Multi-Modal Imaging-Based Foreign Particle Detection System on Coal Conveyor Belt

In various raw materials carrying conveyor belts of steelworks, the occurrence of foreign particles like concrete boulders and iron rods is common during the transfer of material from mines to mills. These particles can damage equipment in the mills while conveying or crushing in grinding mills leading to breakdown and low equipment availability. To detect these particles in the raw coal conveyor belt of the G furnace, an image processing-based solution was developed, and trials were conducted to test the suitability of the product. A multi-modal imaging (Polarization camera)-based system was used to differentiate the material and color properties of foreign objects from native raw materials. Using the polarization camera, we can generate two images having heterogeneous sources of information related to the material and color properties of the scene under consideration. These raw image frames are then parallelly processed using a series of image processing techniques to extract/detect the foreign object from native raw material. Once these objects are detected in any of these heterogeneous images, the decision block tracks the foreign object in subsequent frames. Based on some predefined threshold, a decision is taken to alert the operator of the detected foreign object. While implementing this system in Tata steel, the alerts were generated in a local machine controlled by the operator, and remote alert using a cloud-based platform was also established for real-time detection of foreign objects from native raw materials. The operator can then stop the belt and remove these foreign objects before it causes any damage to the plant's functioning. Two successful trials have been conducted in the G furnace coal conveyer belt, and the system is ready to be implemented.