Tensor-based classification and segmentation of three-dimensional point clouds for organ-level plant phenotyping and growth analysis

Elnashef, Bashar; Filin, Sagi; Lati, Ran Nisim

doi:10.1016/j.compag.2018.10.036

Cited by 64 publications

(51 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Second, Phenomenal provides new algorithms, based on 3D skeleton analysis, to perform individual organ segmentation and morphological analysis of the plant. Because such segmentation is already available for dicotyledons based on planar / tubular segmentation techniques developed elsewhere (Paproki et al, 2012, Elnashelf et al, 2019), our algorithm also targeted monocotyledonous plants, where oblong leaves are directly attached to stems and are partly tubular. (Fig.…”

Section: Resultsmentioning

confidence: 99%

“…Most pipelines are limited to dicotyledonous plants (Paproki et al, 2012), for which leaf segmentation is simplified by easily identifiable petioles. Elnashef et al (2019) showed that tensor-based classification of 3D point cloud allowed stem detection on maize, wheat and cotton. Still, the leaf segmentation task remains complex in monocotyledonous species where leaves are directly connected to the stem, often rolled within each other and frequently crossing neighboring leaves (Das Choudhury et al, 2018, Elnashef et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

“…Currently, existing 3D analysis pipelines are either semi-automatic (Mccormick et al, 2017), automatic but only for a very limited dataset (Elnashef et al, 2019), or implement only a subpart of the phenotyping pipeline, like 3D geometric reconstruction or organ segmentation from a mesh (Paproki et al, 2012; Pound et al, 2014). As a consequence, to our knowledge, none of these 3D solutions has been tested on large datasets.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Phenomenal: An automatic open source library for 3D shoot architecture reconstruction and analysis for image-based plant phenotyping

Artzet

Chen

Chopard

et al. 2019

Preprint

View full text Add to dashboard Cite

In the era of high-throughput visual plant phenotyping, it is crucial to design fully automated and flexible workflows able to derive quantitative traits from plant images. Over the last years, several software supports the extraction of architectural features of shoot systems. Yet currently no end-to-end systems are able to extract both 3D shoot topology and geometry of plants automatically from images on large datasets and a large range of species. In particular, these software essentially deal with dicotyledons, whose architecture is comparatively easier to analyze than monocotyledons. To tackle these challenges, we designed the Phenomenal software featured with: (i) a completely automatic workflow system including data import, reconstruction of 3D plant architecture for a range of species and quantitative measurements on the reconstructed plants; (ii) an open source library for the development and comparison of new algorithms to perform 3D shoot reconstruction and (iii) an integration framework to couple workflow outputs with existing models towards model-assisted phenotyping. Phenomenal analyzes a large variety of data sets and species from images of high-throughput phenotyping platform experiments to published data obtained in different conditions and provided in a different format. Phenomenal has been validated both on manual measurements and synthetic data simulated by 3D models. It has been also tested on other published datasets to reproduce a published semi-automatic reconstruction workflow in an automatic way. Phenomenal is available as an open-source software on a public repository.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Phenomenal: An automatic open source library for 3D shoot architecture reconstruction and analysis for image-based plant phenotyping

Artzet

Chen

Chopard

et al. 2019

Preprint

View full text Add to dashboard Cite

show abstract

“…At stage T2, the overlapping between leaves was so serious that even the manual separation cannot perform well (Figure 7b). As far as we know, fully automatic segmentation has only been realized on broad-leaf plants with few leaves [14] and narrow-leaf crops with few leaves and without tillers [43].…”

Section: Processing Of Point Cloud Of Individual Plantmentioning

confidence: 99%

Image-Based Dynamic Quantification of Aboveground Structure of Sugar Beet in Field

et al. 2020

View full text Add to dashboard Cite

Sugar beet is one of the main crops for sugar production in the world. With the increasing demand for sugar, more desirable sugar beet genotypes need to be cultivated through plant breeding programs. Precise plant phenotyping in the field still remains challenge. In this study, structure from motion (SFM) approach was used to reconstruct a three-dimensional (3D) model for sugar beets from 20 genotypes at three growth stages in the field. An automatic data processing pipeline was developed to process point clouds of sugar beet including preprocessing, coordinates correction, filtering and segmentation of point cloud of individual plant. Phenotypic traits were also automatically extracted regarding plant height, maximum canopy area, convex hull volume, total leaf area and individual leaf length. Total leaf area and convex hull volume were adopted to explore the relationship with biomass. The results showed that high correlations between measured and estimated values with R2 > 0.8. Statistical analyses between biomass and extracted traits proved that both convex hull volume and total leaf area can predict biomass well. The proposed pipeline can estimate sugar beet traits precisely in the field and provide a basis for sugar beet breeding.

show abstract

“…Xu et al [21] designed a computer graphics-based algorithm to segment leaves from TLS point clouds of Ehretia macrophylla, Crape myrtle, and Fatsia japonica plants, and the precision reached 94%, 90.6%, and 88.8%, respectively. Elnashef et al [22] proposed a tensor-based classification method to segment leaves from cotton, corn, and wheat point clouds scanned from a multi-view imaging system, and the average precision reached 92%, 94%, and 95%, respectively. Hu et al [23] developed a 3D point cloud filtering method for leaves.…”

Section: Introductionmentioning

confidence: 99%

An Overlapping-Free Leaf Segmentation Method for Plant Point Clouds

Cao

Shi

et al. 2019

IEEE Access

View full text Add to dashboard Cite

Automatic leaf segmentation, as well as identification and classification methods that built upon it, are able to provide immediate monitoring for plant growth status to guarantee the output. Although 3D plant point clouds contain abundant phenotypic features, plant leaves are usually distributed in clusters and are sometimes seriously overlapped in the canopy. Therefore, it is still a big challenge to automatically segment each individual leaf from a highly crowded plant canopy in 3D for plant phenotyping purposes. In this work, we propose an overlapping-free individual leaf segmentation method for plant point clouds using the 3D filtering and facet region growing. In order to separate leaves with different overlapping situations, we develop a new 3D joint filtering operator, which integrates a Radius-based Outlier Filter (RBOF) and a Surface Boundary Filter (SBF) to help to separate occluded leaves. By introducing the facet over-segmentation and facet-based region growing, the noise in segmentation is suppressed and labeled leaf centers can expand to their whole leaves, respectively. Our method can work on point clouds generated from three types of 3D imaging platforms, and also suitable for different kinds of plant species. In experiments, it obtains a point-level cover rate of 97% for Epipremnum aureum, 99% for Monstera deliciosa, 99% for Calathea makoyana, and 87% for Hedera nepalensis sample plants. At the leaf level, our method reaches an average Recall at 100.00%, a Precision at 99.33%, and an average F-measure at 99.66%, respectively. The proposed method can also facilitate the automatic traits estimation of each single leaf (such as the leaf area, length, and width), which has potential to become a highly effective tool for plant research and agricultural engineering.

show abstract

Tensor-based classification and segmentation of three-dimensional point clouds for organ-level plant phenotyping and growth analysis

Cited by 64 publications

References 23 publications

Phenomenal: An automatic open source library for 3D shoot architecture reconstruction and analysis for image-based plant phenotyping

Phenomenal: An automatic open source library for 3D shoot architecture reconstruction and analysis for image-based plant phenotyping

Image-Based Dynamic Quantification of Aboveground Structure of Sugar Beet in Field

An Overlapping-Free Leaf Segmentation Method for Plant Point Clouds

Contact Info

Product

Resources

About