Robotic System of Optical Control and Data Acquisition for Analyzing the Physical Properties of Industrially Produced Mineral Fertilizer Granules

Dmitrii Iunovidov, Viktor Shabalov, Kirill Menshikov

Abstract: This paper describes a robotic optical system for analyzing the physical properties of mineral fertilizer granules. It is designed for industrial conditions and can be used to accumulate data about the area, size, shape and color of granules. The system includes a rotary sampling device, a vibration unit for moving sample to/from the analysis area and an optical control unit with a computer (including a programmable logic controller, light and data storage system). The equipment operates as a standalone unit and can be integrated into any data network configuration (can be a server itself). System is equipped with a self-diagnostic algorithm with 4 types of errors, works as a standalone Samba server and provides access to any standardized Manufacturing Execution Systems. The software principle is to obtain a digital image with several OpenCV transformations (adaptive equalization of the image, adaptive image filtering, image binarization, morphological transformations, separate nearby granules with distance calculation and approximate granules as closed contours). Further, an ellipse approximation of each found contour with calculation of its physical parameters is performed. The description of the main nodes of the system, the type of calculation of the physical parameters and results comparison with standard industrial control methods are described in the article. The possibility of using this system as standalone device for accumulating data about granule properties is shown. Developed system pass industrial tests and successfully operates in the control of finished products at the JSC “Apatite”, Cherepovets city.