Detection of Basal Stem Rot Disease Using Deep Learning

Palm oil industry is an important economic resource for Malaysia. However, palm tree disease called Basal Stem Rot has impeded the production of oil, which caused significant loss at same time. The by a fungus known as Ganoderma Boninense. Infected trees often have little to no symptoms during early stage infection, made detection difficult. Early necessary allow sanitization and control efforts. Using Terrestrial Laser Scanning technology, 88 grey-distribution canopy images were obtained. pre-processed augmented before being used training testing deep learning models. capabilities Convolution Neural Network models in classification dataset into healthy non-healthy class tested best performing model was identified based on Macro-F1 score. Fine-tuned DenseNet121 model, recorded Macro F1-score 0.798. It also noted that Baseline showed relatively remarkable macro-F1 score 0.747, better than all feature extractor some fine-tuned suffered from overfitting due limitations. For future work, it recommended increase sample size utilize other CNN architectures improve performance progress towards detecting infection classifying multiple classes.