Zenramdhanu Wahyudistyo Sunarta, Muhammad Cahyo Bagaskoro, Aripriharta, Siti Zubaidah, Mokh Sholihul Hadi, Nandang Mufti, Gwo-Jiun Horng
This study focuses on the development of an IoTbased flora monitoring system, called SobatFlora, which integrates flora fingerprint technology and machine learning to identify plant species and monitor environmental changes in the Brantas River Basin. The system utilizes cameras, sensors, and advanced algorithms to analyze plant characteristics and detect alterations in the environment. The findings indicate that SobatFlora successfully achieves high detection accuracy, reaching up to 87% for certain plant species, with stable electrical parameters such as a voltage range of 4.97-4.98 \mathrm{~V} and a current of approximately 197.3 mA. These results highlight the system's capability to provide continuous and reliable monitoring, particularly in areas with limited access to electrical resources. The SobatFlora system proves to be highly effective for environmental monitoring and disaster mitigation, especially in areas prone to natural disasters like floods and landslides. However, the system faces challenges in detecting plants with complex textures, such as moss, which may be influenced by environmental factors like lighting, contrast, and texture similarity. Despite these challenges, the overall performance of the system remains promising. The study suggests that with further optimization, such as improving texture recognition and adapting to varying environmental conditions, the system can achieve even greater accuracy and reliability, making it a valuable tool for ecosystem conservation and disaster prevention. © 2025 IEEE.
Universitas Negeri Malang Malang, Department of Electrical Engineering and Informatics, Indonesia; Universitas Negeri Malang, Department of Biology, Malang, Indonesia; Universitas Negeri Malang, Department of Physics, Malang, Indonesia; Taiwan University of Science and Technology Taiwan, Department of Electronics Engineering Southern, China
