Home // CADAT 2024, The First International Conference on Accessible Digital Agriculture Technologies // View article
Authors:
Sho Nobumoto
Tsuyosi Nakajima
Kota Oshima
Yutaka Kaizu
Keywords: Irrigation ponds; Semantic segmentation; Distance measurement; Autonomous navigation.
Abstract:
Many irrigation ponds are at high risk of collapse due to issues, such as aging caused by insufficient management and sediment accumulation. The use of autonomous boats to measure water depth and estimate water storage capacity in these ponds has proven to be effective. However, there were some problems, such as discrepancies between the navigational map and actual conditions, which resulted in insufficient depth data collection near the water's edge or the boats running aground. In this study, we developed an autonomous boat system by improving the technology to utilize image recognition for identifying shoreline positions, allowing the boat to safely and adequately approach the water's edge. To achieve this, the boat's structure was modified by flattening the hull to reduce the risk of grounding. More importantly, three essential functions were also implemented: shoreline recognition, distance measurement from the boat, and route modification based on the collected data. These functions were evaluated in an actual irrigation pond, and it was verified that the distance estimation was accurate enough for the boat to safely navigate near the shoreline. However, several issues were identified, including the impact of boat sway and lighting conditions on recognition accuracy, as well as the need for improved recognition of obstacles along the route that are not part of the shoreline.
Pages: 11 to 15
Copyright: Copyright (c) IARIA, 2024
Publication date: November 17, 2024
Published in: conference
ISBN: 978-1-68558-324-8
Location: Valencia, Spain
Dates: from November 17, 2024 to November 21, 2024