Research Of Insulator State Recognizing Method Based On The Sparse Representation Algorithm

Insulator in the transmission line plays the role of supporting wires and preventing current back to the ground, the operation process of transmission line is affected by factors such as mechanical load,the electric field and climate,causing the insulator fault. According to state grid, fault splitting or blackout accident in power system caused by insulator fault in the proportion accounted for about 80%.so, periodic inspection of transmission line is an important technical measure to protect safe operation of the power grid. The state of insulator in the transmission line is directly related to security and stable operation of the power grid, the aircraft inspection is an effective method to obtain the status of running transmission line. With the development of digital camera resolution and pattern recognition technology, making insulator fault recognition technology more practical and accurate. The task of recognizing insulator fault from massive aerial image information is very heavy, this paper proposes a new method of using the spare representation algorism to recognize aerial insulator fault.Firstly, we preprocess image operations such as gray processing, image enhancement, increasing quality of insulator image. Then, thin the preprocessing insulator image and detect the straight line of images by Hough transform to realize preliminary positioning of insulator according to the characteristic of insulator string, and use SVM Classifier to realize final positioning. Then, extract the feature vector of cracked and dropped insulator as over complete dictionary of the sparse representation classifier. Finally, uses the sparse representation classifier to recognize insulator fault. The experimental results show that the method is suitable for complicated aerial environment and the effect of cracked and dropped insulator recognition is better.