6th International Azerbaijan Congress on Life, Engineering, Mathematical, and Applied Sciences, Baku, Azerbaycan, 20 Mart 2024, cilt.1, sa.1, ss.247-254
Plant diseases and pests pose significant threats in the agricultural sector, leading to crop damage
and decreased market prices. The focus of this study is the early diagnosis and treatment of diseases
affecting the potato plant, which has a significant consumption volume both globally and in Turkey.
Applying new technological advancements in this area will prevent these damages. This academic work
encompasses the development of a module for real-time plant disease and pest detection supported by
an artificial intelligence algorithm mounted on an unmanned ground agricultural vehicle. The module
consists of an assembly equipped on a 4x4 all-terrain robot, utilizing a model trained with the highly
accurate and fast Yolo algorithm controlled by a Raspberry Pi. It detects fungal diseases on potato
leaves, which are widely prevalent and pose the highest threat of damage to this plant. A Raspberry Pi
compatible camera, mounted on the moving robot in the agricultural field and capable of real-time video
recording, transmits changes observed on the potato leaves for disease detection. A servo motor is used
for an extensive viewing axis. This study compares the performances of the Yolov7, Yolov8 algorithms,
and the Yolov9 algorithm which was introduced as the latest version at the beginning of 2024, regarding
object detection accuracy and speed. The impacts of these models on the study are discussed through
observations. The results demonstrate the precise detection of fungicides on potatoes, which will benefit
the enhancement of product quality and the implementation of more refined agricultural management.
Additionally, the outcomes of the academic study will facilitate the development of an on-site spraying
system supported by a model that performs real-time and highly accurate disease detection