CLASSIFICATION SYSTEM ON SEVERITY OF ANTRAKNOSE DISEASE BASED ON MACHINE LEARNING IN BIG CHILLI

• Main Author: Rizky Abadi Sutoyo, Mochammad
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/74699
• Institution: Institut Teknologi Bandung

Plant growth chamber is a chamber that is isolated the surrounding environment with a microclimate control that’s already deployed within, user can control the microclimate and monitor what’s happened within the chamber. Inside the chamber, there are parameters that can be controlled such as, temperature, humidity, and light intensity for the most optimal range of the plant being grown inside the chamber. growth chamber had various kind of uses by itself, not only to grow a plants, it can also be used to grown disease within for a plant, in this case, we will use Big Red Chilli (capsicum annuum L.) as the main fruit that will be grown inside the chamber, it will be infected with Antraknose virus, there are plenty of disease that can infect a chilli, especially ones that being grown outside of the chamber, thus there’d be a need to cater the need on how to allow the chilli to only get infected by Antraknose disease only, and monitor their growth every time. The growth of the disease will heavily needed an isolated space that can be controlled and moniored at all times, moreover, a machine learning model that can detect the intensity of the disease from the chilli will be needed, there’s also another important factor where the monitoring device should be able to be controlled so the user can sees the object better and the machine learning model can see better, in order to maximize their detection capability. That’s why, the creation of machine learning model will use Object detection, and from what the necessity as it written before, it is produced a model from YOLO version 8l with the success on detecting the target (mAP) 67.4%, and it’s already been deployed to the chamber main system, there’s also the monitoring device that can be controlled within the x and y axis according the user control, in addition, it can also take a picture within a button press. System also allow the user to control the chamber remotely using gateway, cloud server, and website that will show the sensor reading on the onsite chamber.