Classification of lidar point clouds using deep learning and machine learning methods
2025
0 views
0 downloads
Advisor: Dr. Öğr. Üyesi Emrah Benli
Abstract (EN)
The Lidarseg-mini subset of the NuScenes dataset was used in the study, and the point clouds were divided into seven different classes. In the experimental studies, both classical machine learning algorithms (Random Forest, K-Nearest Neighbor, ML-MLP) and deep learning-based methods (DL-MLP, CNN, PointNet Mini) were applied, and the models were compared based on accuracy, F1, sensitivity, and precision metrics.
Author
Dr. Nisa Koban
Institution
How to Cite
Nisa Koban (Master Thesis). Classification of lidar point clouds using deep learning and machine learning methods, 2025, Karadeniz Technical University.
Keywords
License
Tüm Hakları Saklıdır
This work is shared under the specified license terms.
More theses from Karadeniz Technical University
- Optimization of gold recovery from placer deposits using gravity methods(2025)
- Harşit çayından (Tirebolu-Giresun) elde edilen kırılmış dere malzemesinin beton agregası olarak kullanılabilirliğinin incelenmesi(2005)
- Yaşlandırma Süresinin Zn-27Al-1Cu Alaşımının Yapı ve Mekanik Özelliklerine Etkisi(2016)
- Prevalence and associated factors of tobacco use, alcohol consumption, alcohol use disorder among individuals aged 20 and above living in trabzon province(2025)
- Trabzon güney çevre yolu güzergahı Darıca (Akçaabat) - Yalı mahallesi (Trabzon) arasının mühendislik jeolojisi / Investigation of the planned route of the southern highway between Darıca (Akçaabat) - Yalı mahallesi (Trabzon) in terms of engineering geolog(2001)
- EXPERIMENTAL AND NUMERICAL INVESTIGATION OF FATIGUE BEHAVIOUR IN RADIAL JOURNAL BEARINGS MANUFACTURED FROM ZA-27 ALLOY(2025)