Techniques of vision inspection aided by kalman filter in medicine packing production
Hongxia Cui, Jiaqi Liu · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2015
In order to avoid dependence on hardware devices and realize rapid vision detection of medicine defects on packing machine, a vision-based inspection method was proposed by the combination of kalman filter and image processing techniques. A kalman filter model was established to predict medicine movement and used to control the TV camera. The image detection results were adopted to update the kalman filter measurement in real-time inspection process. In this way, the images for medicines packed in the form of aluminum-plastic blister were achieved at a fixed position of the transferring platform. Experimental results indicate that the estimation approach was accurate enough to be capable of triggering vision inspection without proximity sensors used in traditional vision inspection systems.