Lightweight convolutional neural network models for semantic segmentation of in-field cotton bolls

Singh, Naseeb; Tewari, V.K.; Biswas, Prabir Kumar; Dhruw, L.K.

doi:10.1016/j.aiia.2023.03.001

Cited by 10 publications

(1 citation statement)

References 80 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For cotton boll detection, Xu et al used color cotton images acquired by an unmanned aircraft system to design and train convolutional neural networks (CNNs) to detect cotton bolls in the original images and used dense point clouds constructed from aerial images and a motion structure approach to computing the 3D positions of the bolls [14]. Singh et al combined low-level and high-level features of cotton field images and used different filter sizes to construct a neural network model for real-time cotton boll detection [15]. The model can be applied to cotton harvesting robots to provide an effective method for real-time cotton boll identification in the field.…”

Section: Introductionmentioning

confidence: 99%

YOLO-C: An Efficient and Robust Detection Algorithm for Mature Long Staple Cotton Targets with High-Resolution RGB Images

et al. 2023

View full text Add to dashboard Cite

Under complex field conditions, robust and efficient boll detection at maturity is an important tool for pre-harvest strategy and yield prediction. To achieve automatic detection and counting of long-staple cotton in a natural environment, this paper proposes an improved algorithm incorporating deformable convolution and attention mechanism, called YOLO-C, based on YOLOv7: (1) To capture more detailed and localized features in the image, part of the 3 × 3 convolution in the ELAN layer of the backbone is replaced by deformable convolution to improve the expressiveness and accuracy of the model. (2) To suppress irrelevant information, three SENet modules are introduced after the backbone to improve the ability of feature maps to express information, and CBAM and CA are introduced for comparison experiments. (3) A WIoU loss function based on a dynamic non-monotonic focusing mechanism is established to reduce the harmful gradients generated by low-quality examples on the original loss function and improve the model performance. During the model evaluation, the model is compared with other YOLO series and mainstream detection algorithms, and the model mAP@0.5 achieves 97.19%, which is 1.6% better than the YOLOv7 algorithm. In the model testing session, the root mean square error and coefficient of determination (R2) of YOLO-C are 1.88 and 0.96, respectively, indicating that YOLO-C has higher robustness and reliability for boll target detection in complex environments and can provide an effective method for yield prediction of long-staple cotton at maturity.

show abstract

Section: Introductionmentioning

confidence: 99%

YOLO-C: An Efficient and Robust Detection Algorithm for Mature Long Staple Cotton Targets with High-Resolution RGB Images

et al. 2023

View full text Add to dashboard Cite

show abstract

Development of a Deep Learning‐Assisted Mobile Application for the Identification of Nematodes Through Microscopic Images

Singh,

Dhruw

et al. 2024

Modern Agriculture

View full text Add to dashboard Cite

Nematodes are microscopic metazoans, some species of which can be used as biological insecticides, while some other species annually damage 10.0%–20.0% of crops globally. Accurate identification of nematodes is crucial for their effective utilisation or control. Current methods of nematode identification are labour‐intensive, time‐consuming, and prone to false positives, thus necessitating the development of an intelligent system for their identification from microscopic images without technical assistance. In this study, a novel approach was investigated for the identification of nematodes from microscopic images. A novel lightweight convolutional neural network (CNN) was developed to identify the nematodes belonging to different trophic groups (Heterorhabditis indica, Meloidogyne incognita, Helicotylenchus, Anguina tritici, and Xiphinema). The CNN model was trained for 75 epochs using 70.0% of the nematode dataset, with validation on 20.0% of the dataset. To ensure unbiased evaluation, 30 images from each class were randomly selected from the remaining 10.0% of the dataset for testing the classification performance of the trained model. The trained models achieved an average classification accuracy, precision, recall, and F1‐score values of 98.52%, 95.66%, 95.56%, and 95.56%, respectively. The proposed CNN was found to be four times faster and five times lighter against a marginal (< 1.0%) decrease in accuracy when compared with the existing state‐of‐the‐art CNNs. The classification accuracy of the developed mobile application was validated by nematode specialists on freshly captured data and found to be greater than 98.0%. Hence, the developed mobile application can be effectively applied for the identification of targeted nematodes and eliminate the necessity of specialists.

show abstract

Deep Learning-Enabled Mobile Application for On-Site Nitrogen Prediction in Strawberry Cultivation

Singh,

Mahore,

Kaur

et al. 2024

J. Biosyst. Eng.

View full text Add to dashboard Cite

Lightweight convolutional neural network models for semantic segmentation of in-field cotton bolls

Cited by 10 publications

References 80 publications

YOLO-C: An Efficient and Robust Detection Algorithm for Mature Long Staple Cotton Targets with High-Resolution RGB Images

YOLO-C: An Efficient and Robust Detection Algorithm for Mature Long Staple Cotton Targets with High-Resolution RGB Images

Development of a Deep Learning‐Assisted Mobile Application for the Identification of Nematodes Through Microscopic Images

Deep Learning-Enabled Mobile Application for On-Site Nitrogen Prediction in Strawberry Cultivation

Contact Info

Product

Resources

About