AUTOMATIC IDENTIFICATION AND CLASSIFICATION OF FILLING LEVEL OF COCA COLA BOTTLING USING IMAGE PROCESSING

Abstract

Digital image processing and analysis system was developed to detect and count normal and filling bottles in the production flow process. MATLAB (R2013a) image processing and acquisition toolboxes together with night flick camera intex (it-309wc) webcam were used to capture, import and analyze images. Essential size features such as area, perimeter, major axis and minor axis were extracted for detection abnormal bottles due to size defects. The bottles’ images were enhanced and binarized to segment the bottles from the background and subsequently detect and count the defective and normal bottles. For the purpose of referencing, twenty standard brand bottles obtained from the factory quality control office were used. In implementing our solution consider the functionally available in the matlab image processing tool kit. The developed algorithm was used to extract and the range of values were taken as a reference for defect detection of sample bottles. Beside in forty known bottle samples were taken. To verify the accuracy of the algorithm. Among those samples, 12.5% had height or length defects, the flow were interrupted three times and the remaining 87.5% were normal product. Beside on this 100% correctly identified .The research results showed that size level quality inspection, abnormality and irregular flow detection and count of bottles can be done using image processing, analysis and computer vision techniques. Faults with side bottles and missing bottles must be ignored by our system.