[6-1015-D] Other Categories (3)

[6-1015-D-04] Deep Learning and Multiple Sensors Data Acquisition System for Real-time Decision Analysis in Agriculture Using Unmanned Aerial Vehicle

*Yunyan Xie¹, Ryozo Noguchi², Tofael Ahamed² (1. Graduate School of Life and Environmental Sciences, University of Tsukuba(Japan), 2. Faculty of Life and Environmental Sciences, University of Tsukuba(Japan))

キーワード：UAV, Machine Learning, Deep Learning, Multiple Sensors

This research was conducted to develop a user-friendly application to connect multiple sensors while using UAV to collect field data. The onboard and ground sensors were connected in the same application for ease of data collection in one software application. In the onboard sensors, thermal and RGB cameras were connected and transmitted the images within 500 m – 1000 m range. The soil moisture content information, humidity information were collected. In addition, the image analysis and deep-learning algorithm was added to the classification of the objects while landing. Histogram of Oriented Gradient (HOG) and Support Vector Machine (SVM) and YOLOV3 algorithms were implemented for classification of human, vehicle and others obstacle. The Michihibiki module was also connected with IoT application to soil moisture content measurement in the larger fields. The user application is divided into three modules: Hardware Module for Sensors Networks (HMSN), Software Module for Data Acquisition (SMDA), and Deeping Learning for Decision Analysis (DLDA). This research will be extended further with real-time analysis and decision support systems for UAV-based agricultural operations and safety systems.