Zhimin Mei
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zhimin Mei has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Shucai Wang
College of Engineering, Huazhong Agricultural University, Wuhan 430070, China
2 shared publications - Yifan Li
School of Intelligent Manufacturing, Wuchang Institute of Technology, Wuhan 430065, China
2 shared publications - Ben Ye
Hubei Engineering Research Center for Intelligent Detection and Identification of Complex Parts, Wuhan 430205, China
1 shared publication - Jiang Liu
School of Electrical and Information Engineering, Wuhan Institute of Technology, Wuhan 430205, China
1 shared publication - Kai Xiong
School of Intelligent Manufacturing, Wuchang Institute of Technology, Wuhan 430065, China
1 shared publication - Rongbo Zhu
Huazhong Agricultural University
1 shared publication
Latest Publications
- Open Access
- 7 October 2026
- Published by MDPI AG in Agriculture
- Vol. 16 (19) , 2160
Detection and Targeted Pesticide Application for Tea Diseases Based on YOLOv12 and 3D Vision
As a globally significant economic crop, tea yield and quality are strongly affected by disease control. Traditional manual inspection is inefficient and subjective, while non-targeted chemical application may increase pesticide use and environmental burden. To support lesion-level precision plant protection, this study proposes a tea disease detection and targeted pesticide application framework that combines an improved YOLOv12 detector with RGB-D three-dimensional (3D) perception and robotic execution. A self-built dataset containing 2121 source images from four tea disease categories was first partitioned at the original-image level into training, validation, and test subsets, and data augmentation was applied only to the training subset to prevent source-level leakage. The proposed YOLO-EMA-SoftNMS-DSConv (YOLO-EPS) model integrates Efficient Multi-Scale Attention (EMA), Distribution Shift Convolution (DSConv), and Soft Non-Maximum Suppression (Soft-NMS). Under the leakage-controlled protocol, YOLO-EPS achieved 94.28% mAP@0.5 and 87.16% mAP@0.5:0.95. Five independent runs with fixed source-level partitions yielded 94.24 ± 0.20% mAP@0.5, indicating stable training behavior. Detector-only inference time was 1.8 ms per image on an RTX 5090 desktop GPU using a Blackwell-compatible PyTorch/CUDA software stack. The system further maps two-dimensional lesion detections to RGB-D point clouds and robot coordinates and uses local surface geometry to support pose-aware targeted spraying. Controlled 3D localization and edge-platform latency experiments were conducted separately from the desktop detector benchmark. The proposed framework provides a basis for integrating disease recognition, spatial perception, and robotic targeted application in tea plantations.
- 7 October 2026
- Published by MDPI AG in Agriculture
- Vol. 16 (19) , 2160
Open AccessDetection and Targeted Pesticide Application for Tea Diseases Based on YOLOv12 and 3D Vision
As a globally significant economic crop, tea yield and quality are strongly affected by disease control. Traditional manual inspection is inefficient and subjective, while non-targeted chemical application may increase pesticide use and environmental burden. To support lesion-level precision plant protection, this study proposes a tea disease detection and targeted pesticide application framework that combines an improved YOLOv12 detector with RGB-D three-dimensional (3D) perception and robotic execution. A self-built dataset containing 2121 source images from four tea disease categories was first partitioned at the original-image level into training, validation, and test subsets, and data augmentation was applied only to the training subset to prevent source-level leakage. The proposed YOLO-EMA-SoftNMS-DSConv (YOLO-EPS) model integrates Efficient Multi-Scale Attention (EMA), Distribution Shift Convolution (DSConv), and Soft Non-Maximum Suppression (Soft-NMS). Under the leakage-controlled protocol, YOLO-EPS achieved 94.28% mAP@0.5 and 87.16% mAP@0.5:0.95. Five independent runs with fixed source-level partitions yielded 94.24 ± 0.20% mAP@0.5, indicating stable training behavior. Detector-only inference time was 1.8 ms per image on an RTX 5090 desktop GPU using a Blackwell-compatible PyTorch/CUDA software stack. The system further maps two-dimensional lesion detections to RGB-D point clouds and robot coordinates and uses local surface geometry to support pose-aware targeted spraying. Controlled 3D localization and edge-platform latency experiments were conducted separately from the desktop detector benchmark. The proposed framework provides a basis for integrating disease recognition, spatial perception, and robotic targeted application in tea plantations.
- Open Access
- 25 August 2025
- Published by MDPI AG in Electronics
- Vol. 14 (17) , 3372
Research on Medium Voltage Energy Storage Inverter Control Based on Hybrid Variable Virtual Vectors
Medium-voltage energy storage converter equipment is an important component of the new generation of ship power and power systems. Virtual space vector pulse width modulation, as a modulation optimization method to improve the neutral-point voltage imbalance in medium- and high-voltage multilevel energy storage converters, has become a research hotspot for T-type three-level energy storage inverter modulation methods due to its significant balancing effect and simple implementation. However, the current research method of constructing virtual vectors through redundant small vectors has limitations in regulating the neutral-point potential under full (especially high) modulation ratios. This paper proposes a modulation method that uses hybrid variable virtual small vectors and virtual medium vectors through optimization selection and reconstruction of basic vectors. This method ensures that the neutral-point charge change of the vector is zero and the common-mode voltage is minimized within the switching period under the full modulation ratio, achieving the purpose of controlling the neutral-point voltage balance and suppressing the common-mode voltage. Finally, simulation and experimental results show that the proposed method has good neutral-point voltage regulation and common-mode voltage suppression capabilities within the full modulation ratio range, and the system also has strong robustness and adaptability under different load conditions.
- 25 August 2025
- Published by MDPI AG in Electronics
- Vol. 14 (17) , 3372
Open AccessResearch on Medium Voltage Energy Storage Inverter Control Based on Hybrid Variable Virtual Vectors
Medium-voltage energy storage converter equipment is an important component of the new generation of ship power and power systems. Virtual space vector pulse width modulation, as a modulation optimization method to improve the neutral-point voltage imbalance in medium- and high-voltage multilevel energy storage converters, has become a research hotspot for T-type three-level energy storage inverter modulation methods due to its significant balancing effect and simple implementation. However, the current research method of constructing virtual vectors through redundant small vectors has limitations in regulating the neutral-point potential under full (especially high) modulation ratios. This paper proposes a modulation method that uses hybrid variable virtual small vectors and virtual medium vectors through optimization selection and reconstruction of basic vectors. This method ensures that the neutral-point charge change of the vector is zero and the common-mode voltage is minimized within the switching period under the full modulation ratio, achieving the purpose of controlling the neutral-point voltage balance and suppressing the common-mode voltage. Finally, simulation and experimental results show that the proposed method has good neutral-point voltage regulation and common-mode voltage suppression capabilities within the full modulation ratio range, and the system also has strong robustness and adaptability under different load conditions.
- Open Access
- 13 July 2025
- Published by MDPI AG in Agriculture
- Vol. 15 (14) , 1508
Intelligent Fruit Localization and Grasping Method Based on YOLO VX Model and 3D Vision
Recent years have seen significant interest among agricultural researchers in using robotics and machine vision to enhance intelligent orchard harvesting efficiency. This study proposes an improved hybrid framework integrating YOLO VX deep learning, 3D object recognition, and SLAM-based navigation for harvesting ripe fruits in greenhouse environments, achieving servo control of robotic arms with flexible end-effectors. The method comprises three key components: First, a fruit sample database containing varying maturity levels and morphological features is established, interfaced with an optimized YOLO VX model for target fruit identification. Second, a 3D camera acquires the target fruit’s spatial position and orientation data in real time, and these data are stored in the collaborative robot’s microcontroller. Finally, employing binocular calibration and triangulation, the SLAM navigation module guides the robotic arm to the designated picking location via unobstructed target positioning. Comprehensive comparative experiments between the improved YOLO v12n model and earlier versions were conducted to validate its performance. The results demonstrate that the optimized model surpasses traditional recognition and harvesting methods, offering superior target fruit identification response (minimum 30.9ms) and significantly higher accuracy (91.14%).
- 13 July 2025
- Published by MDPI AG in Agriculture
- Vol. 15 (14) , 1508
Open AccessIntelligent Fruit Localization and Grasping Method Based on YOLO VX Model and 3D Vision
Recent years have seen significant interest among agricultural researchers in using robotics and machine vision to enhance intelligent orchard harvesting efficiency. This study proposes an improved hybrid framework integrating YOLO VX deep learning, 3D object recognition, and SLAM-based navigation for harvesting ripe fruits in greenhouse environments, achieving servo control of robotic arms with flexible end-effectors. The method comprises three key components: First, a fruit sample database containing varying maturity levels and morphological features is established, interfaced with an optimized YOLO VX model for target fruit identification. Second, a 3D camera acquires the target fruit’s spatial position and orientation data in real time, and these data are stored in the collaborative robot’s microcontroller. Finally, employing binocular calibration and triangulation, the SLAM navigation module guides the robotic arm to the designated picking location via unobstructed target positioning. Comprehensive comparative experiments between the improved YOLO v12n model and earlier versions were conducted to validate its performance. The results demonstrate that the optimized model surpasses traditional recognition and harvesting methods, offering superior target fruit identification response (minimum 30.9ms) and significantly higher accuracy (91.14%).
- Open Access
- 7 November 2022
- Published by MDPI AG in Biomimetics
- Vol. 7 (4) , 194
Research on Logistics Distribution Center Location Based on Hybrid Beetle Antennae Search and Rain Algorithm
The location of logistics distribution centers is a crucial issue in modern logistics distribution systems. In order to obtain a more reasonable solution, an effective optimization algorithm is essential. This paper proposes a new hybrid method, named the beetle antennae search–rain algorithm (BRA), for the problem of logistics distribution centers’ location. The innovation of the BRA is embodied in three aspects. Firstly, the beetle antennae search (BAS) algorithm is embedded into the rain algorithm (RA); thus, the BAS is improved from an individual search to a swarm intelligent search and the global search ability is improved. Secondly, the search direction strategy of the BAS algorithm is incorporated into the RA, which can improve response speed while ensuring optimization performance. Finally, the search precision is improved by the mechanism of eliminating the inferior solution and generating a new solution. The BRA is tested on 10 benchmark functions and applied to solve the logistics distribution centers’ location problem. The performance of the BRA is compared to that of several classical heuristics by using relevant evaluation indexes and dynamic optimization convergence graphs. Experimental results show that the BRA outperforms the BAS algorithm, the RA and some other classic heuristics. It is also revealed that the BRA is an effective and competitive algorithm for logistics distribution centers’ location.
- 7 November 2022
- Published by MDPI AG in Biomimetics
- Vol. 7 (4) , 194
Open AccessResearch on Logistics Distribution Center Location Based on Hybrid Beetle Antennae Search and Rain Algorithm
The location of logistics distribution centers is a crucial issue in modern logistics distribution systems. In order to obtain a more reasonable solution, an effective optimization algorithm is essential. This paper proposes a new hybrid method, named the beetle antennae search–rain algorithm (BRA), for the problem of logistics distribution centers’ location. The innovation of the BRA is embodied in three aspects. Firstly, the beetle antennae search (BAS) algorithm is embedded into the rain algorithm (RA); thus, the BAS is improved from an individual search to a swarm intelligent search and the global search ability is improved. Secondly, the search direction strategy of the BAS algorithm is incorporated into the RA, which can improve response speed while ensuring optimization performance. Finally, the search precision is improved by the mechanism of eliminating the inferior solution and generating a new solution. The BRA is tested on 10 benchmark functions and applied to solve the logistics distribution centers’ location problem. The performance of the BRA is compared to that of several classical heuristics by using relevant evaluation indexes and dynamic optimization convergence graphs. Experimental results show that the BRA outperforms the BAS algorithm, the RA and some other classic heuristics. It is also revealed that the BRA is an effective and competitive algorithm for logistics distribution centers’ location.