Prediction of SPAD values for paddy leaves using smartphone-based direct contact imaging method

Abstract

The Soil Plant Analysis Development (SPAD) value obtained from the SPAD meter is directly related to leaf chlorophyll content. The chlorophyll content is related to nitrogen content means the amount of fertilizer of a crop. Therefore, determining SPAD value is directly involved with crop health. Minolta SPAD meter can directly measure this value, and this is a well-established method in the research field for measuring chlorophyll content. Still, this instrument is too costly, which is beyond a farmers’ reach in the perspective of Bangladesh. The purpose of this study is to predict the SPAD value for paddy leaves using the smartphone-based direct contact imaging method to estimate the chlorophyll content of a paddy leaf. Numerous features were extracted from each image to predict the SPAD values. The features were then used as parameters in the multiple linear regression model. The models' performance was evaluated using images captured from a paddy field using a Minolta SPAD-502 Chlorophyll Meter. The multiple linear regression model's R2 and root mean square error (RMSE) values were 0.71 and 3.6512, respectively. Therefore, this result confirms that the direct digital contact imaging method has the potential to quantify the SPAD value of paddy leaves accurately. However, these results could be more accurate if the image acquisition was made from the seedling to mature stage of the paddy. In the future, an android app will be developed using this value which can directly measure the chlorophyll content of paddy leaves.