Infrared short-circuit detection for electrolytic copper refining

摘要

This paper proposes an automatic detection method for short-circuit fault that based on the thermal radiation principle of infrared image. During copper electrolytic refining, short circuits between cathode and anode plates will lower the production efficiency. It is necessary to detect short circuits timely to reduce the electricity loss. Firstly, the positive and negative samples were collected that came from the infrared images segmentation of the electrolytic tank images. Then, pixel ordering PCA feature extraction algorithm is proposed to obtain the samples feature. Finally, SVM classifier is used to recognize the short circuits. Experiment results prove that the recognition rate based on proposed method is better than other algorithms, and this method has been applied in the electrolytic copper factory.