Machine Vision Based Method For On-line Measurement Of The Pellet Size In Disc Pelletizer

Iron ore pellets are one of the important raw materials for ironmaking in modern blast furnaces.Disc pelletizer is an important equipment in pelletizing process,which is mainly used to produce iron ore green pellets with a certain size range and mechanical strength.The pellet size distribution(PSD)of green pellets is one of the most important product quality indices in pelletizing process.At present,the PSD in the disc pelletizer is mainly measured off-line by manual sampling and sieving,which is time-consuming and inefficient.Moreover,the measuring result lags behind,which may result in improper control of the pelletizing process and large process fluctuations.Such a situation cannot meet the urgent requirements of iron-making industry for high product quality and production efficiency.In order to solve the above problems and to automatically monitor the quality of green pellets,a machine-vision based automatic measuring method is proposed in the present work for online measurement of size distribution of green pellets in disc pelletizer.By analyzing the characteristics of the pelletizing process of the disc pelletizer,the static layer inside the pelletizer is determined as the image sampling area of the green pellets,and the hardware system for machine vision inspection is designed.The image of the static layer is obtained by industrial camera,and the green pellet region and shadow region are segmented from the image by improved threshold segmentation method.Then the pellets lying on the surface of the sampling area are recognized from the image by use of pellet markers and K-means clustering algorithm.After marking a small amount of shadows in green pellet region,a marker-controlled and region-expanded watershed algorithm is adopted to segment each pellet(including each pellet in overlapping pellets clusters)from the green pellet region,which solves the segmentation problem of overlapping green pellets.After that,the smallest enclosed circle fitting method is used to measure the size of each pellet,and the statistical analysis method is used to obtain the PSD information.Finally,With OpenCV and ALGLIB libraries,the above machine vision inspection algorithm is realized by using C++ programming language,and a set of software for measuring the PSD of green pellets in disc pelletizer online is developed.The method proposed in this paper was tested in a pelletizing plant and compared with the results of manual screening.The comparison results show that the proposed method is suitable for on-line measurement of the PSD of green pellets under different pelletizing conditions.The research results of this paper can be used to replace the manual screening to automatically measure the size of green pellets,and timely feedback the information of PSD to judge the pelletizing condition and the product quality,which is helpful to improve the automation level of pelletizing process in iron and steel industry.