Development of machine vision system by using tensorflow for object detection in manufacturing

Salimi, Muhammad Syazwan (2022) Development of machine vision system by using tensorflow for object detection in manufacturing. Project Report. Melaka, Malaysia, Universiti Teknikal Malaysia Melaka. (Submitted)

	Text (24 Pages) Development of machine vision system by using tensorflow for object detection in manufacturing.pdf - Submitted Version Download (699kB)
	Text (Full Text) Development of machine vision system by using tensorflow for object detection in manufacturing.pdf - Submitted Version Restricted to Repository staff only Download (3MB)

Abstract

This final year project aims to conduct a technical report on developing a machine vision system using TensorFlow for object detection in manufacturing. Before the advent of modern technology, human inspectors cause low consistency on quality control inspection due to fatigue and other disturbances. Next, human vision led to inconstant take up time to detect the defect on product using human sense. Moreover, it difficult to identify the product when it come out with same color and almost the same shape. Therefore, this final year project focuses on developing a high-consistent machine vision system by implementing TensorFlow and Python with constant take-up time for object identification. This project also aims to evaluate the accuracy of machine vision systems in identifying the product with the same color and similar shape quality control process in industrial inspection. The methodology involved in this study is the implementation of phyton, and the TensorFlow software algorithm is used in the development machine vision system for object identification for biscuit inspection. Next, the image collection and labelling will be the benchmark for real-time object detection to identify the defect in this project. Lastly, to evaluate the effectiveness of machine vision systems in identifying quality control processes, the analysis will be performed and measured using the data collection and data interpretation. It is to identify the effectiveness of using a machine vision system. This project's expected outcome is to perform a high-quality control process via object detection in manufacturing. The future recommendation is performing a higher specification of TensorFlow with counterbalance faster detection and higher accuracy. To overcome the problem, future analysis is needed to evaluate the effectiveness of object detection.

Actions (login required)

View Item