Recognition of Main Transformer Oil Conservator Oil Level Representation Number Based on FCM Segmentation

In the traction power supply system, the meter image collected by the inspection robot is difficult to read due to inaccurate positioning and insufficient light. To this, A low-light enhancement fusion algorithm is proposed, which combines adaptive fast bilateral filtering and a low-light image enhancement algorithm based on fractional-order improved Retinex, which effectively solves the problem of enhanced restoration of low-light images; at the same time, it combines perspective transformation and SIFT feature points The registration method corrects the geometric distortion image; finally, for the rectangular scale gauge of the main transformer oil conservator oil level gauge, a main transformer oil conservator oil level gauge recognition algorithm is designed that combines FCM and Otsu to segment. Experiments have proved that the readings obtained by this number recognition algorithm are basically the same as the manual readings, and the average recognition accuracy meets the requirements of practical applications. It can greatly promote the improvement of the recognition accuracy of outdoor instrument readings in traction substations under low-light distortion environments. effect.