Coconut Tree Detection in Coastal Areas with Fast-RCNN Using Resnet-50

• Published in: 2023 Global Conference on Information Technologies and Communications (GCITC) pp. 1 - 4
• Main Authors: Prasad, J. V. D., Chandra, M. Kushwanth, Akash, J., Vivek, K.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2023
• Subjects: Biological system modeling; Coconut Trees; Fast R-CNN; Remote sensing; Residual neural networks; Sea measurements; Task analysis; Training; Tree Detection; Vegetation
• ISBN: 9798350308143
• DOI: 10.1109/GCITC60406.2023.10426612

Remote sensing and image processing are crucial in analyzing and enhancing satellite and aerial imagery data to extract valuable information and improve its quality across various applications. In particular, the coconut tree holds significant economic and ecological importance for many tropical developing countries. Detecting coconut trees in coastal areas is a valuable research endeavor, enabling efficient administration and monitoring of coconut plantations. However, manual detection is impractical and time-consuming. To address this, we propose a Fast-RCNN-based object detection algorithm, incorporating machine learning techniques, to achieve precise coconut tree detection in coastal areas using remote sensing data. To tackle the segmentation task, we utilize the Fast R-CNN model with ResNet50 architectures. Multiple experiments were conducted, varying configuration parameters to determine the most effective settings, ensuring a detection confidence level exceeding 90%.