3D Plant Phenotyping: All You Need is Labelled Point Cloud Data

Chaudhury, Ayan; Boudon, Frédéric; Godin, Christophe

doi:10.1007/978-3-030-65414-6_18

Cited by 11 publications

(12 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There exists many platforms for constructing virtual plant models such as L+C modelling language [40,41] and L-Py framework [42], which are based on the formalism of L-systems [24]. Despite the availability of such platforms capable of generating synthetic plants with complex architectures, employing them as 3D training data in the form of point clouds for plant phenotyping is only considered very recently [23].…”

Section: Related Workmentioning

confidence: 99%

“…The main factor that impedes the 3D deep learning techniques from being applied to plant phenotyping is the lack of large annotated 3D plant data sets [23]. Even moderate size annotated data sets of full plant models are not available.…”

Section: Arxiv:201211489v1 [Cscv] 21 Dec 2020mentioning

confidence: 99%

“…Incorporation of synthetic plants through generative models such as Lindenmayer systems (L-systems) [24,25] into training data are effective with 2D plant phenotyping [26]. The same strategy of creating synthetic 3D plant models can be followed to supply sufficient training data to machine learning frameworks [23].…”

Section: Arxiv:201211489v1 [Cscv] 21 Dec 2020mentioning

confidence: 99%

See 2 more Smart Citations

Segmentation of structural parts of rosebush plants with 3D point-based deep learning methods

et al. 2022

View full text Add to dashboard Cite

Background Segmentation of structural parts of 3D models of plants is an important step for plant phenotyping, especially for monitoring architectural and morphological traits. Current state-of-the art approaches rely on hand-crafted 3D local features for modeling geometric variations in plant structures. While recent advancements in deep learning on point clouds have the potential of extracting relevant local and global characteristics, the scarcity of labeled 3D plant data impedes the exploration of this potential. Results We adapted six recent point-based deep learning architectures (PointNet, PointNet++, DGCNN, PointCNN, ShellNet, RIConv) for segmentation of structural parts of rosebush models. We generated 3D synthetic rosebush models to provide adequate amount of labeled data for modification and pre-training of these architectures. To evaluate their performance on real rosebush plants, we used the ROSE-X data set of fully annotated point cloud models. We provided experiments with and without the incorporation of synthetic data to demonstrate the potential of point-based deep learning techniques even with limited labeled data of real plants. Conclusion The experimental results show that PointNet++ produces the highest segmentation accuracy among the six point-based deep learning methods. The advantage of PointNet++ is that it provides a flexibility in the scales of the hierarchical organization of the point cloud data. Pre-training with synthetic 3D models boosted the performance of all architectures, except for PointNet.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Arxiv:201211489v1 [Cscv] 21 Dec 2020mentioning

confidence: 99%

See 1 more Smart Citation

Segmentation of structural parts of rosebush plants with 3D point-based deep learning methods

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Fruit appearance is a key trait for many crops and can condi-2 tion market viability of fruit products and the success of cul-successfully implemented to measure the shape and size of fruits such as strawberries (4, 12), apples (5), carrot (6, 14), mangoes (28), and many others. More recently, methodologies for 3D reconstruction of plant organs have been developed with approaches that vary in speed, scale, cost, and accuracy; including laser scanners, x-ray computed tomography, and reconstruction from sequences of 2D images from digital cameras (8, [29][30][31][32][33][34][35][36][37][38][39][40][41]. Methods that rely on sequences of 2D images are numerous and variable with their own complexities and nuances that provide different strengths and weaknesses (8,27,37,(40)(41)(42)(43)(44).…”

Section: Introductionmentioning

confidence: 99%

“…Computer vision has shown great potential to quantify external fruit quality and 2D imaging has been successfully implemented to measure the shape and size of fruits such as strawberries (Ishikawa et al, 2018; Feldmann et al, 2020), apples (Migicovsky et al, 2016a), carrot (Horgan, 2001; Turner et al, 2018), mangoes (Naik et al, 2015), and many others. More recently, methodologies for 3D reconstruction of plant organs have been developed with approaches that vary in speed, scale, cost, and accuracy; including laser scanners, x-ray computed tomography, and reconstruction from sequences of 2D images from digital cameras (Gaillard et al, 2020; Chaudhury et al, 2020; Feldman et al, 2021; Dutagaci et al, 2020; Artzet et al, 2020; Teramoto et al, 2020; He et al, 2017; Paulus et al, 2014; Li et al, 2014; Liu et al, 2020; Dowd et al, 2021; Jiang et al, 2019; Sandhu et al, 2019; Hu et al, 2020). Methods that rely on sequences of 2D images are numerous and variable with their own complexities and nuances that provide different strengths and weaknesses (He et al, 2017; Porter et al, 2016; Sandhu et al, 2019; Hu et al, 2020; Wang et al, 2019; Liu et al, 2017, 2020; Warman et al, 2021).…”

Section: Introductionmentioning

confidence: 99%

Cost-effective, high-throughput phenotyping system for 3D reconstruction of fruit form

Feldmann

Tabb

2021

Preprint

View full text Add to dashboard Cite

Reliable phenotyping methods that are simple to operate and inexpensive to deploy are critical for studying quantitative traits in plants. Traditional fruit shape phenotyping relies on human raters or 2D analyses to assess form, e.g., size and shape. Systems for 3D imaging using multi-view stereo have been implemented, but frequently rely on commercial software and/or specialized hardware, which can lead to limitations in accessibility and scalability. We present a complete system constructed of consumer-grade components for capturing, calibrating, and reconstructing the 3D form of small-to-moderate sized fruits and tubers. Data acquisition and image capture sessions are 9 seconds to capture 60 images. The initial prototype cost was $1600 USD. We measured accuracy by comparing reconstructed models of 3D printed ground truth objects to the original digital files of those same ground truth objects. The R2 between length of the primary, secondary, and tertiary axes, volume, and surface area of the ground-truth object and the reconstructed models was > 0.97 and root-mean square error (RMSE) was <3mm for objects without locally concave regions. Measurements from 1mm and 2mm resolution reconstructions were consistent (R2 > 0.99). Qualitative assessments were performed on 48 fruit and tubers, including 18 strawberries, 12 potatoes, 5 grapes, 7 peppers, and 4 Bosch and 2 red Anjou pears. Our proposed phenotyping system is fast, relatively low cost, and has demonstrated accuracy for certain shape classes, and could be used for the 3D analysis of fruit form.

show abstract

Cost‐effective, high‐throughput phenotyping system for 3D reconstruction of fruit form

Feldmann

Tabb

2022

The Plant Phenome Journal

View full text Add to dashboard Cite

Reliable phenotyping methods that are simple to operate and inexpensive to deploy are critical for studying quantitative traits in plants. Traditional fruit shape phenotyping relies on human raters or 2D analyses to assess form, e.g., size and shape. Systems for 3D imaging using multi-view stereo have been implemented, but frequently rely on commercial software and/or specialized hardware, which can lead to limitations in accessibility and scalability. We present a complete system constructed of consumer-grade components for capturing, calibrating, and reconstructing the 3D form of small-to-moderate sized fruits and tubers. Data acquisition and image capture sessions are 9 seconds to capture 60 images. The initial prototype cost was $1600 USD. We measured accuracy by comparing reconstructed models of 3D printed ground truth objects to the original digital files of those same ground truth objects. The R 2 between length of the primary, secondary, and tertiary axes, volume, and surface area of the ground-truth object and the reconstructed models was > 0.97 and root-mean square error (RMSE) was <3mm for objects without locally concave regions. Measurements from 1mm and 2mm resolution reconstructions were consistent (R 2 > 0.99). Qualitative assessments were performed on 48 fruit and tubers, including 18 strawberries, 12 potatoes, 5 grapes, 7 peppers, and 4 Bosc and 2 red Anjou pears. Our proposed phenotyping system is fast, relatively low cost, and has demonstrated accuracy for certain shape classes, and could be used for the 3D analysis of fruit form.

show abstract

3D Plant Phenotyping: All You Need is Labelled Point Cloud Data

Cited by 11 publications

References 43 publications

Segmentation of structural parts of rosebush plants with 3D point-based deep learning methods

Segmentation of structural parts of rosebush plants with 3D point-based deep learning methods

Cost-effective, high-throughput phenotyping system for 3D reconstruction of fruit form

Cost‐effective, high‐throughput phenotyping system for 3D reconstruction of fruit form

Contact Info

Product

Resources

About