TY - JOUR
T1 - Semantic Segmentation of Drone Images Based on Combined Segmentation Network Using Multiple Open Datasets
AU - Song, Ahram
N1 - Publisher Copyright:
Copyright © 2023 by The Korean Society of Remote Sensing.
PY - 2023
Y1 - 2023
N2 - This study proposed and validated a combined segmentation network (CSN) designed to effectively train on multiple drone image datasets and enhance the accuracy of semantic segmentation. CSN shares the entire encoding domain to accommodate the diversity of three drone datasets, while the decoding domains are trained independently. During training, the segmentation accuracy of CSN was lower compared to U-Net and the pyramid scene parsing network (PSPNet) on single datasets because it considers loss values for all datasets simultaneously. However, when applied to domestic autonomous drone images, CSN demonstrated the ability to classify pixels into appropriate classes without requiring additional training, outperforming PSPNet. This research suggests that CSN can serve as a valuable tool for effectively training on diverse drone image datasets and improving object recognition accuracy in new regions.
AB - This study proposed and validated a combined segmentation network (CSN) designed to effectively train on multiple drone image datasets and enhance the accuracy of semantic segmentation. CSN shares the entire encoding domain to accommodate the diversity of three drone datasets, while the decoding domains are trained independently. During training, the segmentation accuracy of CSN was lower compared to U-Net and the pyramid scene parsing network (PSPNet) on single datasets because it considers loss values for all datasets simultaneously. However, when applied to domestic autonomous drone images, CSN demonstrated the ability to classify pixels into appropriate classes without requiring additional training, outperforming PSPNet. This research suggests that CSN can serve as a valuable tool for effectively training on diverse drone image datasets and improving object recognition accuracy in new regions.
KW - Combined segmentation network
KW - Deep learning
KW - Drone image
KW - Semantic segmentation
UR - http://www.scopus.com/inward/record.url?scp=85177554932&partnerID=8YFLogxK
U2 - 10.7780/kjrs.2023.39.5.3.7
DO - 10.7780/kjrs.2023.39.5.3.7
M3 - Article
AN - SCOPUS:85177554932
SN - 1225-6161
VL - 39
SP - 967
EP - 978
JO - Korean Journal of Remote Sensing
JF - Korean Journal of Remote Sensing
IS - 5-3
ER -