SVD-based burning state recognition in rotary kiln using machine learning

How to recognize the burning state efficiently and accurately in sintering process of rotary kiln has become a significant issue. In this paper, a novel recognition method based on SVD and SVM is introduced. It can determine the burning state in real time. The SVD-based feature between real time frame and two libraries, which consist of normal burning state and under burning state respectively, is extracted as the input. Then the SVM classifier is trained offline, which simplifies the computational complexity significantly in practical recognition process. The proposed method has the advantage of high accuracy and low time complexity, which makes it achieve the real time recognition in rotary kiln. Moreover, a multi-label recognition method proposed in this paper can divide burning state into several states. The simulation results show the effectiveness of the proposed method.