Studies on the state recognition of transformer fans based on image processing technique

Processes of digital image processing and image recognition of transformer fan state recognition are expounded in this paper. According to the features of transformer fans, analysis and studies on image processing algorithms such as annulus extraction, image segment and corrosion are given. To solve problem that the identification accuracy is directly affected by image preprocessing effects in the method of whole annulus criterion, a sub-annulus criterion method is proposed to implement fan state recognition in this paper. This method makes full use of the distribution feature of annulus pixels, divides the annulus into equal two parts at pixel number forming sub-annuluses and selects the proportions of fan blade pixels in sub-annuluses as the binary state feature quantities to realize state classification. Experimental results prove that this method can effectively reduce the high demand of recognition accuracy to image processing effects and it has advantages in many other aspects, such as simple computations, strong applicability and perfect recognition performance.