Research On Target Extraction Technology Of Power Equipment Infrared Image Based On Immune Algorithm

With the growth of domestic and industrial power,the load of power equipment is also increasing.Timely fault detection of power equipment is the premise to ensure the stability of the power system.At present,instrument monitoring is widely used in power equipment monitoring in China,and operation information is obtained by parameters of equipment operation.Monitoring technology is gradually developing to non-contact,convenient,multi-functional and intelligent direction.As a mature technology,infrared detection can lock the abnormal part of equipment by the infrared image which could achieve the real-time feedback of temperature distribution without power failure and contact.At the same time,the infrared image which can detect the temperature abnormality in time has high sensitivity to the temperature,and realize the early warning of the power equipment fault to avoid the serious loss.The operation characteristics of power equipment obtained from infrared image target extraction can provide basis for intelligent detection and decision-making.It has theoretical significance and practical value to carry out this research.In this paper,by analyzing the characteristics of infrared image of power equipment in different operation states,three methods of infrared target extraction based on immune algorithm are proposed to achieve the automatic acquisition of effective targets in infrared image of power equipment.The main research contents are as follows:(1)An infrared target extraction algorithm based on the optimal immune domain complement immune network is proposed.It solves the problem of target extraction of power equipment infrared image which has uneven temperature distribution and close distance between fault area and fault warning area.The classification features of power equipment in different fault states are extracted by designing complement immune classifier,so as to achieve the accurate extraction of target areas.(2)A method of infrared image data classification based on complement immune network is proposed.It solves the problem of target extraction from infrared image of power equipment with no obvious fault area.Through Hilbert transform,the relative position features of the image are obtained in the functional space,and the Euclidean distance is used as the training range of the target area to achieve the accurate extraction of the target area of the infrared image.(3)A complement immune clustering algorithm is proposed.It solves the problem of complex background and overlapping traces in infrared detection of power equipment,which makes it difficult to extract the target power equipment effectively.The algorithm establishes several image databases of power equipment template,and determines the contour and position of the target area through the matching degree of template and image.K-means clustering is applied to the image,the clustering center is determined again,and the image of the target area is extracted.Qualitative analysis and quantitative analysis are used to analyze the experimental results of the proposed immune algorithm,and compared with other image extraction algorithms such as threshold method and watershed algorithm.The experimental results show that the proposed immune algorithm has the best segmentation effect.The validity and accuracy of the immune algorithm are further confirmed by the True Positive Rate,False Positive Rate,Dice and Accuracy.