Detection and classification of pills moving at a very high speed using a Bi-IV2 vision system

In this paper, a novel class of information-processing systems called cellular neural networks (CNNs) is overviewed. The CNN provides solutions to some real-life problems which are difficult or time-consuming for classical digital computers. The CNN has emerged as a powerful and practically realizable paradigm of multidimensional, locally connected, nonlinear processor arrays. We will introduce a brief concept, mathematical description of CNN, and develop an application of CNN in the ultra-high frame-rate image processing: detection and classification of pills moving at a very high speed using the Bi-IV2 vision system. We also show you our experiment setup in the laboratory, and then experimental results are introduced. The classification is based on the size and shape parameters of different pills and possible missing parts are identified. The implementation of pills classification algorithm runs on Bi-IV2, the program to be run on Bi-IV2 is loaded over Ethernet, and the host computer PC can write/read data to/from the Bi-IV2 over Ethernet.