External damage prevention detection method, device and equipment for power transmission line and storage medium
By comparing the prediction results of the anti-out-break detection model with the actual results, the problem of difficult to determine the excellent performance model in the prior art is solved, and efficient anti-out-break detection of transmission lines is achieved, which improves the detection effect.
Patent Information
- Application Number
- CN202411706329.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-26
- Publication Date
- 2025-04-18
AI Technical Summary
In the prior art, it is difficult to determine an anti-outbreak detection model with excellent performance, resulting in poor anti-outbreak detection effect of the transmission line.
By obtaining the test data set, input the target anti-outbreak detection model, obtain the detection results, and perform performance evaluation of the detection results based on the test data set. If the performance evaluation results are excellent, this model is used to perform external breach detection on the transmission line.
The performance of the anti-outbreak detection model is quantified, so that users can compare the advantages and disadvantages of different models, and when the model performance evaluation results are excellent, efficient anti-outbreak detection of transmission lines is achieved and the detection effect is improved.
Smart Images
Figure CN119850002A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power transmission line monitoring, and in particular to a method, device, equipment and storage medium for detecting the external damage of a power transmission line. Background Art
[0002] After being laid, transmission lines are vulnerable to external damage, that is, they are vulnerable to external damage. When transmission lines are damaged, they may cause circuit breakage, short circuit and other faults, resulting in power supply interruption, affecting the normal power consumption of users, especially in important industrial, commercial and residential areas, causing serious economic losses. In addition, it may also lead to dangerous situations such as wire breakage, tower collapse and leakage, thus causing fire, explosion and electric shock accidents, posing a threat to personal safety.
[0003] In order to solve this problem, researchers have proposed a variety of anti-external breach monitoring methods, which can detect external breach risks and issue early warnings through anti-external breach monitoring detection models. However, various anti-external breach detection models still have high false alarm and missed alarm rates. For users, it is difficult to judge the performance of various anti-external breach detection models. If a poor-performing anti-external breach detection model is used for anti-external breach detection, it is easy to cause false alarms and missed alarms, affecting the user experience. That is, in related technologies, it is difficult to determine a model with excellent performance for anti-external breach detection, which may lead to poor final detection results. Summary of the invention
[0004] Based on this, it is necessary to address the problem in related technologies that it is difficult to determine a model with excellent performance for anti-external damage detection, which may lead to poor final detection results. A method, device, equipment and storage medium for anti-external damage detection of transmission lines are proposed.
[0005] In a first aspect, an embodiment of the present application provides a method for detecting external damage to a power transmission line, the method comprising:
[0006] Get the test dataset;
[0007] Input the test data set into the target anti-external damage detection model to obtain the detection result;
[0008] According to the test data set, the performance of the detection results is evaluated to obtain the performance evaluation results of the target anti-external damage detection model;
[0009] If the performance evaluation result corresponds to excellent, the target external damage prevention detection model is used to perform external damage prevention detection on the transmission line.
[0010] In a second aspect, an embodiment of the present application provides a transmission line anti-external damage detection device, the device comprising:
[0011] Acquisition module, used to obtain test data set;
[0012] An input module is used to input the test data set into the target anti-external damage detection model to obtain the detection result;
[0013] An evaluation module is used to perform performance evaluation on the detection results according to the test data set to obtain performance evaluation results of the target anti-external damage detection model;
[0014] The detection module is used to perform anti-external damage detection on the transmission line using a target anti-external damage detection model if the performance evaluation result corresponds to excellent.
[0015] In a third aspect, an embodiment of the present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the following steps:
[0016] Get the test dataset;
[0017] Input the test data set into the target anti-external damage detection model to obtain the detection result;
[0018] According to the test data set, the performance of the detection results is evaluated to obtain the performance evaluation results of the target anti-external damage detection model;
[0019] If the performance evaluation result corresponds to excellent, the target external damage prevention detection model is used to perform external damage prevention detection on the transmission line.
[0020] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor performs the following steps:
[0021] Get the test dataset;
[0022] Input the test data set into the target anti-external damage detection model to obtain the detection result;
[0023] According to the test data set, the performance of the detection results is evaluated to obtain the performance evaluation results of the target anti-external damage detection model;
[0024] If the performance evaluation result corresponds to excellent, the target external damage prevention detection model is used to perform external damage prevention detection on the transmission line.
[0025] The beneficial effect of the embodiment of the present application compared with the prior art is: the embodiment of the present application obtains a test data set, and inputs the test data set into the target anti-external damage detection model to obtain a detection result, and then performs a performance evaluation on the detection result according to the test data set to obtain a performance evaluation result for the target anti-external damage detection model. If the performance evaluation result corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line. The embodiment of the present application performs a performance evaluation on the target anti-external damage detection model by comparing the predicted result of the target anti-external damage detection model for the test data set with the actual result of the test data set, and can quantify the performance of the anti-external damage detection model, so that the user can more easily compare the advantages and disadvantages of different anti-external damage detection models, so as to make a reasonable choice, and when the performance evaluation result of the model corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line, so as to achieve anti-external damage detection on the transmission line and improve the detection effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0027] in:
[0028] Figure 1 It is a flow chart of a method for detecting external damage of a power transmission line in one embodiment;
[0029] Figure 2 It is a structural block diagram of a device for detecting external damage of a power transmission line in one embodiment;
[0030] Figure 3 FIG. 4 is a structural block diagram of a computer device in one embodiment. DETAILED DESCRIPTION
[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0032] After being laid, transmission lines are vulnerable to external damage, that is, they are vulnerable to external damage. When transmission lines are damaged, they may cause circuit breakage, short circuit and other faults, resulting in power supply interruption, affecting the normal power consumption of users, especially in important industrial, commercial and residential areas, causing serious economic losses. In addition, it may also lead to dangerous situations such as wire breakage, tower collapse and leakage, thus causing fire, explosion and electric shock accidents, posing a threat to personal safety.
[0033] In order to solve this problem, researchers have proposed a variety of anti-external breach monitoring methods, which can detect external breach risks and issue early warnings through anti-external breach monitoring detection models. However, various anti-external breach detection models still have high false alarm and missed alarm rates. For users, it is difficult to judge the performance of various anti-external breach detection models. If a poor-performing anti-external breach detection model is used for anti-external breach detection, it is easy to cause false alarms and missed alarms, affecting the user experience. That is, in related technologies, it is difficult to determine a model with excellent performance for anti-external breach detection, which may lead to poor final detection results.
[0034] In view of this, the embodiment of the present application evaluates the performance of the target anti-external damage detection model by comparing the predicted results of the target anti-external damage detection model for the test data set with the actual results of the test data set, so as to quantify the performance of the anti-external damage detection model, so that users can more easily compare the advantages and disadvantages of different anti-external damage detection models, and thus make reasonable choices. When the performance evaluation result of the model corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line, thereby realizing anti-external damage detection of the transmission line and improving the detection effect.
[0035] In order to illustrate the technical solution of the present application, a specific embodiment is provided below for illustration.
[0036] Figure 1 The present invention provides a schematic diagram of the implementation process of a method for detecting external damage to a power transmission line provided in an embodiment of the present application. The method can be applied to computer devices, specifically smart phones, tablet computers, laptop computers, ultra-mobile personal computers (UMPC), netbooks, etc.
[0037] Specifically, the above-mentioned transmission line anti-external damage detection method may include the following steps S101 to S103.
[0038] Step S101, obtaining a test data set.
[0039] Among them, the test data set is a data set with known results, which can be used to evaluate the performance of the model on unknown data.
[0040] In the implementation mode of the present application, the computer device can obtain the transmission line scene image collected by the monitoring camera installed on the transmission line tower, and use the annotation software to annotate the transmission line scene image. Specifically, the position of the external damage target in the transmission line scene image can be enclosed by a rectangular frame and the category label can be marked. The external damage target can specifically include construction machinery, tower cranes, cranes, etc. Then the transmission line scene image data set after the above annotation is randomly divided into a training set, a validation set and a test set according to the proportion. The training set and the validation set can be used to train the anti-external damage detection model.
[0041] In addition, data enhancement preprocessing methods can be used to generate new image data by rotating, flipping, scaling, cropping, adding noise, etc. to expand the test data set.
[0042] Step S102: input the test data set into the target anti-external damage detection model to obtain the detection result.
[0043] Among them, the target anti-external damage detection model is a model that needs to be evaluated, which is used to detect whether the transmission line in the transmission line image is at risk of being damaged by external forces.
[0044] In the implementation of the present application, the target anti-external damage detection model can be deployed in a computer device or in a server, and the implementation of the present application does not limit this. The computer device can input the test data set into the target anti-external damage detection model, and the target anti-external damage detection model will process each input sample and output a prediction result (i.e., whether the external damage risk is detected).
[0045] Step S103: performing performance evaluation on the detection results according to the test data set to obtain a performance evaluation result of the target anti-external damage detection model.
[0046] Among them, the performance evaluation result is the result of quantitative evaluation of the performance of the target anti-external damage detection model, which can be a score or a degree of goodness, etc.
[0047] In an embodiment of the present application, a computer device can compare the prediction results of the model with the actual results of the test set, calculate the performance indicators of the model such as accuracy, recall rate, F1 score, etc., and obtain the performance evaluation results of the target anti-external damage detection model based on these performance indicators.
[0048] Step S104: If the performance evaluation result corresponds to excellent, the target external damage prevention detection model is used to perform external damage prevention detection on the transmission line.
[0049] It is understandable that in some embodiments, the performance evaluation results may be divided into grades, such as poor, average, excellent, etc. The excellent grade may mean that the performance of the target anti-external damage detection model is good, and its performance evaluation score may be above a preset value, such as 0.8 (the range of the evaluation score is 0-1).
[0050] In the implementation of the present application, if the performance evaluation result corresponds to excellent, it can be considered that the performance of the target external damage detection model is good enough and can be directly used for external damage prevention detection. Therefore, the computer device can use the target external damage prevention detection model to perform external damage prevention detection on the transmission line when the performance evaluation result corresponds to excellent.
[0051] The beneficial effect of the embodiment of the present application compared with the prior art is: the embodiment of the present application obtains a test data set, and inputs the test data set into the target anti-external damage detection model to obtain a detection result, and then performs a performance evaluation on the detection result according to the test data set to obtain a performance evaluation result for the target anti-external damage detection model. If the performance evaluation result corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line. The embodiment of the present application performs a performance evaluation on the target anti-external damage detection model by comparing the predicted result of the target anti-external damage detection model for the test data set with the actual result of the test data set, and can quantify the performance of the anti-external damage detection model, so that the user can more easily compare the advantages and disadvantages of different anti-external damage detection models, so as to make a reasonable choice, and when the performance evaluation result of the model corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line, so as to achieve anti-external damage detection on the transmission line and improve the detection effect.
[0052] In some embodiments of the present application, the above-mentioned performance evaluation of the detection results is performed based on the test data set to obtain the performance evaluation results of the target anti-external damage detection model, which may specifically include steps S401 to S407.
[0053] Step S401, calculating the recall rate of the detection result based on the test data set.
[0054] The recall rate is a measure of the proportion of positive examples that the model can correctly identify (correctly identifying transmission lines with external damage risks) to all actual positive examples. A high recall rate means that the model can identify more actual external damage events, but it may also lead to more false positives.
[0055] In the implementation of the present application, the computer device can traverse all samples in the test data set. For each sample that is actually a positive example (i.e., a sample that actually has an external failure), check whether the target anti-external failure detection model also predicts it as a positive example, and calculate the ratio of the number of samples correctly predicted as positive examples by the model to the number of actual positive example samples, i.e., the recall rate.
[0056] Step S402, calculating the accuracy of the detection result based on the test data set.
[0057] Among them, precision is a measure of the proportion of samples predicted by the model to be positive examples that are actually positive examples.
[0058] In an embodiment of the present application, the computer device may traverse all samples in the test data set. For samples predicted by the model as positive examples, check whether they are actually positive examples, and calculate the ratio of the number of samples predicted by the model as positive examples and actually positive examples to the number of samples predicted by the model as positive examples, that is, the precision rate.
[0059] Step S403: Calculate the average precision of the detection results based on the test data set.
[0060] Among them, the average precision can measure the detection performance of the model on all categories.
[0061] In an embodiment of the present application, the computer device may first calculate the detection accuracy of the detection result, and then calculate the average accuracy based on the detection accuracy.
[0062] Step S404, calculating the calculation speed of the detection result based on the test data set.
[0063] Among them, the calculation speed is a measure of the speed at which the model processes the input data and gives the prediction results. It is understandable that for applications with high real-time requirements (such as online anti-external damage monitoring), the calculation speed is an important performance indicator.
[0064] In an embodiment of the present application, the computer device may use a timer to record the total time required for the model to process the test data set, and calculate the total time divided by the number of samples in the test data set to obtain the average processing time for each sample.
[0065] Step S405: Calculate the balance score of the detection result according to the recall rate and the precision rate.
[0066] Among them, the balance score is the harmonic mean of precision and recall, which can be used to strike a balance between the two. A high F1 score means that the model performs well in both precision and recall.
[0067] In an embodiment of the present application, the computer device may calculate the balance score using the harmonic mean of the recall rate and the precision rate.
[0068] Step S406, respectively normalize the recall rate, precision rate, average precision, calculation speed and balance score to obtain normalized recall rate, normalized precision rate, normalized average precision rate, normalized calculation speed and normalized balance score.
[0069] In an embodiment of the present application, the computer device can convert the value of each indicator into a value between 0 and 1 (or other specified range) according to a normalization method (such as linear normalization, minimum-maximum normalization, etc.), thereby eliminating the dimensional differences between different indicators so that they can be compared and weighted on the same scale.
[0070] Step S407, calculating a performance evaluation result according to the normalized recall rate, the normalized precision rate, the normalized average precision, the normalized calculation speed, and the normalized balance score.
[0071] In the implementation of the present application, the computer device can select a comprehensive evaluation method (such as weighted average method) according to the normalized index value, and assign a weight to each index according to the actual application requirements. Finally, the weighted average score is calculated as the final performance evaluation result of the model.
[0072] In some specific implementations of the present application, the above-mentioned performance evaluation of the detection results is performed based on the test data set to obtain the performance evaluation results of the target anti-external damage detection model, which may specifically include steps S501 to S504.
[0073] Step S501, obtaining the actual category label of each image in the test data set.
[0074] Among them, the actual category label is in the test data set, each image is assigned one or more labels representing its true content. It can be understood that the actual category label is accurate and is used to compare with the model's prediction results to evaluate the performance of the model.
[0075] In the implementation of the present application, the actual category labels are usually stored in the form of files, database records or arrays / lists in the code. The computer device can read the actual category labels of each image from the metadata of the test data set to compare with the prediction results of the model.
[0076] Step S502, obtaining the predicted category label of each image in the detection result.
[0077] The predicted category label is the prediction result of the target anti-foreign attack detection model for each image in the test data set. It can be understood that the target anti-foreign attack detection model can output one or more predicted labels based on the content of the input image, indicating what it believes the image contains.
[0078] In an embodiment of the present application, the computer device can run the target anti-external damage detection model, make predictions for each image in the test data set, and collect prediction results. The prediction results usually include a predicted category label for each image.
[0079] Step S503, determining the prediction accuracy type of each image in the detection result according to the actual category label and the corresponding predicted category label, where the prediction result types include true positive examples, false positive examples, and false negative examples.
[0080] Among them, a true positive example means that the model correctly classifies an image as a positive example (that is, the actual category label of the image matches the predicted category label, and the label represents the exoconstruction event we are interested in). A false positive example means that the model incorrectly classifies an image as a positive example (that is, the actual category label of the image does not match the predicted category label, but the predicted label represents the exoconstruction event we are interested in). A false negative example means that the model incorrectly classifies an image as a negative example (that is, the actual category label of the image represents the exoconstruction event we are interested in, but the predicted label does not represent this event).
[0081] In an embodiment of the present application, the computer device may compare the actual category label and the predicted category label of each image. If the two match and represent an external damage event, it is recorded as a true positive example; if the predicted label represents an external damage event but the actual label does not match, it is recorded as a false positive example; if the actual label represents an external damage event but the predicted label does not match, it is recorded as a false negative example.
[0082] Step S504, calculating the number of true positive examples, false positive examples, and false negative examples in the prediction results.
[0083] In an implementation of the present application, the computer device may traverse all prediction results, and count the number of each type according to the prediction accuracy type determined in the previous step, so as to subsequently calculate performance indicators such as recall rate and precision rate.
[0084] The implementation method of the present application can further calculate the recall rate, precision rate and other key performance indicators by calculating the number of true positive examples, false positive examples and false negative examples, so as to comprehensively evaluate the performance of the model in the anti-external damage detection task. In addition, understanding the performance of the model on different types of prediction results helps to identify the weaknesses and improvement directions of the model. For example, if the number of false positive examples is high, it may be necessary to adjust the threshold or training strategy of the model to reduce false positives. In addition, based on accurate performance indicators, users can more scientifically select the anti-external damage detection model that best suits their needs, thereby improving the safety and reliability of transmission lines. At the same time, these indicators can also serve as an important reference in the process of model deployment and operation and maintenance.
[0085] In some specific embodiments of the present application, the above-mentioned calculation of the recall rate of the detection result based on the test data set includes:
[0086] Calculate recall based on the number of true positives and false negatives;
[0087] The recall rate is calculated as:
[0088]
[0089] Among them, RP is the recall rate, TP is the number of true positive examples, and FN is the number of false negative examples.
[0090] The embodiment of the present application calculates the recall rate by the above-mentioned calculation method, which can objectively understand the ability of the target anti-external damage detection model in identifying actual external damage events, and provide strong support for the optimization and improvement of the model. In addition, the result of the recall rate can also guide the optimization direction of the target anti-external damage detection model. If the recall rate is low, it means that the model is insufficient in identifying actual external damage events, and it may be necessary to adjust the model parameters, improve the feature extraction method, or increase the training data to improve the recall rate.
[0091] In some specific embodiments of the present application, the accuracy of the detection result is calculated based on the test data set, including:
[0092] Calculate precision based on the number of true positives and the number of false positives;
[0093] The precision is calculated as:
[0094]
[0095] Among them, RR is the precision rate, TP is the number of true positive examples, and FP is the number of false positive examples.
[0096] The implementation method of this application calculates the accuracy rate through the above calculation method, which can more accurately evaluate the performance of the model on a specific task, especially in scenarios where the false alarm rate needs to be strictly controlled. It can also help developers identify problems with the model and then optimize the model to improve its performance on specific tasks.
[0097] In some specific embodiments of the present application, the average accuracy of the detection results is calculated based on the test data set, including:
[0098] Calculate the detection accuracy based on the number of true positives, false positives, and false negatives;
[0099] Calculate the average of the detection accuracies to get the average precision.
[0100] The detection accuracy is calculated as follows:
[0101]
[0102] Among them, AP is the detection accuracy, TP is the number of true positive examples, FP is the number of false positive examples, and FN is the number of false negative examples;
[0103] The average precision is calculated as:
[0104]
[0105] Among them, mAP is the average precision, N is the total number of samples in the test data set, and i represents the i-th sample.
[0106] The implementation method of this application can comprehensively consider the performance of the model under different conditions by calculating the average precision, so as to more comprehensively evaluate the performance of the model. It can also help developers identify the performance differences of the model under different conditions, and then optimize the model to improve its performance in various scenarios. In addition, the average precision provides a standardized evaluation indicator, making the performance comparison between different models more fair and accurate.
[0107] In some specific embodiments of the present application, the calculation speed of calculating the detection result based on the test data set includes:
[0108] The calculation speed is calculated based on the total number of image frames in the test data set and the total computing time.
[0109] The calculation speed is calculated as:
[0110] p=s / t.
[0111] Where s is the total number of image frames, and t is the operation time of the target anti-external damage detection model on the test dataset.
[0112] In some specific implementations of the present application, the balance score is calculated as follows:
[0113]
[0114] Among them, F1 is the balance score, RP is the recall rate, and RR is the precision rate.
[0115] In some specific implementations of the present application, the recall rate, precision rate, average precision, calculation speed and balance score are normalized respectively, and the normalized recall rate, normalized precision rate, normalized average precision rate, normalized calculation speed and normalized balance score are obtained accordingly, including:
[0116] Normalize the recall rate to obtain the normalized recall rate;
[0117] The accuracy is normalized to obtain a normalized accuracy;
[0118] The average precision is normalized to obtain the normalized average precision;
[0119] The calculation speed is normalized to obtain a normalized calculation speed;
[0120] The balance score is normalized to obtain a normalized balance score;
[0121] The normalized recall rate is calculated as:
[0122]
[0123] Among them, RP * is the normalized recall rate, RP is the recall rate, RP(min) is the minimum recall rate, and RP(max) is the maximum recall rate;
[0124] The normalized precision is calculated as:
[0125]
[0126] Among them, RR * is the normalized precision, RR is the precision, RR(min) is the minimum precision, and RR(max) is the maximum precision;
[0127] The normalized average precision is calculated as:
[0128]
[0129] Among them, e AP* is the normalized average precision, e AP is the average precision, e AP (min) is the minimum average precision, e AP (max) is the maximum average accuracy;
[0130] The normalized calculation speed is calculated as follows:
[0131]
[0132] Among them, lgp * is the normalized calculation speed, lgp is the logarithm of the calculation speed, lgp(min) is the minimum value of the logarithm of the calculation speed, and lgp(max) is the maximum value of the logarithm of the calculation speed;
[0133] The normalized balance score is calculated as follows:
[0134]
[0135] in, is the normalized balance score, F1 is the balance score, F1(min) is the minimum balance score, and F1(max) is the maximum balance score.
[0136] The implementation method of the present application uses normalization processing, and the performance evaluation indicators (recall rate, precision rate, average precision, calculation speed, balance score) under different models or different conditions are scaled to the same range, which is convenient for direct comparison and analysis. The normalized indicators can be directly used in the performance evaluation report without additional data conversion or processing.
[0137] In some embodiments of the present application, the performance evaluation results include good performance, average performance, and poor performance. After the performance evaluation of the detection results is performed according to the test data set to obtain the performance evaluation results of the target anti-external damage detection model, the above method further includes the steps of:
[0138] If the performance evaluation result corresponds to poor performance, then determining the parameters to be optimized of the target anti-external damage detection model according to the performance evaluation result;
[0139] Generate corresponding optimization strategies according to the parameters to be optimized;
[0140] The target anti-external damage detection model is optimized according to the optimization strategy to obtain the optimized anti-external damage detection model.
[0141] In the implementation of the present application, the computer device can use the test data set to test the target anti-external damage detection model, evaluate the model according to indicators such as recall rate, precision rate, average precision, etc., and analyze which parameters (such as learning rate, model complexity, feature selection, etc.) have the greatest impact on model performance based on the performance evaluation results, and determine these parameters as parameters to be optimized. Then, based on the parameters to be optimized, generate the corresponding optimization strategy.
[0142] The specific optimization strategies are as follows:
[0143] When the parameter to be optimized is the recall rate, the computer equipment can increase the number of positive examples and collect more samples of external breach events, especially those that are easily missed by the model, to enhance the recognition ability of the model. Adjust the classification threshold and lower the classification threshold of the model to increase the number of positive examples recognized.
[0144] When the parameter to be optimized is accuracy, the computer equipment can perform data cleaning to remove noise and outliers in the training data to reduce the erroneous information learned by the model. More discriminative features can be extracted to help the model better distinguish between positive and negative examples. Regularization techniques, such as L1 and L2 regularization, can be used to prevent overfitting of the model and improve the generalization ability of the model.
[0145] When the parameter to be optimized is the average precision, the computer equipment can optimize the model structure and adjust the number of network layers, activation functions, etc. of the model to improve the model's expressiveness and generalization capabilities. Ensemble learning methods, such as bagging and boosting, combine the prediction results of multiple models to improve overall performance.
[0146] When the parameter to be optimized is the calculation speed, the computer equipment can perform model compression, using pruning, quantization and other technologies to reduce the number of model parameters and the amount of calculation to increase the calculation speed. Hardware acceleration can also be performed, using hardware acceleration technologies such as GPU to improve the calculation efficiency of the model. Algorithm optimization can also be performed to optimize the algorithm implementation of the model and reduce unnecessary calculation overhead.
[0147] When the parameter to be optimized is a balance score, the computer equipment can make comprehensive adjustments, and according to the specific performance of recall rate and precision rate, comprehensively adjust the parameters and structure of the model to find the best balance between the two. You can also use weighted F-score, assign different weights to recall rate and precision rate according to the needs of the actual application scenario, and calculate the weighted F-score as the optimization target.
[0148] Figure 2 The schematic diagram of the structure of a transmission line anti-breakage detection device provided in an embodiment of the present application is shown. The transmission line anti-breakage detection device 2 can be configured on a computer device. Specifically, the transmission line anti-breakage detection device 2 can include:
[0149] An acquisition module 201 is used to acquire a test data set;
[0150] An input module 202 is used to input a test data set into a target anti-external damage detection model to obtain a detection result;
[0151] An evaluation module 203 is used to perform a performance evaluation on the detection result according to the test data set to obtain a performance evaluation result of the target anti-external damage detection model;
[0152] The detection module is used to perform anti-external damage detection on the transmission line using a target anti-external damage detection model if the performance evaluation result corresponds to excellent.
[0153] The beneficial effect of the embodiment of the present application compared with the prior art is: the embodiment of the present application obtains a test data set, and inputs the test data set into the target anti-external damage detection model to obtain a detection result, and then performs a performance evaluation on the detection result according to the test data set to obtain a performance evaluation result for the target anti-external damage detection model. If the performance evaluation result corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line. The embodiment of the present application performs a performance evaluation on the target anti-external damage detection model by comparing the predicted result of the target anti-external damage detection model for the test data set with the actual result of the test data set, and can quantify the performance of the anti-external damage detection model, so that the user can more easily compare the advantages and disadvantages of different anti-external damage detection models, so as to make a reasonable choice, and when the performance evaluation result of the model corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line, so as to achieve anti-external damage detection on the transmission line and improve the detection effect.
[0154] In some embodiments of the present application, the above-mentioned evaluation module 203 may also be used to: calculate the recall rate of the detection result based on the test data set;
[0155] Calculate the accuracy of the detection results based on the test data set;
[0156] Calculate the average precision of the detection results based on the test data set;
[0157] The computational speed of computing the detection results based on the test dataset;
[0158] Calculate the balance score of the test results based on the recall rate and precision rate;
[0159] Normalize the recall rate, precision rate, average precision, calculation speed and balance score respectively, and obtain the normalized recall rate, normalized precision rate, normalized average precision rate, normalized calculation speed and normalized balance score;
[0160] The performance evaluation results are calculated based on the normalized recall rate, normalized precision, normalized average precision, normalized calculation speed and normalized balance score.
[0161] In some embodiments of the present application, the above-mentioned power transmission line anti-external damage detection device 2 further includes a calculation module for: obtaining an actual category label of each image in a test data set;
[0162] Get the predicted category label for each image in the detection results;
[0163] According to the actual category label and the corresponding predicted category label, the prediction accuracy type of each image in the detection result is determined. The prediction result types include true positive examples, false positive examples, and false negative examples.
[0164] Calculate the number of true positives, false positives, and false negatives in the prediction results.
[0165] In some embodiments of the present application, the evaluation module 203 is further used to: calculate the recall rate according to the number of true positive examples and the number of false negative examples;
[0166] The recall rate is calculated as:
[0167]
[0168] Among them, RP is the recall rate, TP is the number of true positive examples, and FN is the number of false negative examples.
[0169] In some embodiments of the present application, the evaluation module 203 is further used to: calculate the precision rate according to the number of true positive examples and the number of false positive examples;
[0170] The precision is calculated as:
[0171]
[0172] Among them, RR is the precision rate, TP is the number of true positive examples, and FP is the number of false positive examples.
[0173] The evaluation module 203 is further used to: calculate the detection accuracy according to the number of true positive examples, the number of false positive examples and the number of false negative examples;
[0174] Calculate the average of the detection accuracies to get the average precision.
[0175] The detection accuracy is calculated as follows:
[0176]
[0177] Among them, AP is the detection accuracy, TP is the number of true positive examples, FP is the number of false positive examples, and FN is the number of false negative examples;
[0178] The average precision is calculated as:
[0179]
[0180] The evaluation module 203 is further used to: normalize the recall rate to obtain a normalized recall rate;
[0181] The accuracy is normalized to obtain a normalized accuracy;
[0182] The average precision is normalized to obtain the normalized average precision;
[0183] The calculation speed is normalized to obtain a normalized calculation speed;
[0184] The balance score is normalized to obtain a normalized balance score;
[0185] The normalized recall rate is calculated as:
[0186]
[0187] Among them, RP * is the normalized recall rate, RP is the recall rate, RP(min) is the minimum recall rate, and RP(max) is the maximum recall rate;
[0188] The normalized precision is calculated as:
[0189]
[0190] Among them, RR *is the normalized precision, RR is the precision, RR(min) is the minimum precision, and RR(max) is the maximum precision;
[0191] The normalized average precision is calculated as:
[0192]
[0193] Among them, e AP* is the normalized average precision, e AP is the average precision, e AP (min) is the minimum average precision, e AP (max) is the maximum average accuracy;
[0194] The normalized calculation speed is calculated as follows:
[0195]
[0196] Among them, lgp * is the normalized calculation speed, lgp is the logarithm of the calculation speed, lgp(min) is the minimum value of the logarithm of the calculation speed, and lgp(max) is the maximum value of the logarithm of the calculation speed;
[0197] The normalized balance score is calculated as follows:
[0198]
[0199] in, is the normalized balance score, F1 is the balance score, F1(min) is the minimum balance score, and F1(max) is the maximum balance score.
[0200] Figure 3 FIG. 1 shows an internal structure diagram of a computer device in an embodiment. The computer device may be a terminal or a server. Figure 3 As shown, the computer device includes a processor, a memory and a network interface connected via a system bus. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program. When the computer program is executed by the processor, the processor may implement the anti-external damage detection method for the power transmission line. The internal memory may also store a computer program. When the computer program is executed by the processor, the processor may implement the anti-external damage detection method for the power transmission line. Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0201] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the following steps:
[0202] Get the test dataset;
[0203] Input the test data set into the target anti-external damage detection model to obtain the detection result;
[0204] According to the test data set, the performance of the detection results is evaluated to obtain the performance evaluation results of the target anti-external damage detection model;
[0205] If the performance evaluation result corresponds to excellent, the target external damage prevention detection model is used to perform external damage prevention detection on the transmission line.
[0206] The beneficial effect of the embodiment of the present application compared with the prior art is: the embodiment of the present application obtains a test data set, and inputs the test data set into the target anti-external damage detection model to obtain a detection result, and then performs a performance evaluation on the detection result according to the test data set to obtain a performance evaluation result for the target anti-external damage detection model. If the performance evaluation result corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line. The embodiment of the present application performs a performance evaluation on the target anti-external damage detection model by comparing the predicted result of the target anti-external damage detection model for the test data set with the actual result of the test data set, and can quantify the performance of the anti-external damage detection model, so that the user can more easily compare the advantages and disadvantages of different anti-external damage detection models, so as to make a reasonable choice, and when the performance evaluation result of the model corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line, so as to achieve anti-external damage detection on the transmission line and improve the detection effect.
[0207] In one embodiment, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the processor performs the following steps:
[0208] Get the test dataset;
[0209] Input the test data set into the target anti-external damage detection model to obtain the detection result;
[0210] According to the test data set, the performance of the detection results is evaluated to obtain the performance evaluation results of the target anti-external damage detection model;
[0211] If the performance evaluation result corresponds to excellent, the target external damage prevention detection model is used to perform external damage prevention detection on the transmission line.
[0212] The beneficial effect of the embodiment of the present application compared with the prior art is: the embodiment of the present application obtains a test data set, and inputs the test data set into the target anti-external damage detection model to obtain a detection result, and then performs a performance evaluation on the detection result according to the test data set to obtain a performance evaluation result for the target anti-external damage detection model. If the performance evaluation result corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line. The embodiment of the present application performs a performance evaluation on the target anti-external damage detection model by comparing the predicted result of the target anti-external damage detection model for the test data set with the actual result of the test data set, and can quantify the performance of the anti-external damage detection model, so that the user can more easily compare the advantages and disadvantages of different anti-external damage detection models, so as to make a reasonable choice, and when the performance evaluation result of the model corresponds to excellent, the target anti-external damage detection model is used to perform anti-external damage detection on the transmission line, so as to achieve anti-external damage detection on the transmission line and improve the detection effect.
[0213] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0214] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0215] In the embodiments provided in the present application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0216] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0217] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0218] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.
[0219] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.
Claims
1. A method for detecting external damage of a power transmission line, characterized in that: The method comprises: Get the test dataset; Inputting the test data set into the target anti-external damage detection model to obtain a detection result; According to the test data set, a performance evaluation is performed on the detection result to obtain a performance evaluation result of the target anti-external damage detection model; If the performance evaluation result corresponds to excellent, the target external damage prevention detection model is used to perform external damage prevention detection on the transmission line.
2. The method for detecting the external damage of a power transmission line according to claim 1, characterized in that: The performing of a performance evaluation on the detection result according to the test data set to obtain a performance evaluation result of the target anti-external damage detection model includes: Calculating the recall rate of the detection result based on the test data set; Calculating the accuracy of the detection result based on the test data set; Calculate the average precision of the detection results based on the test data set; a calculation speed of calculating the detection result based on the test data set; Calculating a balance score of the detection result according to the recall rate and the precision rate; Normalizing the recall rate, the precision rate, the average precision, the calculation speed, and the balance score respectively to obtain a normalized recall rate, a normalized precision rate, a normalized average precision rate, a normalized calculation speed, and a normalized balance score; The performance evaluation result is calculated according to the normalized recall rate, the normalized precision rate, the normalized average precision, the normalized calculation speed and the normalized balance score.
3. The method for detecting the damage of a power transmission line according to claim 2, characterized in that: Before performing performance evaluation on the detection result according to the test data set to obtain a performance evaluation result of the target anti-external damage detection model, the method further includes: Obtaining the actual category label of each image in the test dataset; Obtaining a predicted category label for each image in the detection result; According to the actual category label and the corresponding predicted category label, the predicted accuracy type of each image in the detection result is determined, and the predicted result type includes true positive examples, false positive examples, and false negative examples. Calculate the number of true positive examples, false positive examples, and false negative examples in the prediction result.
4. The method for detecting the power transmission line against external damage according to claim 3, characterized in that: The calculating the recall rate of the detection result based on the test data set includes: Calculating the recall rate based on the number of true positive examples and the number of false negative examples; The recall rate is calculated as: Among them, RP is the recall rate, TP is the number of true positive examples, and FN is the number of false negative examples.
5. The method for detecting the external damage of a power transmission line according to claim 3, characterized in that: The calculating the accuracy of the detection result based on the test data set includes: Calculating the precision rate based on the number of true positive examples and the number of false positive examples; The precision is calculated as: Among them, RR is the precision rate, TP is the number of true positive examples, and FP is the number of false positive examples.
6. The method for detecting the external damage of a power transmission line according to claim 3, characterized in that: The calculating the average precision of the detection result based on the test data set includes: Calculating the detection accuracy based on the number of true positive examples, the number of false positive examples, and the number of false negative examples; An average value of the detection accuracies is calculated to obtain the average accuracy. The detection accuracy is calculated as follows: Among them, AP is the detection accuracy, TP is the number of true positive examples, FP is the number of false positive examples, and FN is the number of false negative examples; The average precision is calculated as:
7. The method for detecting the external damage of a power transmission line according to claim 3, characterized in that: The recall rate, the precision rate, the average precision, the calculation speed and the balance score are respectively normalized to obtain a normalized recall rate, a normalized precision rate, a normalized average precision rate, a normalized calculation speed and a normalized balance score, including: Normalizing the recall rate to obtain a normalized recall rate; Normalizing the accuracy to obtain a normalized accuracy; Normalizing the average precision to obtain a normalized average precision; Normalizing the calculation speed to obtain a normalized calculation speed; Normalizing the balance score to obtain a normalized balance score; The normalized recall rate is calculated as: Among them, RP * is the normalized recall rate, RP is the recall rate, RP(min) is the minimum recall rate, and RP(max) is the maximum recall rate; The normalized precision is calculated as: Among them, RR * is the normalized precision, RR is the precision, RR(min) is the minimum precision, and RR(max) is the maximum precision; The normalized average precision is calculated as: in, is the normalized average precision, e AP is the average precision, e AP (min) is the minimum average precision, e AP (max) is the maximum average accuracy; The normalized calculation speed is calculated as follows: Among them, lgp * is the normalized calculation speed, lgp is the logarithm of the calculation speed, lgp(min) is the minimum value of the logarithm of the calculation speed, and lgp(max) is the maximum value of the logarithm of the calculation speed; The normalized balance score is calculated as follows: in, is the normalized balance score, F1 is the balance score, F1(min) is the minimum balance score, and F1(max) is the maximum balance score.
8. A transmission line anti-external damage detection device, characterized in that: The device comprises: Acquisition module, used to obtain test data set; An input module, used to input the test data set into the target anti-external damage detection model to obtain a detection result; An evaluation module, used to perform a performance evaluation on the detection result according to the test data set to obtain a performance evaluation result for the target anti-external damage detection model; A detection module is used to perform an anti-external damage detection on the power transmission line using the target anti-external damage detection model if the performance evaluation result corresponds to excellent.
9. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the processor executes the steps of the power transmission line anti-external damage detection method according to any one of claims 1 to 7.
10. A computer device, characterized in that: The device comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for detecting the external damage of a power transmission line as claimed in any one of claims 1 to 7.