Contour-based object detection in Automatic Sorting System for a parcel boxes

In the Automatic Sorting System, precision volume and size of the parcel boxes are all concerned. Different box size will certainly have a limited of personnel who visually identify and pick the objects. Computer vision are related to image processing and image analysis tend to focus on 2D image by pixel operation. Contour-based object detection is one alternative can measure the area for objects. In this paper will be in design automation development of a computer vision system that is able to get the volume of the parcel boxes. To get this this value if know the amount of length, width, and height. To find out the dimensional scale of parcel will be used two webcam cameras by calculating the pixels that are captured on camera and making comparison for calibration. The 2D image consists of two images from camera captured with vertical and horizontal view. After getting the length, width, and height of the parcel box, there will be a multiplication program is used to obtain the result of volume. It be separated automatically on the conveyor belt. The system can identify boxes to within 1 - 15 cm in length and width and within 5 - 20 cm in height. For the evaluation, some of boxes were sorted into three categories. The experiment result showed that the Automatic Sorting System is able to sort them out with an accuracy of 87.5%.