Research on Multiple Features Extraction Technology of Insulator Images

Abstract

The insulators are important components of high voltage transmission line. They affect the safety of electric power system. Computer Vision is proposed to extract the kinds of characteristic values of insulator images. These values can provide information for insulator detection and recognition, and then protect the power system safely. The actual insulators are segmented firstly by use of pixel statistical method. The textural features are extracted by use of Gray level cooccurrence matrix. The invariant moment features are extracted in the binary image. The geometric features are computed by local property and pick up the boundary contour. Finally, local features are used for detecting insulator features.