3D Reconstruction of Plant Leaves for High-Throughput Phenotyping

Zhu, Feiyu; Thapa, Suresh; Gao, Tiao; Ge, Yufeng; Walia, Harkamal; Yu, Hongfeng

doi:10.1109/bigdata.2018.8622428

Cited by 14 publications

(23 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, we have added the result of the leaf surface reconstruction step of Zhu et al [ 26 ] from their full process pipeline as the reference. The results from Zhu et al's method ( Figure 6 , column 7) showed the best fit in the model-free method throughout the data.…”

Section: Resultsmentioning

confidence: 99%

“…In addition, we have added the leaf area computation result of the leaf surface reconstruction step of Zhu et al [ 26 ] from their full process pipeline as the reference. The method from Zhu et al [ 26 ] tended to contain many overestimates for both the soybean and sugar beet leaves ( Figure 6 , column 7). Zhu et al's method uses the LOESS method to fit the surface of the point cloud.…”

Section: Discussionmentioning

confidence: 99%

“…We compared the leaf surface reconstruction results of the proposed method with Poisson surface reconstruction [ 32 ] and NURBS surface fitting [ 33 ]. Also, we added the comparison with the “Surface Reconstruction” step from their full process pipeline in Zhu et al [ 26 ] as the reference. We excluded the triangulation method that directly computes the triangular mesh from the data from the comparison because it utilized the noisy leaf point cloud directly to define the vertices of triangles and create the mesh surface.…”

Section: Methodsmentioning

confidence: 99%

“…For the low sensitivity setting, we changed the refinement iterations to 1. For Zhu et al's method, we used the same configuration as in the paper [ 26 ].…”

Section: Methodsmentioning

confidence: 99%

“…First, the model-free methods attempt to reconstruct a leaf surface from the 3D point cloud for a leaf in a bottom-up manner by relying on a leaf surface mesh representation, without requiring a prior mathematical model that describes the leaf. This category includes several methods, including the direct triangulation method that directly connects the points in the data to calculate the triangular mesh [ 15 – 17 ], Poisson surface reconstruction [ 18 – 21 ], nonuniform rational B-spline (NURBS) surface fitting [ 22 – 25 ], and locally weighted scatterplot smoothing (LOESS) [ 26 ]. These methods use local information from the point cloud to reconstruct the surface with high accuracy.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Robust Surface Reconstruction of Plant Leaves from 3D Point Clouds

2021

View full text Add to dashboard Cite

The automation of plant phenotyping using 3D imaging techniques is indispensable. However, conventional methods for reconstructing the leaf surface from 3D point clouds have a trade-off between the accuracy of leaf surface reconstruction and the method’s robustness against noise and missing points. To mitigate this trade-off, we developed a leaf surface reconstruction method that reduces the effects of noise and missing points while maintaining surface reconstruction accuracy by capturing two components of the leaf (the shape and distortion of that shape) separately using leaf-specific properties. This separation simplifies leaf surface reconstruction compared with conventional methods while increasing the robustness against noise and missing points. To evaluate the proposed method, we reconstructed the leaf surfaces from 3D point clouds of leaves acquired from two crop species (soybean and sugar beet) and compared the results with those of conventional methods. The result showed that the proposed method robustly reconstructed the leaf surfaces, despite the noise and missing points for two different leaf shapes. To evaluate the stability of the leaf surface reconstructions, we also calculated the leaf surface areas for 14 consecutive days of the target leaves. The result derived from the proposed method showed less variation of values and fewer outliers compared with the conventional methods.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

“…For the low sensitivity setting, we changed the refinement iterations to 1. For Zhu et al's method, we used the same configuration as in the paper [ 26 ].…”

Section: Methodsmentioning

confidence: 99%