Machine vision system for counting the number of corrugated cardboard

This paper presents a method for counting the number of corrugated cardboard using machine vision system and image processing techniques. The cardboard edge images that are strip or slitter and fluted or cut-off side were used for detection and counting. The gray level of edge images which obtain from the line scan camera are convert to binary images by threshoding. After that the strip lines of the slitter side are extracted by the first-order derivative and morphology then processed with MODE for counting. For the cut-off side, the fluted areas of cardboard are identified and measured using Blob analysis. Then the centers of the detected area are located and counted as the number of cardboards. MATLAB is applied to simulate in the experiments for determine the algorithms and perform the real-time experiments using HALCON to see the accuracy and performance of the system. The cardboard type BC, C and B were used in the tests. As experimental results, it is shown that the system provided the correct counting with tolerance of 2 cm and 3 cm in the case of the cardboard placed away further the camera focus for the slitter and cut-off side respectively.