Advancing precision agriculture: The potential of deep learning for cereal plant head detection

Computers and Electronics in Agriculture, Journal Year: 2023, Volume and Issue: 209, P. 107875 - 107875

Published: May 2, 2023

Language: Английский

---

TasselLFANet: a novel lightweight multi-branch feature aggregation neural network for high-throughput image-based maize tassels detection and counting

Frontiers in Plant Science, Journal Year: 2023, Volume and Issue: 14

Published: April 14, 2023

Accurately and rapidly counting the number of maize tassels is critical for breeding, management, monitoring growth stage plants. With advent high-throughput phenotyping platforms availability large-scale datasets, there a pressing need to automate this task genotype phenotype analysis. Computer vision technology has been increasingly applied in plant science, offering promising solution automated large However, current state-of-the-art image algorithms are hindered by hardware limitations, which compromise balance between algorithmic capacity, running speed, overall performance, making it difficult apply them real-time sensing field environments. Thus, we propose novel lightweight neural network, named TasselLFANet, with an efficient powerful structure accurately efficiently detecting high spatiotemporal sequences. Our proposed approach improves feature-learning ability TasselLFANet adopting cross-stage fusion strategy that balances variability different layers. Additionally, utilizes multiple receptive fields capture diverse feature representations, incorporates innovative visual channel attention module detect features more flexibly precisely. We conducted series comparative experiments on new, highly informative dataset called MrMT, demonstrate outperforms latest batch networks terms flexibility, adaptability, achieving F1 measure value 94.4%, mAP.@5 96.8%, having only 6.0M parameters. Moreover, compared regression-based TasselNetV3-Seg† model, our model achieves superior mean absolute error (MAE) 1.80, root square (RMSE) 2.68, R2 0.99. The meets accuracy speed requirements system tassel detection. Furthermore, method reliable unaffected geographical changes, providing essential technical support computerized field.

Language: Английский

---

High-throughput proximal ground crop phenotyping systems – A comprehensive review

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 224, P. 109108 - 109108

Published: June 8, 2024

Language: Английский

---

Health to Eat: A Smart Plate with Food Recognition, Classification, and Weight Measurement for Type-2 Diabetic Mellitus Patients’ Nutrition Control

Sensors, Journal Year: 2023, Volume and Issue: 23(3), P. 1656 - 1656

Published: Feb. 2, 2023

The management of type 2 diabetes mellitus (T2DM) is generally not only focused on pharmacological therapy. Medical nutrition therapy often forgotten by patients for several reasons, such as difficulty determining the right nutritional pattern themselves, regulating their daily patterns, or even heeding diet recommendations given doctors. Management one important efforts that can be made diabetic to prevent an increase in complexity disease. Setting a with proper will help manage healthy diet. development Smart Plate Health Eat technological innovation helps and users know food, weight, nutrients contained certain foods. This study involved 50 types food total 30,800 foods using YOLOv5s algorithm, where identification, measurement were investigated Chenbo load cell weight sensor (1 kg), HX711 weighing A/D module pressure sensor, IMX219-160 camera (waveshare). results this showed good identification accuracy analysis four food: rice (58%), braised quail eggs soy sauce (60%), spicy beef soup (62%), dried radish (31%), (100%).

Language: Английский

---

WheatLFANet: in-field detection and counting of wheat heads with high-real-time global regression network

Plant Methods, Journal Year: 2023, Volume and Issue: 19(1)

Published: Oct. 4, 2023

Detection and counting of wheat heads are crucial importance in the field plant science, as they can be used for crop management, yield prediction, phenotype analysis. With widespread application computer vision technology monitoring automated high-throughput phenotyping platforms has become possible. Currently, many innovative methods new technologies have been proposed that made significant progress accuracy robustness head recognition. Nevertheless, these often built on high-performance computing devices lack practicality. In resource-limited situations, may not effectively applied deployed, thereby failing to meet needs practical applications.In our recent research maize tassels, we TasselLFANet, most advanced neural network detecting tassels. Building this work, now developed a high-real-time lightweight called WheatLFANet detection. features more compact encoder-decoder structure an effective multi-dimensional information mapping fusion strategy, allowing it run efficiently low-end while maintaining high According evaluation report global detection dataset, outperforms other state-of-the-art with average precision AP 0.900 R2 value 0.949 between predicted values ground truth values. Moreover, runs significantly faster than all by order magnitude (TasselLFANet: FPS: 61).Extensive experiments shown exhibits better generalization ability methods, achieved speed increase accuracy. The success study demonstrates feasibility achieving real-time, devices, also indicates usefulness simple yet powerful designs.

Language: Английский

---

Advancements in Utilizing Image-Analysis Technology for Crop-Yield Estimation

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(6), P. 1003 - 1003

Published: March 12, 2024

Yield calculation is an important link in modern precision agriculture that effective means to improve breeding efficiency and adjust planting marketing plans. With the continuous progress of artificial intelligence sensing technology, yield-calculation schemes based on image-processing technology have many advantages such as high accuracy, low cost, non-destructive calculation, they been favored by a large number researchers. This article reviews research crop-yield remote images visible light images, describes technical characteristics applicable objects different schemes, focuses detailed explanations data acquisition, independent variable screening, algorithm selection, optimization. Common issues are also discussed summarized. Finally, solutions proposed for main problems arisen so far, future directions predicted, with aim achieving more wider popularization image technology.

Language: Английский

---

Improved YOLO-FastestV2 wheat spike detection model based on a multi-stage attention mechanism with a LightFPN detection head

Frontiers in Plant Science, Journal Year: 2024, Volume and Issue: 15

Published: June 19, 2024

The number of wheat spikes has an important influence on yield, and the rapid accurate detection spike numbers is great significance for yield estimation food security. Computer vision machine learning have been widely studied as potential alternatives to human detection. However, models with high accuracy are computationally intensive time consuming, lightweight tend lower precision. To address these concerns, YOLO-FastestV2 was selected base model comprehensive study analysis sheaf In this study, we constructed a target dataset comprising 11,451 images 496,974 bounding boxes. based Global Wheat Detection Dataset Sheaf Dataset, which published by PP Flying Paddle. We three attention mechanisms, Large Separable Kernel Attention (LSKA), Efficient Channel (ECA), Multi-Scale (EMA), enhance feature extraction capability backbone network improve underlying model. First, mechanism added after output phases network. Second, that further improved construct two-phase mechanism. On other hand, SimLightFPN introducing SimConv LightFPN module. results showed YOLO-FastestV2-SimLightFPN-ECA-EMA hybrid model, incorporates ECA in stage introduces EMA combination modules stage, best overall performance. P=83.91%, R=78.35%, AP= 81.52%, F1 = 81.03%, it ranked first GPI (0.84) evaluation. research examines deployment ear counting devices constrained resources, delivering novel solutions evolution agricultural automation precision agriculture.

Language: Английский

---

A fast and lightweight detection model for wheat fusarium head blight spikes in natural environments

Computers and Electronics in Agriculture, Journal Year: 2023, Volume and Issue: 216, P. 108484 - 108484

Published: Dec. 2, 2023

Language: Английский

---

CSNet: A Count-Supervised Network via Multiscale MLP-Mixer for Wheat Ear Counting

Plant Phenomics, Journal Year: 2024, Volume and Issue: 6

Published: Jan. 1, 2024

Wheat is the most widely grown crop in world, and its yield closely related to global food security. The number of ears important for wheat breeding estimation. Therefore, automated ear counting techniques are essential high-yield varieties increasing grain yield. However, all existing methods require position-level annotation training, implying that a large amount labor required annotation, limiting application development deep learning technology agricultural field. To address this problem, we propose count-supervised multiscale perceptive network (CSNet, network), which aims achieve accurate using quantity information. In particular, absence location information, CSNet adopts MLP-Mixer construct perception module with receptive field implements small target attention maps between features. We conduct comparative experiments on publicly available head detection dataset, showing proposed strategy outperforms position-supervised terms mean absolute error (MAE) root square (RMSE). This superior performance indicates approach has positive impact improving counts reducing labeling costs, demonstrating great potential tasks. code at http://csnet.samlab.cn.

Language: Английский

---

Application of improved YOLOv7-based sugarcane stem node recognition algorithm in complex environments

Frontiers in Plant Science, Journal Year: 2023, Volume and Issue: 14

Published: Aug. 23, 2023

Introduction Sugarcane stem node detection is one of the key functions a small intelligent sugarcane harvesting robot, but accuracy severely degraded in complex field environments when shadow confusing backgrounds and other objects. Methods To address problem low arise environments, this paper proposes an improved model based on YOLOv7. First, SimAM (A Simple Parameter-Free Attention Module for Convolutional Neural Networks) attention mechanism added to solve feature loss due image global context information convolution process, which improves case blurring; Second, Deformable Network used replace some traditional layers original Finally, new bounding box regression function WIoU Loss introduced unbalanced sample quality, improve robustness generalization ability, accelerate convergence speed network. Results The experimental results show that mAP algorithm 94.53% F1 value 92.41, are 3.43% 2.21 respectively compared with YOLOv7 model, SOTA method 94.1%, improvement 0.43% achieved, effectively performance target model. Discussion This study provides theoretical basis technical support development may also provide reference types crops similar environments.

Language: Английский

---