Feature Information Extraction And Status Recognition Of Rotary Kiln Based On Digital Image

Rotary kiln is a kind of large-scale thermo-equipments for material sintered, which widely used in cement, papermaking, environmental protection, metallurgy and other industries. The industrial production in the rotary kiln burning zone during in the course of the temperature measurement and monitoring in the entire industrial production process.As the rotary kiln clinker sintering process is a very complex physical and chemical, with strong interference, nonlinear, large time delay and other characteristics, is difficult to accurately describe the mathematical model, the traditional control methods are often unsatisfactory results. To see the actual production of fire are often used in manual mode, but the workers labor-intensive, difficult operating environment, the monitoring results also have considerable randomnes. After long-term observation study found that, from inside the burning zone of rotary kiln flame image collected contained a lot of information on the temperature of burning zone. This article aims to use digital image processing technology to simulate human to see the fire, flames from the rotary kiln for burning the image to extract the image feature information, feature information according to the discriminant model for the establishment of the kiln burning is a more effective intelligent control methods.In this paper, using median filter algorithm Kiln collected a large number of video images to filter out interfering information, to receive follow-up with clear easy to image segmentation and feature extraction image. Went on filter out the images after the adoption of improved Otsu threshold segmentation algorithm for color image segmentation, to be burned with the flame zone, the material area, materials, height of area burned the image, and then be dealt with separately on the basis of two consecutive images over the flame zone, the material area, material altitude regions of HSI were used to improve optical flow algorithm, get the three regions of the optical flow features. In the optical flow computing, this paper first analyzes the optical flow method the basic assumptions and principles of the algorithm has been studied in recent years, research types and presents a HSI color model based on the weighted optical flow algorithm for color image sequence, using this method of image sequence of operation, can increase the recognition accuracy of the moving objects. Finally, based on the establishment of optical flow field vector histogram, statistics vector information, and the establishment of discriminant curve of combustion within the rotary kiln.Actually collected in the multiple color image of the rotary kiln flame application of this method, experimental results show that the use of the method proposed in this paper that the use of image processing techniques to determine the kiln firing zone of the combustion condition is valid.