Identification of Unsound Grains in Wheat Using Deep Learning and Terahertz Spectral Imaging Technology

Jiang, Yuying; Wang, Fei; Ge, Hongyi; Li, Guangming; Chen, Xinyu; Li, Li; Lv, Ming; Zhang, Yuan

doi:10.3390/agronomy12051093

Cited by 7 publications

(3 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With further in-depth research, an increasing number of denoising algorithms based on convolutional neural networks have been proposed. Jiang et al [10] developed a THz spectral image-denoising model called CBDNet-V to address the issues of poor quality and unremarkable features in original THz spectral images of imperfect wheat grains, and the denoised images obtained by this model showed improved peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) compared to the denoising results of traditional models. Balaka et al [11] applied the unsupervised learning network CycleGAN to generate pairs of noisy synthetic images produced by a handwriting generator.…”

Section: Introductionmentioning

confidence: 99%

G-RRDB: An Effective THz Image-Denoising Model for Moldy Wheat

Jiang

Chen

et al. 2023

Foods

Self Cite

View full text Add to dashboard Cite

In order to solve the problem of large image noise and unremarkable features caused by factors such as fluctuations in the power of a light source during the terahertz image acquisition of wheat, this paper proposes a THz image-denoising model called G-RRDB. Firstly, a module called Ghost-LKA is proposed by combining a large kernel convolutional attention mechanism module with a Ghost convolutional structure, which improves the characteristics of the network to acquire a global sensory field. Secondly, by integrating a spatial attention mechanism with channel attention, an attention module called DAB is proposed to enhance the network’s attention to important features. Thirdly, the Ghost-LKA module and DAB module are combined with the baseline model, thus proposing the dense residual denoising network G-RRDB. Compared with traditional denoising networks, both the PSNR and SSIM are improved. The prediction accuracy of G-RRDB is verified through the classification of the VGG16 network, achieving a rate of 92.8%, which represents an improvement of 1.7% and 0.2% compared to the denoised images obtained from the baseline model and the combined baseline model with the DAB module, respectively. The experimental results demonstrate that G-RRDB, a THz image-denoising model based on dense residual structure for moldy wheat, exhibits excellent denoising performance.

show abstract

Section: Introductionmentioning

confidence: 99%

G-RRDB: An Effective THz Image-Denoising Model for Moldy Wheat

Jiang

Chen

et al. 2023

Foods

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recently, deep learning has also achieved impressive results in the field of THz imaging. Deep learning has been applied to segmentation and classification tasks in THz images [209,210] and to low-resolution problem-solving with superresolution techniques based on deep learning [211,212]. In fast THz imaging, deep learning can significantly increase the signalto-noise ratio.…”

mentioning

confidence: 99%

Review of Bioplastics Characterisation by Terahertz Techniques in the View of Ensuring a Circular Economy

Abina,

Korošec,

Puc

et al. 2023

Photonics

View full text Add to dashboard Cite

The increasing scarcity of natural resources, worsening global climate change, environmental degradation, and rising demand for food are forcing the biotechnology and plastics industries to seek and apply circular economy models that would lead to a sustainable transition in the production and use of bioplastics. Circular economy models can improve the economic productivity of bio-based plastics and have a positive impact on the environment by reducing conventional plastic waste and the consumption of petrochemical feedstocks for plastic production. In addition, some agricultural wastes that have the potential to be used as bioplastics can be reused. Terahertz (THz) systems are already used in the plastics and rubber industries for non-destructive testing, detection, imaging, and quality control. Several reports have highlighted the potential applications of THz spectroscopy and imaging in polymer analysis and plastics characterisation. This potential is even greater with chemometric methods and artificial intelligence algorithms. In this review, we focus on applications that support the transformation of the biotechnology sector to the circular economy, particularly via the transition from conventional plastics to bioplastics. In this review, we discuss the potential of THz systems for the characterisation and analysis of bioplastics and biopolymers. The results of previous studies on biopolymers in the THz frequency range are summarised. Furthermore, the potential of using artificial intelligence approaches such as machine learning as advanced analytical methods in THz spectroscopy and imaging, in addition to the conventionally used chemometric methods, is discussed. The results of this review highlight that THz technology can contribute to closed technological circles in important areas of biotechnology and the related plastics and rubber industries.

show abstract

“…Recently, deep learning has achieved impressive results in various fields, including THz imaging 30 . Deep learning has been applied to segmentation and classification tasks in THz images such as impurity detection in wheat 31,32 , breast cancer classification 33 , and heavy-metal detection in soils 34 . The low resolution problem of THz imaging can also be mitigated by deep learning based super-resolution techniques 35,36 .…”

mentioning

confidence: 99%

Deep learning for terahertz image denoising in nondestructive historical document analysis

Dutta

Root

Ullmann

et al. 2022

Sci Rep

View full text Add to dashboard Cite

Historical documents contain essential information about the past, including places, people, or events. Many of these valuable cultural artifacts cannot be further examined due to aging or external influences, as they are too fragile to be opened or turned over, so their rich contents remain hidden. Terahertz (THz) imaging is a nondestructive 3D imaging technique that can be used to reveal the hidden contents without damaging the documents. As noise or imaging artifacts are predominantly present in reconstructed images processed by standard THz reconstruction algorithms, this work intends to improve THz image quality with deep learning. To overcome the data scarcity problem in training a supervised deep learning model, an unsupervised deep learning network (CycleGAN) is first applied to generate paired noisy THz images from clean images (clean images are generated by a handwriting generator). With such synthetic noisy-to-clean paired images, a supervised deep learning model using Pix2pixGAN is trained, which is effective to enhance real noisy THz images. After Pix2pixGAN denoising, 99% characters written on one-side of the Xuan paper can be clearly recognized, while 61% characters written on one-side of the standard paper are sufficiently recognized. The average perceptual indices of Pix2pixGAN processed images are 16.83, which is very close to the average perceptual index 16.19 of clean handwriting images. Our work has important value for THz-imaging-based nondestructive historical document analysis.

show abstract

Identification of Unsound Grains in Wheat Using Deep Learning and Terahertz Spectral Imaging Technology

Cited by 7 publications

References 35 publications

G-RRDB: An Effective THz Image-Denoising Model for Moldy Wheat

G-RRDB: An Effective THz Image-Denoising Model for Moldy Wheat

Review of Bioplastics Characterisation by Terahertz Techniques in the View of Ensuring a Circular Economy

Deep learning for terahertz image denoising in nondestructive historical document analysis

Contact Info

Product

Resources

About