endpoint as a first distance, the distance between the first icing endpoint and the first endpoint of the conductor as a second distance, the second distance being less than the distance between the first icing endpoint and the second endpoint of the conductor, and the distance between the second icing endpoint and the second endpoint of the conductor as a third distance, the third distance being less than the distance between the second icing endpoint and the first endpoint of the conductor.

[0056] Given a first distance, a second distance, and a third distance corresponding to the icing image of the conductor to be analyzed, the computer device can calculate the ratio of the second distance to the first distance as a first ratio, the ratio of the third distance to the first distance as a second ratio, and the absolute value of the difference between the first ratio and the second ratio as a first absolute value.

[0057] Given a first absolute value and a first ratio corresponding to the icing image of the conductor to be analyzed, the computer device can determine the icing category of the conductor based on the first absolute value, the first ratio, a first distance ratio threshold, and a second distance ratio threshold, and output the corresponding icing category label. For example, the computer device can determine an upper limit and a lower limit of the first distance ratio based on the first distance ratio threshold and a preset error calculation formula, and compare the first absolute value with the upper limit and lower limit of the first distance ratio respectively. Based on the comparison result, it can obtain a label indicating whether the icing category of the conductor to be analyzed is mixed rime. If the obtained icing category is not mixed rime, the computer device can determine an upper limit and a lower limit of the second distance ratio based on the second distance ratio threshold and a preset error calculation formula, and compare the first ratio with the upper limit and lower limit of the second distance ratio respectively. Based on the comparison result, it can obtain a label indicating whether the icing category of the conductor to be analyzed is rime ice or hoarfrost.

[0058] In this embodiment, the computer device can input the vertical pixel cumulative histogram of the icing image of the conductor to be analyzed into a trained convolutional neural network to obtain a set of pixel number thresholds and a set of distance ratio thresholds output by the trained convolutional neural network. The computer device can determine the position of the conductor and the ice in the icing image based on the vertical pixel cumulative histogram and the pixel number thresholds, and determine the icing category of the conductor by combining the position of the conductor and the ice with the distance ratio thresholds. Thus, the computer device can automatically and accurately analyze the icing category of the conductor, providing users with a simple and practical analysis method.

[0059] In one embodiment, the distance ratio threshold set includes a first distance ratio threshold for determining whether the conductor icing is mixed rime and a second distance ratio threshold for determining whether the icing type is rime or hoarfrost. Specifically, the diameter of the conductor and the thickness of the icing in the conductor icing image to be analyzed will vary under different shooting angles. To facilitate accurate determination of the distance ratio and thus improve the accuracy of the conductor icing type analysis results, the trained convolutional neural network can output the first distance ratio threshold for determining whether the conductor icing is mixed rime and the second distance ratio threshold for determining whether the icing type is rime or hoarfrost.

[0060] In one embodiment, the step of obtaining an ice category label reflecting the icing category in the icing image of the conductor to be analyzed based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, and the pixel number threshold set further includes: if the icing first endpoint result is the same as the position of the first endpoint or the second endpoint of the conductor, and the icing second endpoint result is the same as the position of the first endpoint or the second endpoint of the conductor, then a fourth label reflecting the icing category of the icing image of the conductor to be analyzed as "no icing" is used as the icing category label.

[0061] Specifically, if the first endpoint of the icing is at the same position as the first endpoint of the conductor, and the second endpoint of the icing is at the same position as the second endpoint of the conductor, it indicates that neither the first nor the second endpoint of the icing exists in the vertical pixel cumulative histogram. In this case, the icing image of the conductor to be analyzed can be considered to be icy. Therefore, when both the first and second endpoints of the icing are at the same position, the computer device can determine that the conductor in the icing image is not icy and assign the fourth label of "no icing" as the icing category label. This improves the accuracy of the conductor icing category analysis results.

[0062] The step of obtaining an ice category label reflecting the icing image of the conductor to be analyzed based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, and the pixel number threshold set includes: inputting the longitudinal pixel cumulative histogram, the first distance ratio threshold, the second distance ratio threshold, the first pixel number threshold, and the second pixel number threshold into a trained decision tree model to obtain the icing category label output by the trained decision tree model.

[0063] Specifically, the computer device can employ a decision tree model to obtain icing category labels. The decision tree model can be pre-trained using a second training dataset. This second training dataset may include multiple sets of training data, each set comprising a longitudinal pixel cumulative histogram of the training conductor icing image, a pre-determined first distance ratio threshold, a pre-determined second distance ratio threshold, a pre-determined first pixel count threshold, a pre-determined second pixel count threshold, and a pre-determined icing category label. By training the decision tree model using this second training dataset, the model can automatically extract if-then rules for determining the icing type of the conductor being analyzed during training.

[0064] After inputting the vertical pixel cumulative histogram of the icing image of the conductor to be analyzed, the first distance ratio threshold output by the convolutional neural network, the second distance ratio threshold output by the convolutional neural network, the first pixel number threshold output by the convolutional neural network, and the second pixel number threshold output by the convolutional neural network into the decision tree model, the decision tree model can make a decision based on the model input data and output the corresponding icing category label.

[0065] In one example, a decision tree model can be used to locate the x-coordinates of points in the vertical pixel cumulative histogram where the number of pixels is greater than a first pixel number threshold. The point with the smallest x-coordinate and the point with the largest x-coordinate are the first endpoint and the second endpoint of the conductor, respectively. The distance between the first endpoint and the second endpoint of the conductor is a first distance. Similarly, in the vertical pixel cumulative histogram, the x-coordinates of points in the vertical pixel cumulative histogram where the number of pixels is greater than a second pixel number threshold are located. The point with the smallest x-coordinate and the point with the largest x-coordinate are the first endpoint and the second endpoint of the icing system. The distance between the first endpoint of the icing system and the first endpoint of the conductor is a second distance, which is less than the distance between the first endpoint of the icing system and the second endpoint of the conductor. The distance between the second endpoint of the icing system and the second endpoint of the conductor is a third distance, which is less than the distance between the second endpoint of the icing system and the first endpoint of the conductor. The ratio of the second distance to the first distance is a first ratio, and the ratio between the third distance and the first distance is a second ratio. The absolute value of the difference between the first ratio and the second ratio is a first absolute value.

[0066] If there are icing first endpoint results and icing second endpoint results in the vertical pixel cumulative histogram, the decision tree can output the conductor icing category label according to the first absolute value, the first ratio, the first distance ratio threshold and the second distance ratio threshold. For details, please refer to the above embodiments, which will not be repeated here.

[0067] In this embodiment, decision trees are used to perform conductor icing category analysis, thereby improving analysis efficiency.

[0068] In one embodiment, prior to the step of obtaining the longitudinal pixel cumulative histogram of the icing image of the conductor to be analyzed, the following steps are included:

[0069] Obtain the original icing image of the conductor, which is a record of the icing to be analyzed;

[0070] The original conductor icing image is binarized to identify the conductor region in the original conductor icing image, and the original conductor icing image is segmented based on the conductor region to obtain a segmented conductor icing image.

[0071] Based on the angle between the conductor and the vertical direction in the segmented conductor icing image, the segmented conductor icing image is corrected to obtain the conductor icing image to be analyzed.

[0072] In this embodiment, the computer device can acquire the original conductor icing image sent by the terminal device, and segment the conductor icing image to be analyzed from the original conductor icing image to reduce the influence of the shooting background on the subsequent conductor icing category analysis, thereby improving the accuracy of the conductor icing category analysis results.

[0073] Optionally, the computer device can identify the icing region of the conductor in the original image of the conductor icing based on edge detection. The edge features can be used to locate the icing region in the original image of the conductor icing; for example, edge detection can be performed using Roberts edge operators, Prewitt operators, Sobel operators, etc.

[0074] In one embodiment, such as Figure 3 As shown, when using convolutional neural networks and decision tree models to perform conductor icing category analysis, the training process of the convolutional neural network and decision tree models may include the following steps:

[0075] S302: Acquire multiple training conductor icing images and build a conductor icing image database with icing category labels based on each training conductor icing image;

[0076] S304: Select multiple training images of icing conductors from the conductor icing image database and generate multiple sets of first training data; each set of first training data includes the vertical pixel cumulative histogram of the training conductor icing images, a pre-determined set of distance ratio thresholds, and a pre-determined set of pixel number thresholds.

[0077] S306: Train the pre-built convolutional neural network using the first training data of each group, and obtain the trained convolutional neural network;

[0078] S308: Select multiple training images of icing conductors from the conductor icing image database, and input the longitudinal pixel cumulative histogram of the multiple training images of icing conductors into the trained convolutional neural network to obtain the first distance ratio threshold, the second distance ratio threshold, the first pixel number threshold and the second pixel number threshold output by the trained convolutional neural network.

[0079] S310: Generate multiple sets of second training data based on the longitudinal pixel cumulative histogram of the training conductor icing image, the first distance ratio threshold, the second distance ratio threshold, the first pixel number threshold, the second pixel number threshold, and the pre-determined icing category label;

[0080] S312: Use multiple sets of second training data to train the decision tree model to obtain the trained decision tree model.

[0081] In one embodiment, the step of calculating the icing thickness based on the first ratio, the second ratio, and the conductor diameter includes:

[0082] The ice thickness on one side of the conductor is calculated by multiplying the first ratio by the conductor diameter, and the ice thickness on the other side of the conductor is calculated by multiplying the second ratio by the conductor diameter.

[0083] The following describes a conductor icing analysis device based on image processing provided in an embodiment of this application. The conductor icing analysis device described below and the conductor icing analysis method based on image processing described above can be referred to in correspondence.

[0084] In one embodiment, such as Figure 4 As shown, this application provides an image processing-based conductor icing analysis device 400, which specifically includes:

[0085] The histogram acquisition module 410 is used to acquire the longitudinal pixel cumulative histogram of the ice-covered image of the conductor to be analyzed;

[0086] The threshold acquisition module 420 is used to input the vertical pixel cumulative histogram into the trained convolutional neural network to obtain the distance ratio threshold set and the pixel number threshold set output by the trained convolutional neural network.

[0087] The icing category label and icing thickness acquisition module 430 is used to obtain icing category labels and icing thickness in the icing image of the conductor to be analyzed based on the longitudinal pixel cumulative histogram, the distance ratio threshold set, the pixel number threshold set, and the conductor diameter.

[0088] In one embodiment, the set of pixel number thresholds includes a first pixel number threshold and a second pixel number threshold for determining the position of the conductor and the position of the icing in the conductor icing image to be analyzed.

[0089] The icing category label and icing thickness acquisition module 430 of this application includes a first distance unit, a second distance unit, a third distance unit, and a label determination unit. The first distance unit is used to locate the horizontal coordinates of points in the vertical pixel cumulative histogram where the number of pixels is greater than a first pixel number threshold. The point with the smallest and largest horizontal coordinates are the first and second endpoints of the conductor, respectively. The distance between the first and second endpoints of the conductor is calculated to obtain a first distance result. The second distance unit is used to locate the horizontal coordinates of points in the vertical pixel cumulative histogram where the number of pixels is greater than the second pixel number threshold. The point with the smallest and largest horizontal coordinates are the first and second endpoints of the icing layer, respectively. The distance between the first endpoint of the icing layer and the first endpoint of the conductor is calculated as a second distance. The second distance is less than the distance between the first endpoint of the icing layer and the second endpoint of the conductor, thus obtaining a second distance result. The third distance unit is used to calculate the distance between the second endpoint of the icing layer and the second endpoint of the conductor. The third distance is less than the distance between the second endpoint of the icing layer and the first endpoint of the conductor, thus obtaining a third distance result.

[0090] The label determination unit is used to determine the icing category label based on the first distance result, the second distance result, the third distance result, and the distance ratio threshold set; the thickness determination unit is used to determine the icing thickness based on the first distance result, the second distance result, the third distance result, the distance ratio threshold set, and the wire diameter.

[0091] In one embodiment, the label determination unit of this application includes a ratio calculation unit, an absolute value calculation unit, and a comparison unit. The ratio calculation unit calculates the ratio of the second distance to the first distance to obtain a first ratio result, and calculates the ratio of the third distance to the first distance to obtain a second ratio. The absolute value calculation unit calculates the absolute value of the difference between the first ratio and the second ratio to obtain a first absolute value result. The comparison unit is configured to: if the first absolute value is greater than the first distance ratio threshold, use a first label reflecting the icing category of the icing image of the conductor to be analyzed as mixed rime as the icing category label; if the first absolute value is less than the first distance ratio threshold and the first ratio is greater than the second distance ratio threshold, use a second label reflecting the icing category of the icing image of the conductor to be analyzed as rime as the icing category label; if the first absolute value is less than the first distance ratio threshold and the first ratio is less than the second distance ratio threshold, use a third label reflecting the icing category of the icing image of the conductor to be analyzed as rime as the icing category label.

[0092] In one embodiment, the label determination unit of this application further includes an endpoint result analysis unit. This endpoint result analysis unit is used to, when the first endpoint result of icing is empty and the second endpoint result of icing is empty, use a fourth label reflecting the icing category of the conductor to be analyzed as "no icing" as the icing category label.

[0093] In one embodiment, the set of pixel number thresholds includes a first pixel number threshold for determining the location of the conductor and a second pixel number threshold for determining the location of the icing.

[0094] The icing category label and icing thickness acquisition module 430 of this application includes a decision tree processing unit. This decision tree processing unit is used to input the vertical pixel cumulative histogram, the first distance ratio threshold, the second distance ratio threshold, the first pixel count threshold, and the second pixel count threshold into a trained decision tree model to obtain the icing category label output by the trained decision tree model.

[0095] The icing category label and icing thickness acquisition module 430 of this application includes an icing thickness calculation unit. This icing thickness calculation unit is used to calculate the icing thickness based on the first ratio, the second ratio, and the wire diameter.

[0096] In one embodiment, the image processing-based conductor icing analysis device 400 of this application may further include an image acquisition module, an image segmentation module, and an image correction module. The image acquisition module acquires an original conductor icing image, which is an image recording the icing to be analyzed. The image segmentation module binarizes the original conductor icing image, identifies conductor regions within the original conductor icing image, and segments the original conductor icing image based on the conductor regions to obtain a segmented conductor icing image. The image correction module corrects the segmented conductor icing image according to the angle between the conductor and the vertical direction in the segmented conductor icing image to obtain the conductor icing image to be analyzed.

[0097] In one embodiment, this application also provides a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform steps of an image processing-based wire icing analysis method as described in any embodiment.

[0098] In one embodiment, this application also provides a computer device storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform steps of the wire icing analysis method as described in any embodiment.

[0099] Indicatively, Figure 5 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of this application. In one example, the computer device can be a server. (Refer to...) Figure 5 The computer device 900 includes a processing component 902, which further includes one or more processors, and memory resources represented by memory 901 for storing instructions, such as application programs, that can be executed by the processing component 902. The application programs stored in memory 901 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 902 is configured to execute instructions to perform the steps of the image processing-based wire icing analysis method described in any of the above embodiments.

[0100] The computer device 900 may also include a power supply component 903 configured to perform power management of the computer device 900, a wired or wireless network interface 904 configured to connect the computer device 900 to a network, and an input / output (I / O) interface 905. The computer device 900 may operate on an operating system stored in memory 901, such as Windows Server™, Mac OS X™, Unix™, Linux™, Free BSD™, or similar.

[0101] Those skilled in the art will understand that the internal structure of the computer device shown in this application is merely a block diagram of a portion of the structure related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0102] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. In this document, "a," "an," "the," "the," and "its" may also include plural forms unless the context clearly indicates otherwise. "Multiple" refers to at least two, such as 2, 3, 5, or 8, etc. "And / or" includes any and all combinations of the related listed items.

[0103] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0104] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for analyzing conductor icing based on image processing, characterized in that, The method comprises: obtaining a longitudinal pixel cumulative histogram of an ice-coated conductor image to be analyzed; inputting the longitudinal pixel cumulative histogram into a trained convolutional neural network to obtain a distance ratio threshold set and a pixel number threshold set output by the trained convolutional neural network; obtaining an ice class label and an ice thickness in the ice-coated conductor image to be analyzed according to the longitudinal pixel cumulative histogram, the distance ratio threshold set, the pixel number threshold set, and the conductor diameter; the pixel number threshold set comprises a first pixel number threshold and a second pixel number threshold for determining the conductor position and the ice position in the ice-coated conductor image to be analyzed; the step of obtaining the ice class label in the ice-coated conductor image to be analyzed according to the longitudinal pixel cumulative histogram, the distance ratio threshold set, and the pixel number threshold set comprises: locating the abscissa with a pixel number greater than the first pixel number threshold in the longitudinal pixel cumulative histogram, wherein the point with the smallest abscissa and the point with the largest abscissa are the first end point of the conductor and the second end point of the conductor, and the distance between the first end point of the conductor and the second end point of the conductor is a first distance; locating the abscissa with a pixel number greater than the second pixel number threshold in the longitudinal pixel cumulative histogram, wherein the point with the smallest abscissa and the point with the largest abscissa are the first end point of the ice and the second end point of the ice, the distance between the first end point of the ice and the first end point of the conductor is a second distance, the second distance is smaller than the distance between the first end point of the ice and the second end point of the conductor, the distance between the second end point of the ice and the second end point of the conductor is a third distance, and the third distance is smaller than the distance between the second end point of the ice and the first end point of the conductor; determining the ice class label according to the first distance, the second distance, and the third distance; the step of determining the ice class label according to the first distance, the second distance, and the third distance comprises: the distance ratio threshold set comprises a first distance ratio threshold and a second distance ratio threshold for determining the conductor position and the ice position in the ice-coated conductor image to be analyzed; calculating the ratio of the second distance to the first distance as a first ratio, the distance ratio between the ratio of the third distance to the first distance as a second ratio, the absolute value of the difference between the first ratio and the second ratio as a first absolute value, and if the first absolute value is greater than the first distance ratio threshold, a first label reflecting the ice class of the ice-coated conductor image to be analyzed as mixed rime is taken as the ice class label; if the first absolute value is less than the first distance ratio threshold and the first ratio is greater than the second distance ratio threshold, a second label reflecting the ice class of the ice-coated conductor image to be analyzed as glaze is taken as the ice class label; if the first absolute value is less than the first distance ratio threshold and the first ratio is less than the second distance ratio threshold, a third label reflecting the ice class of the ice-coated conductor image to be analyzed as hoar is taken as the ice class label; The pixel quantity threshold set comprises a first pixel quantity threshold for determining the conductor position and a second pixel quantity threshold for determining the ice position; The step of obtaining the ice category label in the to-be-analyzed conductor icing image according to the longitudinal pixel cumulative histogram, the distance ratio threshold set and the pixel quantity threshold set comprises: The longitudinal pixel cumulative histogram, the first distance ratio threshold, the second distance ratio threshold, the first pixel quantity threshold and the second pixel quantity threshold are input into the trained decision tree model to obtain the ice category label output by the trained decision tree model.

2. The method of claim 1, wherein, The step of obtaining the ice category label in the to-be-analyzed conductor icing image according to the longitudinal pixel cumulative histogram, the distance ratio threshold set and the pixel quantity threshold set further comprises: If the ice first endpoint result and the conductor first endpoint or the conductor second endpoint position are the same and the ice second endpoint result and the conductor first endpoint or the conductor second endpoint position are the same, a fourth label reflecting that the ice category of the to-be-analyzed conductor icing image is no ice is taken as the ice category label.

3. The method of claim 1, wherein, According to the first ratio, the second ratio and the conductor diameter, the ice thickness is calculated; The value of the first ratio multiplied by the conductor diameter is the ice thickness on one side of the conductor, and the value of the second ratio multiplied by the conductor diameter is the ice thickness on the other side of the conductor.

4. The method according to any one of claims 1 to 3, characterized in that, Before the step of obtaining the longitudinal pixel cumulative histogram of the to-be-analyzed conductor icing image, the method comprises: The original conductor icing image is obtained, and the original conductor icing image is an image recording the to-be-analyzed ice; The original conductor icing image is binarized to identify a conductor region in the original conductor icing image, and the original conductor icing image is segmented based on the conductor region to obtain a segmented conductor icing image; 5. An image processing-based conductor icing analysis apparatus for performing the method of claim 1, characterized by The segmented conductor icing image is corrected according to an included angle between the conductor and the vertical direction to obtain the to-be-analyzed conductor icing image. The method comprises: The histogram obtaining module is configured to obtain a longitudinal pixel cumulative histogram of a to-be-analyzed conductor icing image; The threshold obtaining module is configured to input the longitudinal pixel cumulative histogram into a trained convolutional neural network to obtain a distance ratio threshold set and a pixel quantity threshold set output by the trained convolutional neural network; 6. A storage medium, characterized by The ice category label and ice thickness obtaining module is configured to obtain an ice category label and ice thickness in the to-be-analyzed conductor icing image according to the longitudinal pixel cumulative histogram, the distance ratio threshold set, the pixel quantity threshold set and a conductor diameter.

7. A computer device, comprising: The storage medium stores computer readable instructions, and the computer readable instructions are executed by one or more processors to cause the one or more processors to perform the steps of the image processing-based conductor icing analysis method in any one of claims 1 to 4. The method comprises: One or more processors and a memory; The memory stores computer readable instructions which, when executed by the one or more processors, perform the steps of the image processing-based conductor icing analysis method of any one of claims 1 to 4.

Images