Development of autonomous drones to detect diseases on plant leaves of durian trees

Name: Jiranuwat Piriyasupakij

Organization : Huahin Vitthayalai School

Email : jiranuwatpiriyasupakij@gmail.com

Name: Ratchada Prasitphan

Organization : Huahin Vitthayalai School

Email : meteemaker@gmail.com

LCSH: Drone aircraft -- Automation

LCSH: Detectorst -- Automation

Abstract: Durian is a very important crop in the agricultural industry and Thailand's number one export crop. To produce the high-quality durian requires care that can quickly resolve problems of the durian trees. The use of technology can take care of and solve the problem of durian trees for farmers. It was found that the technology is expensive, making the farmers in the area inaccessible to the technology currently used, causing a direct impact on the quality of the produce, causing the product to not be in demand in the market. The researcher has designed and developed an autonomous drone to detect diseases on the leaves of durian plants. It provides agriculture with information to prevent and solve problems caused by durian leaf diseases leading to higher yield quality of durian. The structure of the autonomous drone measures 860 mm in length and 320 mm in height, made of plastic and carbon fiber, in the shape of a six-rotor drone. The control unit is divided into the flight control unit and the durian leaf disease detection camera control unit. The flight control unit uses the Pixhawk board as a controller with GPS modules, Telemetry modules, Receiver modules, and ESC modules through the Mission Planner control program. The camera control unit uses the Jetson nano board to connect with the 8MP USB Camera module. Roboflow platform to create a dataset and then train Yolov5 with Google Colab for object detection of diseased durian leaves. The automatic operation of the drone is divided into 2 operations. Manual operation is using Stabilizer mode and Loiter mode. In Automatic operation, use Auto mode to follow the waypoint In the test of autonomous drone to detect disease in durian plant leaves, the test was divided into two sections, the first section is the flight test with 2 flight modes, Manual flight test and Auto-flight test. The second part is an automatic flight in conjunction with a camera to detect diseases from durian leaves. Found that the manual flight in Loiter mode was the most accurate, accounting for 97.50%, and the automatic flying in a straight line at a distance of 100 meters, and a height of 5 meters, was the most accurate, accounting for 98.00%. Auto flight test together with the detection of plant leaf diseases of durian trees at a height of 3 meters has the highest accuracy, accounting for 80.66%. The installation is separate from and detection of the autonomous drone camera has a problem in terms of the focus distance of the autonomous drone and the speed of the autonomous drone must be appropriate for the detection to work in best efficiency

King Mongkut's University of Technology North Bangkok. Central Library

Address: BANGKOK

Email: library@kmutnb.ac.th

Created: 2023

Modified: 2024-04-23

Issued: 2024-04-23

บทความ/Article

application/pdf

BibliograpyCitation : In Institute of Electrical and Electronics Engineers. The 27th International Computer Science and Engineering Conference 2023 (ICSEC 2023) (pp.258-265). Bangkok : Institute of Electrical and Electronics Engineers

eng

©copyrights King Mongkut's University of Technology North Bangkok

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	ICSEC 2023pp.258-265.pdf	848.56 KB

ใช้เวลา

0.03308 วินาที