DEVELOPMENT OF LEMON GRADING SYSTEM BASED ON IMAGE PROCESSING

• Main Author: Reza Arifandie, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/53731
• Institution: Institut Teknologi Bandung

Lemon is a fruit with many benefits and can be graded based on ripeness, size, and defects. Color and size are the primary parameter for a consumer to grade a lemon quality. Nowadays, the process of grading the quality of lemon in Indonesia is still using manual process. Grading lemon manually is a costly process and consumes a lot of time and resources. An automatic lemon grading sistem can reduce the farmer’s cost for manpower. Using image processing techniques, the process of grading lemon manually can be digitalized. On this research, a system that could grade a lemon quality according to criterion by analyzing the color and size of the lemon. The system consists of Raspberry Pi, camera, and a lighting system for taking and processing the image of a lemon. Color and contour detection based on tresholding technique will be used to determine the quality of the lemon by size, ripeness, and defects. Three classifications type is done by the system which are classification based on color, size, and defects. Classification by color is done by applying HSI thresholding to lemon image while classification by defects is done by applying histogram equalization and binary thresholding. To classify the lemon by size, thresholding and contour detection is used. Using this method, the developed system have adequate perfomance. However, the system is still unable to replace the manual grading process because of it’s low grading performance. The system is able to measure the diameter of a lemon with 98,79% accuracy. The grading accuracy of the system is 47,08%. The system is able to grade a lemon in 0.343 second, which, relatively speaking, is quite fast.