Insulator Recognition And Location Of The High Voltage Transmission Line Image

As an important infrastructure for the building of our people's livelihood,the power grid plays a decisive role in the transmission of electrical energy.In recent years,as the scale of the power grid continuous to expand,transmission lines continue to increase,and the area covered by EHV and high-capacity transmission lines will also continue to increase.Therefore,the security risks faced by the power grid are increasing.Insulator string as a very important safety component in the transmission line,in order to ensure its normal operation in the power grid operation,we need to focus on inspection of insulators in the inspection process of grid.The traditional manual inspection methods has been unable to meet the actual needs for the current EHV and high-capacity transmission lines,and poses a major threat to the life safety of patrol personnel.For check the insulators more convenient and safety,it is necessary to identify and locate the insulators from the aerial image under complex background.However,the identification and location of the target with a complicated background is a very hot and difficult issue in the field of pattern recognition.In order to solve the practical problem of insulator identification and positioning.In this thesis,according to the characteristics of insulator strings in high voltage transmission lines,research has been conducted from the following aspects.(1)The thesis first analyzes and studies the commonly used methods of insulator segmentation and extraction.On this basis,combined with the characteristics of complex image background of high voltage transmission lines,the method for detecting the saliency of images is studied.Therefore,a more accurate insulator string identification and location algorithm based on frequency tuning is proposed in this thesis.First of all,the inspection image containing the insulators is transformed from RGB color space into HSV color space,and uses the frequency tuning method according to the low-frequency and high-frequency information in the three spatial components of H,S and V to calculate the saliency information of the insulator strings in the H,S and V component images,and merged to obtain a saliency image of the insulator string.(2)After obtaining the saliency image of the insulator string,in order to extract the insulator string segmentation,this thesis continues to analyze the commonly used algorithm of image segmentation.The saliency image of the insulator string is processed using the Otsu segmentation algorithm and the morphological image processing dilation erosion operation to obtain the segmented insulator string.(3)For the segmented insulator string binary image,the edge detection algorithm of the image is used to detect the edge profile of the insulator string in the image,and then the insulator string is connected to the connected region for the identification of the high voltage transmission line through the connectivity of the insulator string.(4)Aiming at the tilt problem of insulator string in aerial image of transmission line,in this thesis,the mathematical model is constructed by Hough transform,the straight line where the insulator string is located is simulated,the angle of inclination of the insulator string and the horizontal plane is calculated from the information of the straight line,and then the tilt problem is solved by rotating the insulator string.(5)The effectiveness of the proposed algorithm is verified by comparison experiments.Under the Visual Studio development environment,using the OpenCV3 and Qt5 software development tools,the insulator string automatic identification and location interface of the aerial grid image under complex environment was developed.The experimental results show that the proposed method can obtain the information of the insulator string from the patrol image with complex background,and can accurately identify and locate the insulator string from the image,which can solve the tilt correction problem of the insulator string to some extent.It has high engineering application value and can provide important reference information for fault detection of insulator string.