Video tracking for high-similarity drug tablets based on reflective lightness intensity and fuzzy recognition system

Video tracking of drug tablet exerts important influences on the efficiency and reliability of its mass production; this topic also becomes a difficult and targeted focus for pharmaceutical production monitory in the past several years due to the high similarity and random distribution of those objectives to be searched for. By measuring the reflective lightness intensity of illumination lightness on tablet surface, reflective lightness intensity matrix was established and demonstrated in the form of grey image, presenting its shape topology and topography details in return. On this basis, a series of mathematical properties for describing reflective lightness intensity images were proposed, thereafter a set of fuzzy recognition system and its identification rules can be employed to classify those moving tablets with inputted image properties, which facilitates the determination of their instantaneous coordinate positions on given image frame accordingly. By repeating identical operations on the next frame, the real-time motions of tablet objectives were traced successfully. Orthogonal tracking experiment and performance comparisons verified the accuracy and reliability of this new method in pharmaceutical industry. With original suggestions concerning imaging arrangements, tablet descriptions and video tracking, this article provides reliable references and new research ideas for tablet tracking performance in high-yielding production domain.