Performance Analysis of Surface Defect Detection Algorithm of Aerospace Parts via Confusion Matrix Approach
DOI:
https://doi.org/10.37934/araset.55.1.141153Keywords:
Surface defect detection (SDD), Confusion matrix, Surface quality, Injection moulding, Aerospace partsAbstract
This paper examines the development of surface defect detection algorithm of a real plastic injection moulding parts from an aerospace part supplier company. Plastic injection moulding has evolved into one of the most essential and commonly utilized polymer processing activities in the plastics industry today. The work being done attempts to accomplish three goals: identify the frequency of defective injection moulded parts produced, develop an algorithm to detect defective parts using image processing techniques, and evaluate the effectiveness of image processing techniques for quality inspection in terms of accuracy and processing time. According to the protocol, there are five phases to this project: fault frequency detection, technique selection for image processing, MATLAB coding development, MATLAB coding debugging, and automatic inspection analysis. In this project, the algorithm and Graphical User Interface (GUI) are generated using the MATLAB Simulink program. A total of 100 parts samples were inspected, and the accuracy attained was 96% while processing time was reduced by 8.81 seconds.