Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Real-time hyperspectral stereo processing for the generation of 3D depth information
We present a local stereo matching method for hyperspectral camera data, allowing multiple usage of camera hardware and imaging data such as for object classification or spectral analysis and multichannel input to the correspondence problem. The matching process combines correlation-based similarity measures for pixel windows utilizing all 16 spectral channels followed by a consistency check for disparity selection. We evaluate stereo-processing methods focusing on effectiveness and runtime of the processing on a CPU and analyze parallelization possibilities. Based on the results of the evaluation on the CPU, we implement the optimized stereo matching for images with 16 channels on a graphics processing unit (GPU) utilizing the Compute Unified Device Architecture (CUDA). The parallel processing of the calculation steps to obtain the disparity image on the GPU achieves more than 27× speed up, resulting in calculation and post-processing of hyperspectral images with 8 – 13 Hz, depending on the selection of maximum disparity. The 3D reconstruction achieves a mean square error of 0.0267 m 2 in distance measurements from 5 – 10 m2.
Real-time hyperspectral stereo processing for the generation of 3D depth information
We present a local stereo matching method for hyperspectral camera data, allowing multiple usage of camera hardware and imaging data such as for object classification or spectral analysis and multichannel input to the correspondence problem. The matching process combines correlation-based similarity measures for pixel windows utilizing all 16 spectral channels followed by a consistency check for disparity selection. We evaluate stereo-processing methods focusing on effectiveness and runtime of the processing on a CPU and analyze parallelization possibilities. Based on the results of the evaluation on the CPU, we implement the optimized stereo matching for images with 16 channels on a graphics processing unit (GPU) utilizing the Compute Unified Device Architecture (CUDA). The parallel processing of the calculation steps to obtain the disparity image on the GPU achieves more than 27× speed up, resulting in calculation and post-processing of hyperspectral images with 8 – 13 Hz, depending on the selection of maximum disparity. The 3D reconstruction achieves a mean square error of 0.0267 m 2 in distance measurements from 5 – 10 m2.
Real-time hyperspectral stereo processing for the generation of 3D depth information
Heide, Nina (Autor:in) / Frese, Christian (Autor:in) / Emter, Thomas (Autor:in) / Petereit, Janko (Autor:in)
01.01.2018
Fraunhofer IOSB
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
DDC:
710
Development of a real-time stereo TEM
| British Library Online Contents | 2006
Information Processing - Spatiotriangulation With Multisensor HR Stereo-Images
| Online Contents | 2006
Information Processing - Methodological Proposal for Multispectral Stereo Matching
| Online Contents | 2006
Real-Time Simplex Growing Algorithms for Hyperspectral Endmember Extraction
| Online Contents | 2010