Claim · model-written summary
The proposed method can be adopted to build stereo matching datasets and the workflow can be used for depth perception in plant phenotyping.
Not reportedMaterialGeneCompoundOrgan / stageInteractorInteractor roleSecond interactorSecond interactor roleTreatmentComparatorSettingDirectionValueQualifier
Verbatim
e6“The proposed method can be adopted to build stereo matching datasets, and the workflow can be used for depth perception in plant phenotyping”
Title
PlantStereo: A High Quality Stereo Matching Dataset for Plant Reconstruction
Qingyu Wang; Dihua Wu; Wei Liu; et al. Agriculture 2023. doi:10.3390/agriculture13020330
Other claims from this paper (8)
- Method / resource The PlantStereo dataset outperforms other representative stereo matching datasets in data size, disparity accuracy, disparity density and data type.
- Intervention effect Sub-pixel level disparity accuracy improves matching accuracy compared with pixel level disparity accuracy.
- Method / resource For PSMNet, sub-pixel disparity accuracy decreased the EPE by 0.30 pixels.
- Method / resource For PSMNet, sub-pixel disparity accuracy decreased the bad−3 error by 2.13%.
- Method / resource For GwcNet, sub-pixel disparity accuracy decreased the EPE by 0.08 pixels.
- Method / resource For GwcNet, sub-pixel disparity accuracy decreased the bad−3 error by 0.42%.
- Method / resource The stereo matching workflow achieves competitive depth error (2.5 mm at 0.7 m) compared with ToF and structured light.
- Method / resource The stereo matching workflow achieves real-time performance of 50 fps at 1046 × 606 compared with ToF and structured light.