Research On The Key Technology Of Image Identification Of Insulator Fouling Level

Insulators mainly play a role in supporting wires and electrical insulation in transmission lines,and their operating status is an important factor affecting the safe operation of power systems.During the long-term operation of the insulator,dust particles in the air will continue to deposit on the surface of the insulator under the combined action of gravity and electric field force,and eventually form a polluted insulator.Polluted insulators can easily cause pollution flashover after being damp,affecting the normal operation of transmission lines,and even causing large-scale power outages in severe cases.Therefore,it is very important to study an efficient method for identifying the pollution level of insulators.With the continuous construction of smart grids,the method of identifying pollution levels of insulators has begun to shift from manual inspection to intelligent inspection.However,the current image recognition method of insulator pollution level does not consider the influence of ambient light.In this paper,digital image processing technology,combined with environmental lighting factors,is used to study the key technology of insulator pollution level image recognition.The main research work of this paper is as follows:First,based on the Kubelka-Munk theory,a model of insulator pollution color characteristics is established to verify the effectiveness of insulator pollution grade classification based on color characteristics,and to analyze the influencing factors of insulator pollution grade identification.In order to eliminate the deviation of the color characteristics of low-light insulator images,a low-light image color restoration algorithm based on improved retinal theory is proposed.The algorithm decomposes the original image into high-frequency information sub-bands and low-frequency information sub-bands through dual-tree complex wavelet transform,using an improved optimal estimation method of illuminance information to remove illuminance information in low frequency subbands,and use Butterworth filter to remove the noise in the high-frequency subband,and achieve the purpose of color restoration of the low-illumination polluted insulator image.Then,the U~2-Net network is used to construct the insulator target recognition model for insulator target recognition,and then the color features of the contaminated insulators are extracted to obtain a total of 36-dimensional feature quantities.In order to improve the recognition accuracy of the insulator pollution level,the Fisher criterion is used to screen and obtain the characteristic quantity with the strongest classification ability,which is used as the basis for the recognition of the insulator pollution level;a self-learning sparrow search algorithm with escape mechanism is proposed.The algorithm first uses the SPM chaotic mapping rule to initialize the population,which lays the foundation for global optimization;The golden sine algorithm is used to improve the location update rules of discoverers,and solve the problem of the smaller search dimension of sparrow;at the same time,guide the individual sparrow to conduct self-learning,so that the historical location information can be fully utilized;and use the escape mechanism to scatter the gathered individuals to the rest of the search space.The simulation experiment of the benchmark test function shows that the optimization performance of the proposed algorithm is improved significantly,which verifies the effectiveness of the improved strategy.Then the improved sparrow search algorithm is used to solve the input parameters of the support vector machine,and then the insulator contamination level recognition model is constructed through algorithm training.Finally,the insulator pollution level recognition model is applied to the polluted insulator inspection of transmission lines,and the recognition accuracy under normal light and harsh low light environment reaches 92%and 90%,respectively.The experimental results show that the recognition accuracy of the insulator pollution level recognition model constructed in this paper is good,which can provide a basis for the recognition of the transmission line insulator contamination level.