期刊名称：IAENG International Journal of Computer Science
印刷版ISSN：1819-656X
电子版ISSN：1819-9224
出版年度：2021
卷号：48
期号：2
语种：English
出版社：IAENG - International Association of Engineers
摘要：The rapid detection of tree trunks is key to forest automation, inventory, and monitoring, enabling the use of tree-harvesting robots capable of navigation, tree counting, and tree measurement. In this paper, we propose a method called yolov3_trunk_model (Y3TM) to detect trunks rapidly using a convolutional neural network (CNN) and transfer learning. We use an enhanced yolov3 for object detection and an improved prediction strategy using feature pyramid networks (FPNs) for classification and boundary box determination of the tree trunks. Experimental results showed that our Y3TM offers a greatly improved recall rate of over 93% with a drastically average detection time of 0.3 s.
