Determination Of The Optimum Shooting Conditions In Estimating Paddy Leaf Color Level By Using a Handphone Cameras

Abstract

Leaf color is an indicator that can be used to predict paddy fertilizer need. Handphone cameras can be used to measure the leaf color replacing the function of leaf color chart as the handphone technology has spread to farmer level. This research continues the previous research that still has low accuracy. The objective of the research is to determine the optimum shooting conditions in the estimation of paddy leaf color level using a handphone cameras. Rice leaf is slipped between user's finger on his/her palm, and the image is taken under body shadow. This image is then processed to get the RGB (red, green, blue) color components and then formulated with KNN to determine the color level. The KNN formulation was for every mobile phone type. It was found that the accuracy is affected by light intensity and the full fillment image frame. The best accuracy was found at 800-5000 lux light intensity and with a full frame image palms. The best accuracy was found to best shooting conditions with a image frame full of palm skin image at low light intensity, Samsung Ace was 87%, Sony Ericsson SK 17i was 93%, LG P698 was 91%, Samsung GT was 90%, Nexian was 93%.