Enhanced Temperature Monitoring System for Sinter in a Rotatory Cooler

Abstract

Temperature monitoring of moving material is a complex procedure that provides relevant information about thermal behavior. Temperature monitoring requires accurate tracking of moving material to calculate its temperature so temperature curves can be calculated. This paper proposes a temperature monitoring system for sinter material, with multiple potential applications in a number of different areas. First, a calibration procedure that translates images to real world units is developed. Calibration enables the calculation of images without perspective projection that provide reliable geometric information. Moreover, with these transformed images, the model used to estimate movement is greatly simplified. Second, a tracking procedure consisting of feature detection, feature matching, and model estimation is applied to the rectified images. Finally, multiple temperature decay curves are calculated simultaneously, providing relevant information about cooling behavior. Features extracted from these curves can be used to detect abnormal cooling behaviors, and to control and regulate the sintering process. Tests of the proposed system indicate excellent performance in terms of calibration accuracy, temperature measurement, description of the cooling patterns, and detection of abnormal conditions.