Summary of Research Introduction

126 CHAPTER 5 CONCLUSION

5.0 Introduction

This chapter discusses the conclusion and suggests future work to further improve the development of the system. It also stresses the significance and potential application from the research output.

5.1 Summary of Research

Based on this research works, this thesis presented the development of machine vision with additional of KUKA arm robot as assistive robot for windscreen gluing application. This system is presented with the Graphical User Interface GUI to ease users’ utilization. This thesis consists of two main parts, first mostly on vision system and the other is robotic system. For vision system, techniques of edge detection and template matching are used to create a shape-based matching method. It involves two phase processes of training and recognition phase. Fixed template creation of ROI replacing other feature extraction methods enables the system to be train to recognize various type of unknown objects instead of pre-determined objects. Median filtering technique helps in providing better image display by removing noise for better recognition rate. Template matching is introduced in performing pattern recognition for comparing between source images with 127 template images saved from training phase. Recognition rate known as the result for template matching where pixels equal recorded from both images. Two cameras are introduced in determining 3D transformation generate from two 2D images capture from both cameras. Harris Corner Detector HCD used to determine edge of the object with its location stated in x and y coordinates. The recognised edges then used to generate additional point by using mathematical equation. The combinations of two 2D images in generate 3D transformations by integrating HCD and mathematical equation of straight line to define Multiplication Factor MultFactor. MultFactor is the value in pixel location used to locate z coordinate for defects occur in the system by multiplying it with y coordinate generate from top camera. Robotic system, the automatic robot application is integrating in the system to add services of gluing process of the object. Data generate from vision system then being sent to robot system to do task according to the information given. Data generate must be transform to robot coordinate for ensuring the location recognised by the robot itself. It also able to do correction for each defect occur to ensure that reject gluing object still can be reproduced as perfect gluing object. This system tested by three experiments conducted to evaluate the effectiveness of the system. 1. 2D Matching Experimental setup based on classifying the types of process object to cope with the further process in the system. For example, this system running using pyramid object developed using white cardboard with dimension of 15cm x 15cm. So, the system will recognise the object as pyramid and pyramid algorithm being used until the system detect other object. 128 2. Classification of Types of Defects Experimental done in testing the effectiveness of the system in detecting glue defects occur in gluing pyramid. 50 tested images being run using this system with 236 defects appeared in the images. The purposed are detecting defects occur in the image and providing location for each defects in x, y and z coordinates. 3. Correction of Defects Experimental tested by using 15 tested images occur from inspection system. This application used to provide altering the defects in order to reduced rejected product. Correction done in two steps where 1 st CoD introduced first time correction and 2 nd CoD provides additional correction from 1 st CoD.

5.2 Achievement of Research Objectives