Optimasi Deteksi Dini Masalah Kesehatan Bibit Kelapa Sawit dengan JST MLP Berbasis Citra Termal

Optimizing Oil Palm Seedling Health Detection with Thermal Image-Based MLP Neural Network

Abstract

Palm oil has an important role in the palm oil industry, but health problems in the seeds threaten production results. This research advocates an innovative approach by combining thermal imaging technology and artificial intelligence, especially Multilayer Perceptron Artificial Neural Networks (MLP ANN), for early detection of health problems in oil palm seedlings. The use of thermal cameras makes it easier to measure the temperature of plants and the surrounding environment. Thermal image analysis helps in evaluating thermal characteristics, especially plant temperature, which may be associated with health problems. Temperature data is classified into normal plants and plants affected by health problems, using statistical analysis to strengthen the relationship. A predictive model using MLP ANN was formulated to correlate thermal characteristics with the health condition of oil palm seedlings. The research results show that this model has high validity, with R2 of 0.933 for calibration and 0.930 for validation. In essence, this method accurately predicts the health condition of oil palm seedlings based on thermal images. This approach has the potential to provide early detection of plant health problems quickly, accurately, and efficiently. Through the application of this method, it is hoped that it can reduce losses due to health problems in oil palm seedlings, thereby making a major contribution to increasing productivity and welfare in the palm oil industry.