Tree line fault identification and early warning method based on DCN-YOLO model

Abstract

The invention discloses a DCN-YOLO model-based tree line fault identification and early warning method, which comprises the steps of simulating real tree line discharge through a tree line discharge experiment platform, shooting a video, exporting a picture according to frames, and labeling the picture to obtain an initial data set; acquiring a flame, cable and branch data set, expanding the initial data set to form an enhanced data set, and dividing the enhanced data set into a training set, a verification set and a test set; data enhancement, channel attention mechanism improvement and loss function sample weight balance improvement are carried out based on a DCN-YOLO model; the enhanced data set is input into the improved model for training and verification, the maximum evaluation index mAP50 and mAP50-90 values are obtained, and a tree line fault monitoring and early warning model is obtained; the tree line fault monitoring and early warning model is connected to a power transmission tower monitoring system, and when it is monitored that branches and flames or one of the branches and the flames exist on a power transmission line, an alarm system is triggered; according to the invention, accurate early warning in the early stage of tree line fault fire is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power disaster prevention, and in particular to a tree line fault identification and early warning method based on a DCN-YOLO model. Background Art

[0002] The model detection of tree-line faults is based on image processing and deep learning technology. Transmission lines are critical power transmission facilities in the power system. Tree-line contact or discharge faults in transmission lines passing through forests may cause forest fires or power line tripping accidents. Rapid identification and warning of tree-line faults and taking effective measures to prevent faults can effectively improve the power supply stability of the power system and reduce the probability of forest fires. At present, there is a lack of target detection for tree-line faults, and the target detection algorithm based on deep learning has better robustness and accuracy than the traditional feature-based target detection algorithm. Based on the characteristics of fast computing speed and easy training of the YOLO series, it can better meet the real-time requirements of fault identification and warning.

[0003] Existing technologies for transmission line research are mostly focused on forest fire image recognition under the line, and are unable to accurately warn of tree-line faults on the line in the early stages of fires, thereby reducing the probability of tree-line faults causing forest fires. At present, when the three-phase voltage of the power grid is unbalanced and a tree-line fault occurs, it is difficult to accurately identify the faulty line, resulting in the inability to take timely and effective measures to cut off the faulty line, affecting the safety and stability of the power system. Summary of the invention

[0004] In order to solve the shortcomings of the existing image recognition technology, the present invention provides a tree-line fault recognition and early warning method based on the DCN-YOLO model. Aiming at the problem that the current transmission line image recognition technology is lacking in tree-line fault recognition technology, an improved DCN-YOLO model is proposed in the existing target detection technology. By enhancing the data set and adjusting the improved attention mechanism and loss function parameters, a tree-line fault monitoring and early warning model is obtained and connected to the early warning system, so as to realize real-time online recognition, monitoring and early warning of transmission line tree-line faults.

[0005] To achieve the above object, the technical solution adopted by the present invention is: a tree line fault identification and early warning method based on the DCN-YOLO model, comprising the following steps:

[0006] Step 1: Use the tree line discharge experimental platform to simulate real tree line discharge, shoot videos and export pictures by frame, and label the pictures to obtain the initial data set;

[0007] Step 2: Obtain the flame, cable, and branch data sets to expand the initial data set to form an enhanced data set, and divide it into a training set, a validation set, and a test set;

[0008] Step 3: Based on the DCN-YOLO model, data enhancement, channel attention mechanism improvement, and loss function sample weight balance improvement are performed;

[0009] Step 4: Input the enhanced data set into the improved model for training and verification, obtain the highest evaluation index mAP50 and mAP50-90 values, and obtain the tree line fault monitoring and early warning model;

[0010] Step 5: Connect the tree-line fault monitoring and early warning model to the transmission tower monitoring system. When branches and flames or one of the two are detected on the transmission line, the alarm system will be triggered.

[0011] As a further improvement of the present invention, in step 1, image data when a tree line fault occurs is obtained by using image acquisition equipment installed around the transmission line, and the image acquisition equipment includes a movable camera and a visual monitoring device set at a fixed node of the tower; the initial data set refers to the data set of flame, cable, and branch labels after labeling.

[0012] As a further improvement of the present invention, the step 3 specifically comprises the following steps:

[0013] Step 3.1, build the DCN-YOLO model, including the backbone network, which is used to extract features from the input image, the head network, which is used to make decisions for the target detection model and generate the final detection results, and the neck network, which is used for feature fusion and enhancement. The deformable convolution DCNv3 is introduced to improve the performance of processing complex deformed targets.

[0014] Step 3.2: Randomly adjust the color attributes of the dataset to enhance the DCN-YOLO model's ability to recognize targets under different lighting and color conditions. At the same time, randomly rotate the image by up to 15 degrees to simulate the different orientations of the target in the image. Randomly crop an area from the original image of the dataset and adjust it to a size of 224x224 for the DCN-YOLO model to learn targets of different sizes and proportions. Apply Gaussian blur to the dataset image to simulate images at different focal lengths and enhance the DCN-YOLO model's invariance to details.

[0015] Step 3.3: Integrate a channel attention mechanism into the DCN-YOLO model, obtain the global information of the feature map through global average pooling and global maximum pooling, and then learn the weight relationship between channels through two 1×1 convolutional fully connected layers and ReLU activation function. Finally, use the Sgmoid function to compress the output to the [0,1] interval as the channel attention weight;

[0016] Step 3.4: Improve the DCN-YOLO model through the focal loss function, reduce the weight of easy-to-classify samples and increase the weight of difficult-to-classify samples.

[0017] As a further improvement of the present invention, in step 3.3, the global average pooling is specifically as follows:

[0018]

[0019] Where: F is the input feature map, H and W are the height and width of the feature map respectively, and c is the channel index;

[0020] The global maximum pooling is as follows:

[0021]

[0022] The ReLU activation function is as follows:

[0023] ReLU(x)=max(0,x)

[0024] The Sigmoid function is as follows:

[0025]

[0026] Where: e is the base of the natural logarithm and x is the input value;

[0027] In step 3.4, the focal loss function is as follows:

[0028] FocalLoss(p t )=-α(1-p t ) γ log(p t )

[0029] Where: p t is the predicted probability of the model for each category. For positive samples, p t is the confidence of the model predicting the positive class; for negative samples, p t is the confidence of the model prediction as negative class; α t is the coefficient that balances the weights of positive and negative samples; γ is the focusing parameter that adjusts the weights of easy-to-classify samples and difficult-to-classify samples.

[0030] As a further improvement of the present invention, in step 4, the model is made to reach an ideal state by adjusting the data enhancement, attention mechanism and loss function parameters in the improved algorithm, and the trained and optimized DCN-YOLO model is obtained and deployed in practical applications to perform tree line fault identification tasks.

[0031] As a further improvement of the present invention, in step 5, the alarm system refers to outputting flame and branch respectively when the trained and optimized DCN-YOLO model recognizes flame and branch, and taking flame and branch as input signals for triggering the alarm into the alarm system to complete the early warning of tree line failure.

[0032] As a further improvement of the present invention, the early warning for tree line failure includes a primary early warning and a secondary alarm. When the trained and optimized DCN-YOLO model only recognizes branches and outputs branch, a primary early warning is triggered. When the trained and optimized DCN-YOLO model recognizes flames and outputs flame, a secondary alarm is triggered.

[0033] The beneficial effects of the present invention are:

[0034] 1. The present invention simulates real tree-line discharge through a tree-line discharge experimental platform, shoots videos and exports pictures by frame, uses labelimg to label the pictures to obtain an initial data set; obtains flame, cable, and branch data sets through roboflow to expand the initial data to form an enhanced data set, and divides it into a training set, a validation set, and a test set; performs data enhancement, channel attention mechanism improvement, and loss function Focal Loss sample weight balance improvement based on the DCN-YOLO model; inputs the enhanced data set into the improved DCN-YOLO model for training and verification, obtains the highest mAP50 and mAP50-90 values, and obtains a tree-line fault monitoring and early warning model; connects the tree-line fault monitoring and early warning model to the transmission tower monitoring system, and triggers the alarm system when branches and flames or one of the two are detected on the transmission line. The present invention realizes accurate early warning of tree-line fault fires in the early stages by performing image recognition and early warning on tree-line faults, thereby reducing the probability of fires caused by tree-line faults.

[0035] 2. The present invention uses an enhanced data set containing a large number of real tree-line discharge images for marking and inputs the DCN-YOLO model for training, and uses the improved DCN-YOLO model with the highest mAP50 and mAP50-90 values ​​as a tree-line fault monitoring and early warning model. It can quickly and accurately identify whether there is a tree-line fault in the transmission line, realize real-time monitoring of the transmission line, and promptly discover possible tree-line fault hazards. Through the alarm system, it can more quickly and clearly confirm whether there is a tree-line discharge in the transmission line, improve the integrated automation of fault identification and early warning, assist staff to promptly and effectively avoid the occurrence and expansion of tree-line faults, and improve the power supply reliability and safety of the transmission line. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0038] Example

[0039] like Figure 1 As shown, a tree line fault identification and warning method based on the DCN-YOLO model includes the following steps:

[0040] S1: Use the treeline discharge experimental platform to simulate real treeline discharge, shoot videos and export pictures by frame, and use labelimg to label the pictures to obtain the initial data set;

[0041] The method for obtaining the initial data set is:

[0042] The image data when the tree line fault occurs is obtained by installing image acquisition equipment around the transmission line, and the image acquisition equipment includes a movable camera and a visual monitoring device set up at a fixed node of the tower. The initial data set refers to the data set marked with flame, cable, and branch labels.

[0043] S2: Use roboflow to obtain flame, cable, and branch datasets to expand the initial data to form an enhanced dataset, and divide it into training set, validation set, and test set;

[0044] The training set, validation set and test set refer to dividing the enhanced data set into a train training set, a val validation set and a test test set in a ratio of 8:1:1.

[0045] S3: Data enhancement, channel attention mechanism improvement, and Focal Loss sample weight balance improvement based on the DCN-YOLO model;

[0046] The improvement based on the DCN-YOLO model includes the following steps:

[0047] S31: Build the DCN-YOLO model, including the backbone network backbone responsible for extracting features from the input image, the head network Head responsible for making decisions on the target detection model and generating the final detection results, the neck network Neck responsible for feature fusion and enhancement, and introduce the deformable convolution DCNv3 to improve the performance of processing complex deformed targets.

[0048] S32: Randomly adjust the color attributes (brightness, contrast, saturation, and hue) of the dataset in S2 to enhance the model's ability to recognize targets under different lighting and color conditions. At the same time, randomly rotate the image by up to 15 degrees to simulate the different orientations of the target in the image, and randomly crop an area from the original image of the dataset and resize it to 224x224 to help the model learn different sizes and proportions of the target. Apply Gaussian blur to the dataset images to simulate images at different focal lengths and enhance the model's invariance to details.

[0049] S33: A channel attention mechanism is integrated into the DCN-YOLO model. The global information of the feature map is obtained through global average pooling and global maximum pooling respectively. The weight relationship between channels is learned through two 1×1 convolutional fully connected layers and ReLU activation function. Finally, the Sgmoid function is used to compress the output to the [0,1] interval as the channel attention weight.

[0050] In S33, the attention module performs global average pooling, as shown in the following formula:

[0051]

[0052] Where: F is the input feature map, H and W are the height and width of the feature map respectively, and c is the channel index.

[0053] The global maximum pooling is shown in the following formula:

[0054]

[0055] Where: F is the input feature map, H and W are the height and width of the feature map respectively, and c is the channel index.

[0056] The ReLU activation function is shown in the following formula:

[0057] ReLU(x)=max(0,x)

[0058] The formula means that the ReLU function sets all negative values ​​to 0, while all non-negative values ​​remain unchanged.

[0059] The Sigmoid function is shown in the following formula:

[0060]

[0061] Where: e is the base of the natural logarithm, approximately equal to 2.71828, and x is the input value.

[0062] S34: Improve the model through the focal loss function, reduce the weight of easy-to-classify samples and increase the weight of difficult-to-classify samples.

[0063] In S34, the focal loss function is shown in the following formula:

[0064] Focal Loss(pt)=-α(1-pt)γlog(pt)

[0065] Where: pt is the predicted probability of each category by the model. For positive samples, pt is the confidence of the model predicting the positive class; for negative samples, pt is the confidence of the model predicting the negative class. αt is the coefficient for balancing the weights of positive and negative samples. Different αt values ​​can be set for positive and negative samples respectively. Usually, a smaller value is taken for positive samples to solve the problem of class imbalance. γ is a focusing parameter for adjusting the weights of easy-to-classify samples and difficult-to-classify samples. It is usually a number greater than 0, which is used to reduce the loss contribution of easy-to-classify samples and increase the loss contribution of difficult-to-classify samples.

[0066] S4: Input the enhanced data set into the improved DCN-YOLO model for training and verification, obtain the highest mAP50 and mAP50-90 values, and obtain the tree line fault monitoring and early warning model;

[0067] In S4, the data enhancement, attention mechanism and loss function parameters in the improved algorithm are adjusted to make the model reach the ideal state, and the trained and optimized DCN-YOLO model is obtained, which is then deployed in practical applications to perform tree line fault identification tasks.

[0068] S5: Connect the tree line fault monitoring and early warning model to the transmission tower monitoring system, and trigger the alarm system when branches and flames or one of the two are detected on the transmission line. The present invention realizes accurate early warning of tree line fault fires by image recognition and early warning of tree line faults, thereby reducing the probability of fires caused by tree line faults.

[0069] In S5, when the presence of branches and flames or one of the two is detected on the transmission line, the alarm system will be triggered. The alarm system means that when the model recognizes flames and branches, it will output flame and branch respectively. Both flame and branch will enter the alarm system as input signals for triggering alarms, completing the early warning of tree line faults. The early warning is divided into primary early warning and secondary alarm. When the model only recognizes branches and outputs branch, the primary early warning is triggered. When the model recognizes flames and outputs flame, the secondary alarm is triggered.

[0070] In this example, we first collect images of real tree line faults through image acquisition equipment set up around the transmission line, and then use labelimg to label the images to obtain the initial data set, and then use the roboflow public data set to expand the three types of data sets in the initial data set, namely flames, cables, and branches, to obtain an enhanced data set to improve the generalization ability of the model. The enhanced data set is divided into train training set, val validation set, and test test set in a ratio of 8:1:1. Construct a DCN-YOLO model, including the backbone network backbone responsible for extracting features from the input image, the head network Head responsible for the decision-making of the target detection model and generating the final detection results, the neck network Neck responsible for feature fusion and enhancement, and introduces deformable convolution DCNv3 to improve the performance of processing complex deformed targets. The color attributes (brightness, contrast, saturation and hue) of the augmented dataset are randomly adjusted to enhance the model's ability to recognize targets under different lighting and color conditions. The image is randomly rotated up to 15 degrees to simulate the different orientations of the target in the image. A region is randomly cropped from the original image of the dataset and resized to 224x224 to help the model learn different sizes and proportions of the target. Gaussian blur is applied to the dataset images to simulate images at different focal lengths and enhance the model's invariance to details. A channel attention mechanism is integrated into the yolov8 model. The global information of the feature map is obtained through global average pooling and global maximum pooling respectively. The weight relationship between channels is learned through two 1×1 convolutional fully connected layers and ReLU activation function. Finally, the output is compressed to the interval [0,1] using the Sgmoid function as the channel attention weight. The model is improved by the focal loss function to reduce the weight of easy-to-classify samples and increase the weight of difficult-to-classify samples. The enhanced data set is input into the improved DCN-YOLO model for training, and the model is made to reach the ideal state by adjusting the data enhancement, attention mechanism and loss function parameters in the improved algorithm, and the trained and optimized DCN-YOLO model is obtained and deployed in practical applications for tree line fault identification tasks. When the optimized model detects the presence of branches and flames or one of the two on the transmission line, the alarm system will be triggered. The alarm system means that when the model recognizes flames and branches, it will output flame and branch respectively. Both flame and branch will enter the alarm system as input signals to trigger the alarm, completing the early warning of tree line faults.

[0071] This example has the following advantages: the present invention uses an enhanced data set containing a large number of real tree-line discharge images for marking and inputs the DCN-YOLO model for training, and uses the improved DCN-YOLO model with the highest mAP50 and mAP50-90 values ​​as a tree-line fault monitoring and early warning model, which can quickly and accurately identify whether there is a tree-line fault in the transmission line, realize real-time monitoring of the transmission line, and promptly discover possible tree-line fault hazards; through the alarm system, it can more quickly and clearly confirm whether there is a tree-line discharge in the transmission line, improve the integrated automation of fault identification and early warning, assist staff to promptly and effectively avoid the occurrence and expansion of tree-line faults, and improve the power supply reliability and safety of the transmission line.

[0072] The above-mentioned embodiments only express the specific implementation of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A tree line fault identification and early warning method based on the DCN-YOLO model, characterized in that: The following steps are involved: Step 1: Use the tree line discharge experimental platform to simulate real tree line discharge, shoot videos and export pictures by frame, and label the pictures to obtain the initial data set; Step 2: Obtain the flame, cable, and branch data sets to expand the initial data set to form an enhanced data set, and divide it into a training set, a validation set, and a test set; Step 3: Based on the DCN-YOLO model, data enhancement, channel attention mechanism improvement, and loss function sample weight balance improvement are performed; Step 4: Input the enhanced data set into the improved model for training and verification, obtain the highest evaluation index mAP50 and mAP50-90 values, and obtain the tree line fault monitoring and early warning model; Step 5: Connect the tree-line fault monitoring and early warning model to the transmission tower monitoring system. When branches and flames or one of the two are detected on the transmission line, the alarm system will be triggered.

2. The tree line fault identification and early warning method based on the DCN-YOLO model according to claim 1 is characterized in that: In step 1, image data when a tree line fault occurs is obtained by using image acquisition equipment installed around the transmission line, wherein the image acquisition equipment includes a movable camera and a visual monitoring device installed at a fixed node of a tower; The initial data set refers to the data set of flame, cable, and branch labels after labeling.

3. The tree line fault identification and early warning method based on the DCN-YOLO model according to claim 1 is characterized in that: The step 3 specifically comprises the following steps: Step 3.1, build the DCN-YOLO model, including the backbone network, which is used to extract features from the input image, the head network, which is used to make decisions for the target detection model and generate the final detection results, and the neck network, which is used for feature fusion and enhancement. The deformable convolution DCNv3 is introduced to improve the performance of processing complex deformed targets. Step 3.2: Randomly adjust the color attributes of the dataset to enhance the DCN-YOLO model's ability to recognize targets under different lighting and color conditions. At the same time, randomly rotate the image by up to 15 degrees to simulate the different orientations of the target in the image. Randomly crop an area from the original image of the dataset and adjust it to a size of 224x224 for the DCN-YOLO model to learn targets of different sizes and proportions. Apply Gaussian blur to the dataset image to simulate images at different focal lengths and enhance the DCN-YOLO model's invariance to details. Step 3.3: Integrate a channel attention mechanism into the DCN-YOLO model, obtain the global information of the feature map through global average pooling and global maximum pooling, and then learn the weight relationship between channels through two 1×1 convolutional fully connected layers and ReLU activation function. Finally, use the Sgmoid function to compress the output to the [0,1] interval as the channel attention weight; Step 3.4: Improve the DCN-YOLO model through the focal loss function, reduce the weight of easy-to-classify samples and increase the weight of difficult-to-classify samples.

4. The tree line fault identification and early warning method based on the DCN-YOLO model according to claim 3 is characterized in that: In step 3.3, the global average pooling is as follows: Where: F is the input feature map, H and W are the height and width of the feature map respectively, and c is the channel index; The global maximum pooling is as follows: The ReLU activation function is as follows: ReLU(x)=max(0,x) The Sigmoid function is as follows: Where: e is the base of the natural logarithm and x is the input value; In step 3.4, the focal loss function is as follows: Focal Loss(p t )=-α(1-p t ) γ log(p t ) Where: p t is the predicted probability of the model for each category. For positive samples, p t is the confidence of the model predicting the positive class; for negative samples, p t is the confidence of the model prediction as negative class; α t is the coefficient that balances the weights of positive and negative samples; γ is a focusing parameter that adjusts the weights of easy-to-classify samples and hard-to-classify samples.

5. The tree line fault identification and early warning method based on the DCN-YOLO model according to claim 3 is characterized in that: In step 4, the model is made to reach an ideal state by adjusting the data enhancement, attention mechanism, and loss function parameters in the improved algorithm, and the trained and optimized DCN-YOLO model is obtained and deployed in practical applications for tree line fault identification tasks.

6. The tree line fault identification and early warning method based on the DCN-YOLO model according to claim 5 is characterized in that: In step 5, the alarm system outputs flame and branch respectively when the trained and optimized DCN-YOLO model recognizes flame and branch, and uses flame and branch as input signals to trigger the alarm and enter the alarm system to complete the early warning of tree line failure.

7. The tree line fault identification and early warning method based on the DCN-YOLO model according to claim 6 is characterized in that: The early warning for tree line failure includes primary warning and secondary alarm. When the trained and optimized DCN-YOLO model only recognizes branches and outputs branch, the primary warning is triggered. When the trained and optimized DCN-YOLO model recognizes flames and outputs flame, the secondary alarm is triggered.

Images