A vision-based recognition method for transformer based on AdaBoost and multi-template matching

With the smart monitoring being widely concerned recently, substations have been introducing smart monitoring system. In this paper, we propose a vision-based recognition method for transformers in substation via combining with AdaBoost and a multi-template matching method. The proposed method works by dividing the whole process into two parts, namely coarse detection and fine recognition. In coarse detection, haar features of training samples in each sub-region are extracted and then AdaBoost algorithm is utilized for training and detecting. After coarse detection, we then perform fine recognition using multi-template matching with histogram intersection. Experimental results demonstrate that our method has a higher recognition precision and it is superior and more effective than the conventional AdaBoost method.