Development of Algorithm for the Tomato Identification Using MATLAB

Khong, Chung Hang (2018) Development of Algorithm for the Tomato Identification Using MATLAB. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Khong Chung Hang.pdf Restricted to Registered users only Download (2MB)

Abstract

Tomato, a member of the nightshade family, most of us will classify this as a vegetable because of the affinity savory ingredients, but in fact this is fruit. Skin, seed, and flesh can all be eaten, yet the leave is consist of toxic. Harvest process always is a tough event, because labors picker need to spend half of the day to collect all the ripe tomato. After that, classification of the tomato can be done according to the ripeness of the tomato. This is because the ripeness of the tomato will directly affect the shipping time and place. Sometime the whole process is not that efficient due to laziness and mistaken made by human, but this cannot be avoid. So, farming robotic had been invented to reduce the manpower and increases the productivity of the farm because robot will not make mistake. Harvest robot is a type of a farming robot, for a harvest robot, the machine vision is very important part of it, which use to determine and detected locality of tomatoes. In Malaysia, harvesting robot was not yet extensively implemented thus this is a challenge to let form more cost effective in harvest process, and increase the productivity. The objective of this research is design an algorithm system for identify the tomato locality. MATLAB will be used as the software to design the tomato identification algorithm. After image was capturing, it will undergo few images processing such as segmentation, deblurring and edge detection to determine the locality of the tomato. Noise remover also will be implemented in the beginning of the design in order to enhance the images. In order to enhance the edge detected operation, 2R-B-G color values technique was applied, tested, and verified in this project. This is because it aid in the red color pixels in look significantly and more easily recognize by the algorithm. A GUI had been designed for the conveniences of the user in the aspect of image selection, image processing and results comparison.