Inspeksi Kualitas Untuk Pendeteksian Cacat Bentuk Pada Botol Minuman Plastik Berbasis Visi Komputer

Faisal Karim, Feri Candra


Product quality inspection system plays an important role in industrial production. Manual inspection process based on human tends to have several deficiencies in recognizing defects, such as workers subjectivity, inconsistency of work, and boredom level. Therefore, this paper presents an automated computer vision systems of plastic bottle shape defect detection for quality inspection system as a solution for the problem that has been raised. In this study, Mizone Lychee Lemon 500ml was used as sample. Digital Image Processing Technique is used to extract shape feature of plastic bottle. Through this technique, the defects of the bottle structure is described from the feature set such as Area, Perimeter, Major Axis Length, and Solidity. Then, the bottle is classified whether it is passed or rejected from the estimated parameters using Backpropagation Neural Network Method. A total of 100 data of bottle images are used in this study, consisting of 70 Training Images and 30 Testing Images. The result of this study is that the system can be used to differentiate plastic bottles according to shape with 100% accurary rate.
Keywords: Quality Inspection, Plastic Bottle, Computer Vision, Digital Image Processing, Backpropagation Neural Network

Full Text:



  • There are currently no refbacks.