Nondestructive high-throughput sugar beet fruit analysis using X-ray CT and deep learning

Looverbosch, Tim Van De; Vandenbussche, B.; Verboven, Pieter; Nicolaı̈, Bart

doi:10.1016/j.compag.2022.107228

Cited by 16 publications

(8 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our results are compared with the ones obtained in the closest literature. [ 40 ] employed X-ray CT imaging for physical quality evaluation, which may not be feasible for all research or production cases as it’s time-consuming and costly. In our method, not only 2D X-ray imaging was utilized to ensure rapid seed screening, but also our 2D-detection models can effectively provide a rapid physical quality assessment on Z-projections of 3D X-ray CT images.…”

Section: Discussionmentioning

confidence: 99%

“…In the most related state-of-the-art, images of single seeds are classified based on their viability/non-viability [ 37 ], their vigor [ 38 ] or the presence of insects [ 39 ]. By contrast, and like in [ 40 ], we propose a single pass end-to-end architecture based on joint object detection and classification while the two tasks were performed in two steps in previous works. Also, we especially focus on the data augmentation approach to deal with the class imbalance that was not considered in [ 40 ].…”

Section: Introductionmentioning

confidence: 99%

“…By contrast, and like in [ 40 ], we propose a single pass end-to-end architecture based on joint object detection and classification while the two tasks were performed in two steps in previous works. Also, we especially focus on the data augmentation approach to deal with the class imbalance that was not considered in [ 40 ]. In addition, other minor differences can also be underlined.…”

Section: Introductionmentioning

confidence: 99%

“…In addition, other minor differences can also be underlined. While [ 40 ] dealt with sugar beet only, we demonstrate our approach on two species without limitation of transferability to other species. Also, we consider a larger amount of defects in our approach.…”

Section: Introductionmentioning

confidence: 99%

“…Also, we consider a larger amount of defects in our approach. We deal, like in [ 37 – 39 , 41 ] with 2D images while [ 40 ] dealed with 3D images. Indeed, this is important as 2D X-ray imaging is faster and more cost-effective than 3D tomography.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Toward robust and high-throughput detection of seed defects in X-ray images via deep learning

Hamdy,

Charrier,

Corre

et al. 2024

Plant Methods

View full text Add to dashboard Cite

Background The detection of internal defects in seeds via non-destructive imaging techniques is a topic of high interest to optimize the quality of seed lots. In this context, X-ray imaging is especially suited. Recent studies have shown the feasibility of defect detection via deep learning models in 3D tomography images. We demonstrate the possibility of performing such deep learning-based analysis on 2D X-ray radiography for a faster yet robust method via the X-Robustifier pipeline proposed in this article. Results 2D X-ray images of both defective and defect-free seeds were acquired. A deep learning model based on state-of-the-art object detection neural networks is proposed. Specific data augmentation techniques are introduced to compensate for the low ratio of defects and increase the robustness to variation of the physical parameters of the X-ray imaging systems. The seed defects were accurately detected (F1-score >90%), surpassing human performance in computation time and error rates. The robustness of these models against the principal distortions commonly found in actual agro-industrial conditions is demonstrated, in particular, the robustness to physical noise, dimensionality reduction and the presence of seed coating. Conclusion This work provides a full pipeline to automatically detect common defects in seeds via 2D X-ray imaging. The method is illustrated on sugar beet and faba bean and could be efficiently extended to other species via the proposed generic X-ray data processing approach (X-Robustifier). Beyond a simple proof of feasibility, this constitutes important results toward the effective use in the routine of deep learning-based automatic detection of seed defects.

show abstract

Section: Discussionmentioning

confidence: 99%