Yanfei Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yanfei Zhang has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Guangyao Yuan
Dalian University of Technology
2 shared publications - Jinliang Gong
School of Agricultural Engineering and Food Science, Shandong University of Technology, Zibo/China
2 shared publications - Kazi Mostafa
School of Intelligent Manufacturing Ecosystem, Xi’an Jiaotong-Liverpool University, Suzhou 215123, China
2 shared publications - Jiaheng Zhang
School of Mechanical Engineering, Shandong University of Technology, Zibo 255049, China
1 shared publication - Kai Cui
Nanjing Dragonfly Smart Agri Research Corp. LTD, Nanjing 210019, China
1 shared publication
Latest Publications
- Open Access
Weed Identification in Maize Fields Based on Improved Swin-Unet
The maize field environment is complex. Weeds and maize have similar colors and may overlap, and lighting and weather conditions vary. Thus, many methods for the automated differentiation of maize and weeds achieve poor segmentation or cannot be used in real time. In this paper, a weed recognition model based on improved Swin-Unet is proposed. The model first performs semantic segmentation of maize seedlings and uses the resulting mask to identify weeds. U-Net acts as the semantic segmentation framework, and a Swin transformer module is introduced to improve performance. DropBlock regularization, which randomly hides some blocks in crop feature maps, is applied to enhance the generalization ability of the model. Finally, weed areas are identified and segmented with the aid of an improved morphological processing algorithm. The DeepLabv3+, PSANet, Mask R-CNN, original Swin-Unet, and proposed models are trained on a dataset of maize seedling images. The proposed Swin-Unet model outperforms the others, achieving a mean intersection over union of 92.75%, mean pixel accuracy of 95.57%, and inference speed of 15.1 FPS. Our model could be used for accurate, real-time segmentation of crops and weeds and as a reference for the development of intelligent agricultural equipment.
Open AccessWeed Identification in Maize Fields Based on Improved Swin-Unet
The maize field environment is complex. Weeds and maize have similar colors and may overlap, and lighting and weather conditions vary. Thus, many methods for the automated differentiation of maize and weeds achieve poor segmentation or cannot be used in real time. In this paper, a weed recognition model based on improved Swin-Unet is proposed. The model first performs semantic segmentation of maize seedlings and uses the resulting mask to identify weeds. U-Net acts as the semantic segmentation framework, and a Swin transformer module is introduced to improve performance. DropBlock regularization, which randomly hides some blocks in crop feature maps, is applied to enhance the generalization ability of the model. Finally, weed areas are identified and segmented with the aid of an improved morphological processing algorithm. The DeepLabv3+, PSANet, Mask R-CNN, original Swin-Unet, and proposed models are trained on a dataset of maize seedling images. The proposed Swin-Unet model outperforms the others, achieving a mean intersection over union of 92.75%, mean pixel accuracy of 95.57%, and inference speed of 15.1 FPS. Our model could be used for accurate, real-time segmentation of crops and weeds and as a reference for the development of intelligent agricultural equipment.
- Open Access
- 17 November 2022
- Published by MDPI AG in Micromachines
- Vol. 13 (11) , 1996
Performance Test and Parameter Optimization of Trichogramma Delivery System
Trichogramma is a small wasp that is a natural enemy of many agricultural pests. Although Trichogramma can be used in sustainable crop production, conventional methods of delivering Trichogramma to fields are expensive and may cause pollution. In this study, the feasibility of using unmanned aerial vehicles (UAVs) for Trichogramma delivery was investigated. A six-rotor plant protection UAV was equipped with a Trichogramma delivery device, and a Box-Behnken experimental design was carried out with the Trichogramma pills as the test material, the launch height, the launch speed and the launch interval as the experimental factors, and the Trichogramma pills’ landing accuracy as the test index. The data were analyzed by ANOVA using the Design-Expert software, and the influence of each experimental factor on the accuracy of the Trichogramma pills bolus landing was explored through response surface analysis. The regression model between the experimental factors and the experimental indicators was established, and the parameters were optimized based on the response surface method, and the optimal combination parameters were obtained. The ANOVA revealed that the launch height A had the greatest effect on the accuracy, followed by launch interval C and launch velocity B. The results demonstrated that the optimal parameter combination of the Trichogramma delivery system is the launch height of 147.95 cm, the launch speed of 3.7745 m/s, and the launch interval of 2.98 s. At this moment, the accuracy of Trichogramma pills’ bolus landing was the highest, with an accuracy of 93.29%. The average relative error between the experimental value and the optimization result was 1.71%, indicating that the Trichogramma delivery system could meet the requirements of delivery. This study provides theoretical references and technical support for verifying the feasibility of the Trichogramma delivery system.
- 17 November 2022
- Published by MDPI AG in Micromachines
- Vol. 13 (11) , 1996
Open AccessPerformance Test and Parameter Optimization of Trichogramma Delivery System
Trichogramma is a small wasp that is a natural enemy of many agricultural pests. Although Trichogramma can be used in sustainable crop production, conventional methods of delivering Trichogramma to fields are expensive and may cause pollution. In this study, the feasibility of using unmanned aerial vehicles (UAVs) for Trichogramma delivery was investigated. A six-rotor plant protection UAV was equipped with a Trichogramma delivery device, and a Box-Behnken experimental design was carried out with the Trichogramma pills as the test material, the launch height, the launch speed and the launch interval as the experimental factors, and the Trichogramma pills’ landing accuracy as the test index. The data were analyzed by ANOVA using the Design-Expert software, and the influence of each experimental factor on the accuracy of the Trichogramma pills bolus landing was explored through response surface analysis. The regression model between the experimental factors and the experimental indicators was established, and the parameters were optimized based on the response surface method, and the optimal combination parameters were obtained. The ANOVA revealed that the launch height A had the greatest effect on the accuracy, followed by launch interval C and launch velocity B. The results demonstrated that the optimal parameter combination of the Trichogramma delivery system is the launch height of 147.95 cm, the launch speed of 3.7745 m/s, and the launch interval of 2.98 s. At this moment, the accuracy of Trichogramma pills’ bolus landing was the highest, with an accuracy of 93.29%. The average relative error between the experimental value and the optimization result was 1.71%, indicating that the Trichogramma delivery system could meet the requirements of delivery. This study provides theoretical references and technical support for verifying the feasibility of the Trichogramma delivery system.