Determination of water stress for pepper from spectral reflections through artificial learning methods

Irrigation is one of the important issues that determine growing crops and soil quality in agriculture. With this in mind, irrigation should be done considering soil and climate factors. Plants and fruits need different amount of water every period. The amount of water that is not given is negatively affecting agriculture. In this study, it is tried to determine whether the pepper plant is in water stress by spectral reflections. For this purpose, spectral reflections from pepper leaves exposed to water stress, were recorded. Classification of these reflection data has been performed. The data set used in the study is divided into three groups. The first group (A) contains reflections from pepper leaves exposed to 100% water. The second group (B) contains reflections from pepper leaves exposed to 75% water. The third group (C) contains reflections from pepper leaves exposed to 50% water. Water stress detection consists of two steps. In the second stage, classification of the feature vectors related to the data is performed. The K nearest neighbors (KNN) and Artificial Neural Networks methods were used as the classification.