CN105138976A - A Genetic Wavelet Neural Network Based Icing Thickness Recognition Method for Transmission Lines - Google Patents

Abstract

The invention relates to a method for identifying the icing thickness of a power transmission line of a genetic wavelet neural network, which comprises the following steps: acquiring an icing image of the power transmission line through an image sensor; preprocessing the acquired image of the power transmission line; identifying the edge of the power transmission line image by using a genetic wavelet neural network; and calculating the icing thickness of the power transmission line. The method for identifying the icing thickness of the power transmission line has the advantages of high identification precision, reliable identification result and the like, can effectively calculate the icing thickness of the power transmission line, and can be applied to processing of icing disasters of the power transmission line.

Description

A Genetic Wavelet Neural Network Based Icing Thickness Recognition Method for Transmission Lines

technical field

The invention relates to the field of icing disaster monitoring of transmission lines, in particular to a method for identifying the icing thickness of transmission lines using a genetic wavelet neural network.

Background technique

With the economic development of the country, transmission lines play an increasingly important role. Since the working environment of the transmission line is mostly in the wild, it is easily affected by various environmental and meteorological factors, especially in some high-altitude and cold areas, the transmission line is easily affected by ice-covered disasters, resulting in failure of the transmission line, making the power system unsafe and reliable. Operation, affecting the normal production and life of the people, causing economic losses and social impact. Therefore, it has important practical significance and practical value for the safety and stability of power transmission to grasp the icing status of transmission lines in time and carry out the prevention and treatment of icing disasters.

Most of the existing image-based identification methods for transmission line ice thickness use traditional image edge detection operators. However, these operators are only suitable for limited types of detection, and are sensitive to noise and have poor adaptability.

The present invention proposes a genetic wavelet neural network method for identifying the icing thickness of transmission lines, which can effectively calculate the icing thickness of transmission lines through the identification of icing images on transmission lines and provide reference for the monitoring and treatment of icing disasters on transmission lines , has broad application prospects and economic value.

Contents of the invention

The invention proposes a genetic wavelet neural network-based identification method for icing thickness of transmission lines, which provides reference for monitoring and processing icing disasters of transmission lines, thereby ensuring safe and stable power operation.

In order to achieve the above object, the technical scheme adopted in the present invention is:

A genetic wavelet neural network identification method for transmission line ice thickness, comprising the following steps:

Step 1: Obtain the image of the icing of the transmission line through the image sensor;

Step 2: Preprocessing the acquired transmission line image;

Step 3: Use the genetic wavelet neural network to identify the edge of the transmission line image;

Step 4: Calculate the ice thickness of the transmission line.

The following specific steps should be included in the step 2:

Step 2.1: grayscale processing of the image acquired in step 1 described in claim 1:

(1)

In the formula, is the gray value of the corresponding pixel after being converted into a grayscale image, , , are the red component value, green component value and blue component value of the original image;

Step 2.2: To grayscale the image with Perform median filtering for the window.

The following specific steps should be included in the step 3:

Step 3.1: Normalize the preprocessed image obtained in step 2 of claim 1:

(2)

In the formula, is the normalized pixel gray value, is the preprocessed pixel gray value

Step 3.2: Using the Canny operator The window is used as the magnitude of the binary image detection window, and the 9-dimensional feature vector is obtained as the input of the genetic wavelet neural network. The pixel point is the edge point. The output of the wavelet neural network is 1, otherwise it is 0. The hidden layer neurons in the wavelet neural network The function is:

(3)

in, and is the scale factor for stretching and translation, , Take Morlet wavelet:

(4)

The neurons in the output layer select the Sigmoid function:

(5)

It can be expressed as:

(6)

in,

(7)

(8)

(9)

(10)

(11)

In the formula is the number of samples, , , , is the network learning rate, is the network momentum factor. Step 3.3: Specifically include:

Step 3.3.1: Initialize genetic algorithm parameters, including population size , cross probability , mutation probability and the maximum algebraic ;

Step 3.3.2: the number of hidden layer nodes in the wavelet neural network , initial weight , , network learning rate , , , , scaling factor , and the network momentum factor Use the real number encoding method as a set of chromosomes;

Step 3.3.3: Random generation group of initial chromosomes, constituting the initial population ;

Step 3.3.4: Define the fitness function as:

,

In the formula for the first The actual output of the sample output neuron, for the first The ideal output of the sample output neuron, calculate the population the fitness of each individual;

Step 3.3.5: Select the fitness proportional selection method for genetic operation;

Step 3.3.6: According to the crossover probability , perform cross operation;

Step 3.3.7: Select Mutation Probability , to perform a mutation operation;

Step 3.3.8: Get the next generation population ;

Step 3.3.9: Determine whether the maximum number of times or If the fitness of the optimal individual is greater than or equal to the predetermined value, if so, stop the operation, and decode the optimal individual into the parameters of the wavelet neural network, if otherwise, repeat steps 3.3.3-3.3.9;

Step 3.4: Use the pixels detected as edges by the Canny operator as training samples to train the wavelet neural network.

Step 4 should include the following specific steps:

Step 4.1: According to the output of the wavelet neural network, calculate the average profile width of the conductor after the transmission line is iced:

(12)

In the formula, for the first The width of the edge pixel outline;

Step 4.2: Calculate the average thickness of icing on transmission lines:

(13)

In the formula, is the known original diameter of the transmission line.

The beneficial effects of the present invention are:

A transmission line icing thickness recognition method based on a genetic wavelet neural network of the present invention can achieve the purpose of effectively calculating the transmission line icing thickness. Provide reference for disaster monitoring and handling to ensure the safety and stability of the power grid system.

Description of drawings

Fig. 1 is a flowchart of the present invention.

Fig. 2 is a structure diagram of the wavelet neural network in the present invention.

Fig. 3 is a flowchart of genetic algorithm optimization in the present invention.

Detailed ways

Preferred embodiments of the present invention will be described below in conjunction with accompanying drawings 1-3. It should be understood that the preferred embodiments described here are only used to illustrate and explain the present invention, and are not intended to limit the present invention.

A genetic wavelet neural network identification method for transmission line ice thickness, comprising the following steps:

Step 1: Obtain the image of the icing of the transmission line through the image sensor;

Step 2: Preprocessing the acquired transmission line image

Step 3: Use the genetic wavelet neural network to identify the edge of the transmission line image;

Step 4: Calculate the ice thickness of the transmission line.

The following specific steps should be included in the step 2:

Step 2.1: grayscale processing of the image acquired in step 1 described in claim 1:

(1)

In the formula, is the gray value of the corresponding pixel after being converted into a grayscale image, , , are the red component value, green component value and blue component value of the original image;

Step 2.2: To grayscale the image with Perform median filtering for the window.

The following specific steps should be included in the step 3:

Step 3.1: Normalize the preprocessed image obtained in step 2:

(2)

In the formula, is the normalized pixel gray value, is the preprocessed pixel gray value

Using Canny operator The window is used as the magnitude of the binary image detection window, and the 9-dimensional feature vector is obtained as the input of the genetic wavelet neural network. The pixel point is the edge point. The output of the wavelet neural network is 1, otherwise it is 0. The hidden layer neurons in the wavelet neural network The function is:

(3)

in, and is the scale factor for stretching and translation, , Take Morlet wavelet:

(4)

The neurons in the output layer select the Sigmoid function:

(5)

It can be expressed as:

(6)

in,

(7)

(8)

(9)

(10)

(11)

In the formula is the number of samples, , , , is the network learning rate, is the network momentum factor.

Step 3.3: Specifically include:

Step 3.3.1: Initialize genetic algorithm parameters, including population size , cross probability , mutation probability and the maximum algebraic ;

Step 3.3.2: the number of hidden layer nodes in the prediction model based on the wavelet neural network described in step 3 described in claim 1 , initial weight , , network learning rate , , , , scaling factor , and the network momentum factor Use the real number encoding method as a set of chromosomes;

Step 3.3.3: Random generation group of initial chromosomes, constituting the initial population ;

Step 3.3.4: Define the fitness function as

In the formula for the first The actual output of the sample output neuron, for the first The ideal output of the sample output neuron, calculate the population the fitness of each individual;

Step 3.3.5: Select the fitness proportional selection method for genetic operation;

Step 3.3.6: According to the crossover probability , perform cross operation;

Step 3.3.7: Select Mutation Probability , to perform a mutation operation;

Step 3.3.8: Get the next generation population ;

Step 3.3.9: Determine whether the maximum number of times or If the fitness of the optimal individual is greater than or equal to the predetermined value, if so, stop the operation, and decode the optimal individual into the parameters of the wavelet neural network, if otherwise, repeat steps 3.3.3-3.3.9;

Step 3.4: Take 1500 pixels of edge pixels detected by Canny operator as training samples, and 500 as training wavelet neural network.

Step 4 should include the following specific steps:

Step 4.1: According to the output of the wavelet neural network, calculate the average profile width of the conductor after the transmission line is iced:

(12)

In the formula, for the first The width of the edge pixel outline;

Step 4.2: Calculate the average thickness of icing on transmission lines:

(13)

In the formula, is the known original diameter of the transmission line.

The above-described embodiments are preferred embodiments of the present invention. It should be pointed out that those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. All modifications or equivalent replacements made within the spirit and principles of the present invention shall fall within the scope of the claims of the present invention.

Claims (3)

1. a transmission line icing thickness identification method of genetic wavelet neural network, is characterized in that, described method comprises: Step 1, acquire the icing image of the transmission line through the image sensor; Step 2, preprocessing the acquired transmission line image; Step 3, use the genetic wavelet neural network to identify the edge of the transmission line image, as follows: Step 3.1: Normalize the preprocessed image obtained in step 2: (2) In the formula, is the normalized pixel gray value, is the preprocessed pixel gray value; Step 3.2: Using the Canny operator The window is used as the magnitude of the binary image detection window, and the 9-dimensional feature vector is obtained as the input of the genetic wavelet neural network. The pixel point is the edge point. The output of the wavelet neural network is 1, otherwise it is 0. The hidden layer neurons in the wavelet neural network The function is: (3) in, and is the scale factor for stretching and translation, , Take Morlet wavelet: (4) The neurons in the output layer select the Sigmoid function: (5) It can be expressed as: (6) in, (7) (8) (9) (10) (11) In the formula is the number of samples, , , , is the network learning rate, is the network momentum factor; Step 3.3: the number of hidden layer nodes in the wavelet neural network , initial weight , , network learning rate , , , , scaling factor , and the network momentum factor Using the real number encoding method as a group of chromosomes and using the genetic algorithm to obtain the optimal individual decoding as the parameters of the wavelet neural network; Step 3.4: Use the pixels detected as edges by the Canny operator as training samples to train the wavelet neural network; Step 4, calculate the ice thickness of the transmission line. 2. the transmission line icing thickness identification method of genetic wavelet neural network according to claim 1, is characterized in that, in described step 2, should comprise following concrete steps: Step 2.1: grayscale processing of the image acquired in step 1 described in claim 1: (1) In the formula, is the gray value of the corresponding pixel after being converted into a grayscale image, , , are the red component value, green component value and blue component value of the original image; Step 2.2: To grayscale the image with Perform median filtering for the window. 3. according to the icing thickness identification method of the transmission line of the genetic wavelet neural network described in claim 1 or 2, it is characterized in that, in the described step 4, the following specific steps should be included: Step 4.1: According to the output of the wavelet neural network, calculate the average profile width of the conductor after the transmission line is iced: (12) In the formula, for the first The width of the edge pixel outline; Step 4.2: Calculate the average thickness of icing on transmission lines: (13) In the formula, is the known original diameter of the transmission line.