Yazarlar
- Akbulut, Zeynep(2)
- Karsli, Fevzi(2)
- Akbulut, Z.(1)
- Bayram, B.(1)
- Dihkan, M.(1)
- Dihkan, Mustafa(1)
- Isikdag, U.(1)
- Karsli, F.(1)
- Kose, Oguz(1)
- Oktay, Tugrul(1)
- Ozdemir, S.(1)
Yayınlar
- MakaleAçık Erişim2025Scopus SJRQ1DAPNet++: density adaptive PointNet + + for airborne laser scanning data
Akbulut, Zeynep; Karsli, Fevzi
Addressing the challenges arising from the irregularity and varying density of Airborne Laser Scanning (ALS) point clouds, which particularly affect the performance and generalization ability of 3D Deep Neural Networks (DNNs), is essential for their effective application in direct semantic segmentation tasks. In this study, we investigated the underexplored aspect of adapting PointNet + + for semantic segmentation of ALS point clouds. We introduced Density-Adaptive PointNet++ (DAPNet++) to enhance robustness against variable point densities through modifications made to the original PointNet + + architecture. Our methodology includes controlled block partitioning based on point density replacing the original batching strategy. Another key advancement in the study is the automation of the initial value of the neighborhood search radius by taking into account the characteristics of the dataset. This approach optimizes receptive field determination, crucial for effective semantic segmentation. The effectiveness of DAPNet + + is validated through extensive experiments on various datasets, including ISPRS Vaihingen, DALES, subsampled DALES, and OpenGF. Notable improvements include up to 11% increase in weighted mean Intersection-over-Union (mIoU) on the highly variable OpenGF test dataset and 3% increase in mIoU on the subsampled DALES dataset. Furthermore, the generalization capability of the DAPNet + + was tested, revealing an approximately 5% improvement in evaluation metrics compared to PointNet++. In summary, DAPNet + + minimizes trial-and-error in the selection of parameters for block partitioning and radius, and it enhances robustness against variable point density. With the proposed approach, it has been observed that there is a significant improvement in accuracy for underrepresented classes, mitigating class imbalance in ALS point clouds.
AtıfScopus9WoS5OpenAlex6
İncele - MakaleYalnız künye2024Scopus SJRQ2A New Approach for Ground Filtering of Airborne Laser Scanning Data Using PointNet++
Akbulut, Zeynep; Karsli, Fevzi; Dihkan, Mustafa
Ground filtering is an essential process that separates ground and non-ground points for generating digital elevation models (DEMs) from airborne laser scanning (ALS) point cloud data. Traditionally, ground filtering was carried out using rule-based approaches, but with the advancements in deep learning networks, ground filtering can now be performed using these methods. Deep neural networks (DNNs) require large amounts of training samples to reveal key information within the dataset. OpenGF is a new ultra-large-scale dataset which has nearly half a billion points from nine different scenes ranging from highly dense urban areas to steep vegetation covered mountains for ground filtering task. In this study, PointNet++ architecture was employed for ground filtering on the OpenGF dataset. Firstly, we slightly modified PointNet++ architecture’s sampling strategy to decrease the number of duplicate points in the batching process. We proposed a new adaptive tiling strategy based on input point cloud density to cope with highly varying point sampling of the OpenGF dataset. Modified PointNet++ architecture was trained with both adaptive tiling strategy and original OpenGF tiles to investigate the advantages and drawbacks of the proposed strategy on the ground filtering task. To evaluate the performance of the proposed strategy, the results for different test areas were compared with those obtained using the original PointNet++ and rule-based cloth simulation filtering (CSF) algorithms. Despite using fewer epochs, the proposed data collection strategy achieved up to 13% higher accuracy on test scenes with the modified PointNet++ architecture compared to the network trained with original OpenGF tiles.
AtıfScopus2WoS1OpenAlex2İncele - BildiriAçık Erişim2024AN ANALYSIS OF NEIGHBOURHOOD TYPES FOR POINTNET++ IN SEMANTIC SEGMENTATION OF AIRBORNE LASER SCANNING DATA
Akbulut, Z.; Ozdemir, S.; Karsli, F.; Dihkan, M.; Isikdag, U.; Bayram, B.
The objective of the study is to conduct a comprehensive examination of how different neighbourhood types, namely spherical, cylindrical, and k-nearest neighbour (kNN), influence the feature extraction capabilities of the PointNet++ architecture in the semantic segmentation of Airborne Laser Scanning (ALS) point clouds. Two datasets are utilized for semantic segmentation analysis: the Dayton Annotated LiDAR Earth Scan (DALES) and the ISPRS 3D Semantic Labelling Benchmark datasets. In the experiments, the kNN method exhibited approximately 1% higher accuracy in weighted mean F1 and intersection over union (IoU) metrics compared to the spherical and cylindrical neighbourhood types on the DALES dataset. However, in the generalization experiment conducted on the ISPRS dataset, the spherical neighbourhood achieved the best results in these metrics, outperforming the cylindrical neighbourhood by a small margin. Notably, the kNN method was the least accurate, with a decrease in accuracy of approximately 1% in both weighted mean IoU and F1 scores. These findings suggest that the features extracted from spherical and cylindrical neighbourhood types are more generalizable compared to those from the kNN method.
AtıfScopus1WoS1OpenAlex1İncele - MakaleYalnız künye2023Scopus SJRQ2Simultaneous design of morphing hexarotor and autopilot system by using deep neural network and SPSA
Kose, Oguz; Oktay, Tugrul
PurposeThe purpose of this paper is to optimize the simultaneous flight performance of a hexarotor unmanned aerial vehicle (UAV) by using simultaneous perturbation stochastic approximation (i.e. SPSA), deep neural network and proportional integral derivative (i.e. PID) according to varying arm length (i.e. morphing). Design/methodology/approachIn this paper, proper PID gain coefficients and morphing ratio were obtained using the stochastic optimization method, also known as SPSA to maximize flight efficiency. Because it is difficult to establish an analytical connection between the morphing ratio and hexarotor moments of inertia, the deep neural network was used to obtain the moments of inertia according to the morphing ratio. By using SPSA and deep neural network, the best performance indexes were obtained and both longitudinal and lateral flight simulations were performed with the obtained data. FindingsWith SPSA, the best PID coefficients and morphing ratio are obtained for both longitudinal and lateral flight. Because the hexarotor solid body model changes according to the morphing ratio, the moment of inertia values used in the simulations also change. According to the morphing ratio, the moment of inertia values was obtained with the deep neural network over a created data set. Research limitations/implicationsIt takes a long time to obtain the morphing ratio suitable for the hexarotor model and the PID gain coefficients suitable for this morphing ratio. However, this situation can be overcome with the proposed SPSA. In addition, it takes a long time to obtain the appropriate moments of inertia according to the morphing ratio. However, in this case, it was overcome using the deep neural network. Practical implicationsDetermining the morphing ratio and PID gain coefficients using the optimization method, as well as determining the moments of inertia using the deep neural network, is very useful as it can increase the efficiency of hexarotor flight and flight efficiently with different arm lengths. With the proposed method, the hexarotor design performance criteria (i.e. rise time, settling time and overshoot) values were significantly improved compared to similar studies. Social implicationsDetermining the hexarotor flight parameters using SPSA and deep neural network provides advantages in terms of time, cost and applicability. Originality/valueThe hexarotor flight efficiency is improved with the proposed SPSA and deep neural network approaches. In addition, the desired flight parameters can be obtained more quickly and reliably with the proposed approaches. The design performance criteria were also improved, enabling the hexarotor UAV to follow the given trajectory in the best way and providing convenience for end users. SPSA was preferred because it converged faster than other methods. While other methods perform 2n operations per iteration, SPSA only performs two operations. To obtain the moment of inertia, many physical parameter values of the UAV are required in the existing methods. In the proposed method, by creating a date set, only arm length and moment of inertia were estimated without the need to obtain physical parameters with the deep neural network structure.
AtıfWoS48OpenAlex47İncele