PHYSICAL DESIGNS AND HARDWARE COMPONENTS IN IOT SYSTEM FOR EARLY/LATE BLIGHT DISEASE SEVERITY LEVEL IDENTIFICATION ON TOMATO PLANTS

• Main Author: Rozin, Fauzan
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/58375
• Institution: Institut Teknologi Bandung

Early/late blight disease is a disease that attacks tomato plants so that it can cause crop failure. To overcome this problem, superior varieties obtained from research for 3-5 years in a screenhouse can be used. The current research process is still manual, the operator takes plant photos and records its development every day. This process requires a lot of time, effort and is prone to inconsistencies when taking the photos and recording the progress. To answer these problems, the proposed solution is an IoT System and Machine Learning System for Early/Late Blight Disease Severity Level Identification on Tomato Plants. The solution consists of IoT and Machine Learning systems. In developing an IoT system, it is necessary to design physical designs (robotic trains, robotic arms, and camera systems) as well as hardware components. In the robotic train and robotic arms, a NEMA 23 57HS82 stepper motor is used, while in the camera system a MG996R servo motor is used. Data collection of plant photos used Arducam OV5647. All commands executed on the hardware are processed by the Raspberry Pi 4. This system has 1.5 meters horizontal range, 98 centimeters vertical range, and 175? and 170? slope range of upper dan side camera respectively. The system also has 4.9 centimeters horizontal resolution, 9,8 centimeters vertical resolution, and 8.2 centimeters and 8.6 centimeters slope resolution of upper and side camera respectively. At last, the data acquisition speed of this system is 8.23 seconds per photo.