Looking behind occlusions: A study on amodal segmentation for robust on-tree apple fruit size estimation

Gené-Mola, Jordi; Ferrer-Ferrer, Mar; Gregorio, Eduard; Blok, Pieter M.; Hemming, J.; Morros, Josep Ramon; Rosell-Polo, Joan R.; Ruiz-Hidalgo, Javier

doi:10.1016/j.compag.2023.107854

Cited by 22 publications

(12 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1 ). This subset is suitable for training, validating and testing fruit detection and sizing methods, as demonstrated in previous studies [ 1 , 4 ]. The availability of modal and amodal masks also enables the estimation of apples visibility, which is of interest in fruit sizing and harvesting robotics [7] .…”

Section: Data Descriptionmentioning

confidence: 94%

“…The dataset is prepared to be applied to train and fine-tune existing instance segmentation neural networks such as Mask R-CNN [3] . Additionally, the dataset includes ground truth fruit sizes, enabling the implementation and reproduction of results obtained in the fruit sizing methods proposed in [1] and [4] . Rich and Diverse Dataset: With an extensive collection of RGB-D images featuring both modal and amodal segmentation masks, researchers can delve into advanced fruit detection and sizing techniques.…”

Section: Value Of the Datamentioning

confidence: 99%

Section: Value Of the Datamentioning

confidence: 99%

“…However, these methods encounter challenges in effectively handling partially occluded fruits. Recently, the authors proposed a solution to this challenge introducing the use of amodal instance segmentation, predicting both visible and occluded regions of fruits [1] . The present data article adds significant value to the published research paper by providing a comprehensive and publicly available resource for training and testing the proposed modal-amodal segmentation algorithm.…”

Section: Introductionmentioning

confidence: 99%

“…Computers and Electronics in Agriculture 209 (2023) 107,854. DOI: https://doi.org/10.1016/j.compag.2023.107854 [1] …”

mentioning

confidence: 99%

See 4 more Smart Citations

AmodalAppleSize_RGB-D dataset: RGB-D images of apple trees annotated with modal and amodal segmentation masks for fruit detection, visibility and size estimation

Gené-Mola,

Ferrer-Ferrer,

Hemming

et al. 2024

Data in Brief

Self Cite

View full text Add to dashboard Cite

Section: Data Descriptionmentioning

confidence: 94%

Section: Value Of the Datamentioning

confidence: 99%

Section: Value Of the Datamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…Computers and Electronics in Agriculture 209 (2023) 107,854. DOI: https://doi.org/10.1016/j.compag.2023.107854 [1] …”

mentioning

confidence: 99%

See 3 more Smart Citations

AmodalAppleSize_RGB-D dataset: RGB-D images of apple trees annotated with modal and amodal segmentation masks for fruit detection, visibility and size estimation

Gené-Mola,

Ferrer-Ferrer,

Hemming

et al. 2024

Data in Brief

Self Cite

View full text Add to dashboard Cite

Instance segmentation of faces and mouth-opening degrees based on improved YOLOv8 method

Fan,

Zhang,

Zheng

et al. 2024

Multimedia Systems

View full text Add to dashboard Cite

A computer vision system for apple fruit sizing by means of low-cost depth camera and neural network application

Bortolotti,

Piani,

Gullino

et al. 2024

Precision Agric

View full text Add to dashboard Cite

Fruit size is crucial for growers as it influences consumer willingness to buy and the price of the fruit. Fruit size and growth along the seasons are two parameters that can lead to more precise orchard management favoring production sustainability. In this study, a Python-based computer vision system (CVS) for sizing apples directly on the tree was developed to ease fruit sizing tasks. The system is made of a consumer-grade depth camera and was tested at two distances among 17 timings throughout the season, in a Fuji apple orchard. The CVS exploited a specifically trained YOLOv5 detection algorithm, a circle detection algorithm, and a trigonometric approach based on depth information to size the fruits. Comparisons with standard-trained YOLOv5 models and with spherical objects were carried out. The algorithm showed good fruit detection and circle detection performance, with a sizing rate of 92%. Good correlations (r > 0.8) between estimated and actual fruit size were found. The sizing performance showed an overall mean error (mE) and RMSE of + 5.7 mm (9%) and 10 mm (15%). The best results of mE were always found at 1.0 m, compared to 1.5 m. Key factors for the presented methodology were: the fruit detectors customization; the HoughCircle parameters adaptability to object size, camera distance, and color; and the issue of field natural illumination. The study also highlighted the uncertainty of human operators in the reference data collection (5–6%) and the effect of random subsampling on the statistical analysis of fruit size estimation. Despite the high error values, the CVS shows potential for fruit sizing at the orchard scale. Future research will focus on improving and testing the CVS on a large scale, as well as investigating other image analysis methods and the ability to estimate fruit growth.

show abstract

Looking behind occlusions: A study on amodal segmentation for robust on-tree apple fruit size estimation

Cited by 22 publications

References 33 publications

AmodalAppleSize_RGB-D dataset: RGB-D images of apple trees annotated with modal and amodal segmentation masks for fruit detection, visibility and size estimation

AmodalAppleSize_RGB-D dataset: RGB-D images of apple trees annotated with modal and amodal segmentation masks for fruit detection, visibility and size estimation

Instance segmentation of faces and mouth-opening degrees based on improved YOLOv8 method

A computer vision system for apple fruit sizing by means of low-cost depth camera and neural network application

Contact Info

Product

Resources

About