Past, present and future of deep plant leaf disease recognition: A survey

Computers and Electronics in Agriculture, Год журнала: 2025, Номер 234, С. 110128 - 110128

Опубликована: Март 12, 2025

Язык: Английский

---

A high-precision jujube disease spot detection based on SSD during the sorting process

PLoS ONE, Год журнала: 2024, Номер 19(1), С. e0296314 - e0296314

Опубликована: Янв. 5, 2024

The development of automated grading equipment requires achieving high throughput and precise detection disease spots on jujubes. However, the current algorithms are inadequate in accomplishing these objectives due to their density, varying sizes shapes, limited location information regarding This paper proposes a method called JujubeSSD, boost precision identifying jujubes based single shot multi-box detector (SSD) network. In this study, diverse dataset comprising varied densities, multiple details was created through artificial collection data augmentation. parameter obtained from transfer learning into backbone feature extraction network SSD model, which reduced time spot 0.14 s. To enhance target detail features improve recognition weak information, traditional convolution layer replaced with deformable convolutional networks (DCNs). Furthermore, address challenge shapes regions jujubes, path aggregation pyramid (PAFPN) balanced (BFP) were integrated Experimental results demonstrate that mean average at IoU (intersection over union) threshold 0.5 ( [email protected] ) JujubeSSD reached 97.1%, representing an improvement approximately 6.35% compared original algorithm. When existing algorithms, such as YOLOv5 Faster R-CNN, improvements 16.84% 8.61%, respectively. Therefore, proposed for detecting jujube achieves superior performance surface meets requirements practical application agricultural production.

Язык: Английский

---

Transfer learning in agriculture: a review

Artificial Intelligence Review, Год журнала: 2025, Номер 58(4)

Опубликована: Янв. 25, 2025

Язык: Английский

---

Research on Insect Pest Identification in Rice Canopy Based on GA-Mask R-CNN

Agronomy, Год журнала: 2023, Номер 13(8), С. 2155 - 2155

Опубликована: Авг. 17, 2023

Aiming at difficult image acquisition and low recognition accuracy of two rice canopy pests, stem borer leaf roller, we constructed a GA-Mask R-CNN (Generative Adversarial Based Mask Region Convolutional Neural Network) intelligent model for combined it with field monitoring equipment them. Firstly, based on the biological habits variety pest collection methods were used to obtain images roller pests. different segmentation algorithms, segmented extract single samples. Secondly, bug generator generative adversarial network strategy improves sensitivity classification information, generates information in real environment, obtains sample dataset deep learning through multi-way augmentation. Then, adding channel attention ECA module improving connection residual blocks backbone ResNet101, is improved. Finally, was tested multi-source an average precision (AP) 92.71%, recall (R) 89.28% balanced score F1 90.96%. The precision, recall, are improved by 7.07, 7.65, 8.83%, respectively, compared original R-CNN. results show that performance indexes all better than R-CNN, Faster SSD, YOLOv5, other models, which can provide technical support remote

Язык: Английский

---

Can artificial intelligence understand our emotions? Deep learning applications with face recognition

Current Psychology, Год журнала: 2025, Номер unknown

Опубликована: Янв. 25, 2025

Язык: Английский

---

A Detection Method for Sweet Potato Leaf Spot Disease and Leaf-Eating Pests

Agriculture, Год журнала: 2025, Номер 15(5), С. 503 - 503

Опубликована: Фев. 26, 2025

Traditional sweet potato disease and pest detection methods have the limitations of low efficiency, poor accuracy manual dependence, while deep learning-based target can achieve an efficient accurate detection. This paper proposed leaf method SPLDPvB, as well a low-complexity version SPLDPvT, to identification spots pests, such hawk moth wheat moth. First, residual module containing three depthwise separable convolutional layers skip connection was effectively retain key feature information. Then, extraction integrating attention mechanism designed significantly improve capability. Finally, in model architecture, only structure backbone network decoupling head combination retained, traditional replaced by module, which greatly reduced complexity. The experimental results showed that mAP0.5 mAP0.5:0.95 SPLDPvB were 88.7% 74.6%, respectively, number parameters amount calculation 1.1 M 7.7 G, respectively. Compared with YOLOv11S, increased 2.3% 2.8%, 88.2% 63.8%, achieves higher complexity, demonstrating excellent performance detecting pests diseases. realizes automatic diseases provides technical guidance for spraying

Язык: Английский

---

From Convolutional Networks to Vision Transformers: Evolution of Deep Learning in Agricultural Pest and Disease Identification

Agronomy, Год журнала: 2025, Номер 15(5), С. 1079 - 1079

Опубликована: Апрель 29, 2025

Traditional pest and disease identification methods mainly rely on manual detection or traditional machine learning techniques, but they have obvious deficiencies in terms of their accuracy generalisation ability. In recent years, deep has gradually become the preferred solution for intelligent pests diseases by virtue its powerful automatic feature extraction complex data-processing capabilities. this paper, we systematically present application limitations, focus research progress methods, covering three mainstream architectures: convolutional neural network (CNN), Vision Transformer model CNN–Transformer hybrid model. addition, paper provides an in-depth analysis key challenges currently faced field recognition, including problems small-sample learning, background interference lightweighting, further propose solutions future to provide theoretical references technical guidance development related fields.

Язык: Английский

---

Aphid Recognition and Counting Based on an Improved YOLOv5 Algorithm in a Climate Chamber Environment

Insects, Год журнала: 2023, Номер 14(11), С. 839 - 839

Опубликована: Окт. 28, 2023

Due to changes in light intensity, varying degrees of aphid aggregation, and small scales the climate chamber environment, accurately identifying counting aphids remains a challenge. In this paper, an improved YOLOv5 detection model based on CNN is proposed address recognition counting. First, reduce overfitting problem insufficient data, uses image enhancement method combining Mosaic GridMask expand dataset. Second, convolutional block attention mechanism (CBAM) backbone layer improve accuracy targets. Subsequently, feature fusion bi-directional pyramid network (BiFPN) employed enhance neck, further improving speed aphids; addition, Transformer structure introduced front head investigate impact aggregation intensity accuracy. Experiments have shown that, through methods, recall rate can reach 99.1%, value [email protected] 99.3%, inference time 9.4 ms, which significantly better than other YOLO series networks. Moreover, it has strong robustness actual tasks provide reference for pest prevention control chambers.

Язык: Английский

---

MSGV-YOLOv7: A Lightweight Pineapple Detection Method

Agriculture, Год журнала: 2023, Номер 14(1), С. 29 - 29

Опубликована: Дек. 23, 2023

In order to optimize the efficiency of pineapple harvesting robots in recognition and target detection, this paper introduces a lightweight detection model, namely MSGV-YOLOv7. This model adopts MobileOne as innovative backbone network uses thin neck network. The enhancements these architectures have significantly improved ability feature extraction fusion, thereby speeding up rate. Empirical results indicated that MSGV-YOLOv7 surpassed original YOLOv7 with 1.98% increase precision, 1.35% recall rate, 3.03% mAP, while real-time speed reached 17.52 frames per second. Compared Faster R-CNN YOLOv5n, mAP increased by 14.89% 5.22%, respectively, approximately 2.18 times 1.58 times, respectively. application image visualization testing has verified results, confirming successfully precisely identified unique features pineapples. proposed method presents significant potential for broad-scale implementation. It is expected notably reduce both time economic costs associated operations.

Язык: Английский

---

A Deep Learning-based Pest Insect Monitoring System for Ultra-low Power Pocket-sized Drones

Опубликована: Апрель 29, 2024

Язык: Английский

---