Page 50 - FYP Abstract BETA Mech, Mechatronics, MD_Neat

Final Year Project 2020/2021

Multiple Object Detection using Deep

Learning in Manufacturing Workshop

Student’s Name: Nurdiyana Binti Othman

Supervisor’s Name: Mohamed Yusof Radzak

Supervisor’s Email: myusofr@unikl.edu.my

Abstract Object detection is become important for understanding plans in manufacturing
workshop. In this work, development of own dataset and object detection architecture
originally designed to detect two classes of machine in images and trained the dataset. The
two classes are lathe and vertical milling machine. In this paper, the proposed method is using
to detect the object is by using Faster RCNN object detector because this method suitable for
detection of big object with high accuracy. The objective of the study is to develop dataset
machine in manufacturing workshop and to develop deep learning architecture. Besides,
Region proposal networks is used to proposed algorithm based on Convolution Neural Network
for the detection and the detection of object by using own dataset has many of step of
development. The RPN generates region proposals, which give the region if interests into the
RCNN network as input. The two networks can then be combined into a single network by
sharing their convolutional features in order to detect a specific object in a given image.

Keywords Object detection, Deep learning, Faster RCNN, Faster RCNN architecture.

Bachelor of Engineering Technology (Hons) in Mechanical (Automotive) 48
Bachelor of Engineering Technology (Hons) in Mechatronics (Automotive)
Bachelor of Engineering Technology (Hons) in Mechanical Design