Çok Ölçekli Eğrilik Sınıflandırmasının CUDA ile Hızlandırılması

Year 2021, Issue: 28, 1022 - 1027, 30.11.2021

Mustafa Oğuz , Sercan Demirci

https://doi.org/10.31590/ejosat.1012445

Abstract

Günümüzde mekansal ölçüm teknolojileri gelişerek büyümektedir. Yeni teknolojiler ile birlikte çok daha hızlı ve doğru ölçümler yapmak mümkün hale gelmiştir. Hata oranlarının azalması, yanında verinin yoğunluğunun artması sorununu getirmiştir. Daha yoğun bir veri her ne kadar daha doğru sonuç verse de işlem süreçlerinde dezavantaja neden olmuştur. Verilerin işlenme sürelerinde büyük artışlar meydana gelmiştir. Bunlarla birlikte günümüzde paralel programlama üzerine çalışmalar yapılmaktadır. Paralel programlama işlemci üzerinde yapılabileceği gibi ekran kartı üzerinde de yapmak mümkündür. Ekran kartları üzerinde paralel programlama yapmak için kütüphaneler mevcuttur. Bunlardan en popüleri çok iyi bilinen Nvidia CUDA kütüphanesidir. CUDA kütüphanesi ile CUDA çekirdekleri üzerinde paralel programlama yapmak mümkün hale gelmiştir. Yapılan çalışmada yer sınıflandırma algoritması üzerinde hızlanma elde edilmesi hedeflenmiştir. Yer sınıflandırma algoritması olan MCC algoritması CUDA çekirdekleri üzerine dağıtılmış ve paralel hesaplanması sağlanmıştır. Çalışma sonunda 21 kat hızlanma elde edilmiştir.

Keywords

Filtreleme algoritmaları, Nokta bulutu, CUDA, Paralel programlama, İnce plaka spline, Çok ölçekli eğrilik sınıflandırması

Accelerating Multiscale Curvature Classification with CUDA

Abstract

Today, spatial measurement technologies are growing and developing. With new technologies, it has become possible to make much faster and more accurate measurements. Along with the decrease in error rates, the increase in the density of the data has brought the problem. Although a denser data gives more accurate results, it has caused a disadvantage in the processing steps. There has been a great increase in the processing times of the data. Along with these, studies on parallel programming are carried out today. Parallel programming can be done on the processor as well as on the graphics card. Many libraries are available on the internet for parallel programming on graphics cards. The most popular of these is the well-known Nvidia CUDA library. Parallel programming on CUDA cores has become possible with the CUDA library. In the study, it is aimed to achieve acceleration on the location classification algorithm. The MCC algorithm, which is a location classification algorithm, is distributed on CUDA cores and its parallel calculation is provided. At the end of the study, 21 times acceleration was obtained.

Keywords

Filtering Algorithms, Point cloud, CUDA, Parallel programming, Thin-plate spline, Multiscale Curvature Algorithm

References

• Chen, Q., Wang, H., Zhang, H., Sun, M., & Liu, X. (2016). A Point Cloud Filtering Approach to Generating DTMs for Steep Mountainous Areas and Adjacent Residential Areas. Remote Sensing, 8(1), 71. doi:10.3390/rs8010071
• Chen Z, Gao B, Devereux B. State-of-the-Art: DTM Generation Using Airborne LIDAR Data. Sensors. 2017; 17(1):150. https://doi.org/10.3390/s17010150
• Cheng, J., Grossman, M., & McKercher, T. (2014). Professional CUDA C Programming (1st ed.). Wrox.
• Cook, S. (2012). CUDA Programming: A Developer’s Guide to Parallel Computing with GPUs (Applications of Gpu Computing) (1st ed.). Morgan Kaufmann.
• Garland, M., le Grand, S., Nickolls, J., Anderson, J., Hardwick, J., Morton, S., Phillips, E., Zhang, Y., & Volkov, V. (2008). Parallel Computing Experiences with CUDA. IEEE Micro, 28(4), 13–27. https://doi.org/10.1109/mm.2008.57
• J. S. Evans and A. T. Hudak, "A Multiscale Curvature Algorithm for Classifying Discrete Return LiDAR in Forested Environments," in IEEE Transactions on Geoscience and Remote Sensing, vol. 45, no. 4, pp. 1029-1038, April 2007, doi: 10.1109/TGRS.2006.890412.
• Keqi Zhang, Shu-Ching Chen, Whitman, D., Mei-Ling Shyu, Jianhua Yan, & Chengcui Zhang. (2003). A progressive morphological filter for removing nonground measurements from airborne LIDAR data. IEEE Transactions on Geoscience and Remote Sensing, 41(4), 872–882. https://doi.org/10.1109/tgrs.2003.810682
• Meng, X., Lin, Y., Yan, L., Gao, X., Yao, Y., Wang, C., & Luo, S. (2019). Airborne LiDAR Point Cloud Filtering by a Multilevel Adaptive Filter Based on Morphological Reconstruction and Thin Plate Spline Interpolation. Electronics, 8(10), 1153. https://doi.org/10.3390/electronics8101153
• Mongus, D., & Zalik, B. (2014). Computationally Efficient Method for the Generation of a Digital Terrain Model From Airborne LiDAR Data Using Connected Operators. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 7, 340-351.
• Otepka, J., Ghuffar, S., Waldhauser, C., Hochreiter, R., & Pfeifer, N. (2013). Georeferenced Point Clouds: A Survey of Features and Point Cloud Management. ISPRS International Journal of Geo-Information, 2(4), 1038–1065. https://doi.org/10.3390/ijgi2041038
• Sibson, R., & Stone, G. (1991). Computation of Thin-Plate Splines. SIAM Journal on Scientific and Statistical Computing, 12(6), 1304–1313. https://doi.org/10.1137/0912070
• Sithole, G., & Vosselman, G. (2004). Experimental comparison of filter algorithms for bare-Earth extraction from airborne laser scanning point clouds. ISPRS Journal of Photogrammetry and Remote Sensing, 59(1–2),85–101. https://doi.org/10.1016/j.isprsjprs.2004.05.004
• Soyata, T. (2018). GPU Parallel Program Development Using CUDA (Chapman & Hall/CRC Computational Science) (1st ed.). Chapman and Hall/CRC.
• Vosselman, G., & Maas, H.-G. (2010). Airborne and terrestrial laser scanning. Dunbeath, Scotland: Whittles.