Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: point cloud classification

Sortuj według:

Ogranicz wyniki do:

Classification of objects in a point cloud using neural networks

Daszuta Marcin, Napieralska-Juszczak Ewa

Journal of Applied Computer Science

2019

Vol. 27, nr 2

7--16

3-dimensional scans captured in shape of point clouds are widely used in many different areas. Every such area use different kinds of sensors to ac-quire point clouds and do the analysis of the data but each of those needs some preanalysis to be done. One of the most important is segmentation and classification of points into types of objects. Such information considerably widens possibilities of usage for further purposes. There are many classifiers and many features based on which labeling can be done. In this paper few most commonly used approaches were chosen to check the influence of neighboring points acquisition on classification process. Results proof signif-icant relation between those two steps of point cloud analysis. Visualization of analyzed point cloud also shown that precision of predictions not always comes with better visibility of certain types of objects. Additionally, color-less analysis of geometrical features seems to be promising way for further research.

Development of LiDAR Data Classification Algorithms based on Parallel Computing using nVidia CUDA Technology

Bratuś R., Musialik P., Prochaska M., Rzonca A.

Measurement Automation Monitoring

2016

Vol. 62, No. 11

387--393

The paper presents an innovative data classification approach based on parallel computing performed on a GPGPU (General-Purpose Graphics Processing Unit). The results shown in this paper were obtained in the course of a European Commission-funded project: “Research on large-scale storage, sharing and processing of spatial laser data”, which concentrated on LIDAR data storage and sharing via databases and the application of parallel computing using nVidia CUDA technology. The paper describes the general requirements of nVidia CUDA technology application in massive LiDAR data processing. The studied point cloud data structure fulfills these requirements in most potential cases. A unique organization of the processing procedure is necessary. An innovative approach based on rapid parallel computing and analysis of each point’s normal vector to examine point cloud geometry within a classification process is described in this paper. The presented algorithm called LiMON classifies points into basic classes defined in LAS format: ground, buildings, vegetation, low points. The specific stages of the classification process are presented. The efficiency and correctness of LiMON were compared with popular program called Terrascan. The correctness of the results was tested in quantitive and qualitative ways. The test of quality was executed on specific objects, that are usually difficult for classification algorithms. The quantitive test used various environment types: forest, agricultural area, village, town. Reference clouds were obtained via two different methods: (1) automatic classification using Terrascan, (2) manually corrected clouds classified by Terrascan. The following coefficients for quantitive testing of classification correctness were calculated: Type 1 Error, Type 2 Error, Kappa, Total Error. The results shown in the paper present the use of parallel computing on a GPGPU as an attractive route for point cloud data processing.