Classification Of Nutrient Deficiency In Lettuce Plants (Lactuca Sativa ) Using Machine Learning Algorithm

Abstract

Plants require appropriate nutrients or nutrients for their growth and development. Inappropriate nutrient levels
can interfere with the plant growth process, resulting in less-than-optimal harvest results. Therefore, it is very important
for farmers to know the nutrient levels of their plants, neither excessive nor lacking. Identification of nutrient deficiencies
in plants such as Lettuce (Lactuca Sativa) traditionally requires careful observation of the physical characteristics of the
plant, which is often long-drawn out and stand in need of a high level of accuracy. Leaf color is often used as an indication,
for example if it is pale or yellow it can indicate a lack of nitrogen or iron. This requires expertise and experience in
cultivation for lettuce cultivators. So, a tool is needed that can identify nutrient deficiencies accurately, quickly, and easily.
This study aims to overcome this challenge, namely identifying nutrient deficiencies in lettuce plants. This approach utilizes
machine learning technology to distinguish four main classes of deficiencies, namely: nitrogen (N), phosphorus (P), and
potassium (K), as well as normal or healthy lettuce leaf conditions. The proposed research method consists of the following
stages: 1). Lettuce leaf image dataset collection, 2). Preprocessing dataset, 3). Implementation of machine learning using
the Support Vector Machine (SVM) algorithm. In the implementation of SVM, experiments were carried out by applying
various SVM kernel spesifically: Linear, Polynomial, Radial Basis Function (RBF), and Sigmoid, 4). Evaluation of model
performance. Model performance was evaluated by measuring its level of accuracy in classifying nutrient deficiencies in
Lettuce leaf image data. The results of the experiment showed that SVM with the RBF kernel had the best accuracy, namely:
92%. The findings of this study provide valuable insights into the effectiveness of machine learning approaches in
classifying nutrient deficiencies in Lettuce plants. This study can help farmers to optimize their crop production more
efficiently and accurately.