Condition Monitoring of Substation Equipment Based on Machine Vision

Abstract

This article aims to reach a global level by selecting a Harr-like function with the participation of bubbles and light and shadow on a digital display, improving the fitness function of the genetic algorithm and the crossover change function. The classifier support vector machine parameters were optimized by an improved genetic algorithm. Finally, compared to other identification algorithms, basic component analysis reduces the number of characters through support vector machines and recognizes the target character. The results of the experiment showed that eight sets of experiments were performed on each character. Experiments have shown that character recognition is best achieved using Harr-like software and then SVM classification based on improved gene algorithms. For small samples, the sorting speed is also fast, which can meet the time requirement. In addition, the historical state data of the transformer are analyzed, which is consistent with the manual monitoring results, but the time is shorter; machine vision has been shown to be effective in monitoring the condition of substation equipment.