Yuhang Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yuhang Zhang has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Youdong Jia
Faculty of Mechanical and Electrical Engineering, Kunming University, Kunming 650214, China
3 shared publications - Zhengfang Li
School of Mechanical and Electrical Engineering, Kunming University, Kunming 650214, China
3 shared publications - Zhengyang Qiao
Faculty of Mechanical and Electrical Engineering, Kunming University, Kunming 650214, China
2 shared publications - Haiwen Wu
Institute of Ecological Conservation and Restoration, Chinese Academy of Forestry, Beijing 100091, China
1 shared publication - Huihua Wang
School of Mechanical and Electronic Engineering, Kunming University, Kunming 650214, China
1 shared publication Jiankun Yang
Kunming Univ, Sch Mech & Elect Engn, Kunming 650214, Peoples R China
1 shared publication
Latest Publications
- Open Access
- 2 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-14
Hybrid-YOLOv5 for object detection of non-ferrous metals in end-of-life vehicles
In recent years, the detection of non-ferrous metals in end-of-life vehicles (ELVs) has become essential for improving recycling in the circular economy. Traditional methods struggle with accurate detection due to the variety of metals and challenging industrial environments. This study proposes a Hybrid-YOLOv5-based algorithm for efficiently detecting copper, aluminum, and stainless steel in ELVs. The goal is to enhance detection accuracy and computational efficiency in metal sorting. By integrating the Coarse-to-Fine (C2F) module, Squeeze-and-Excitation (SE) module, and MobileNetV3 backbone, we significantly improve performance and speed. On a dataset of 2,500 infrared images, Hybrid-YOLOv5 achieves 84.2% mAP@0.5 and 60 FPS inference speed, outperforming YOLOv3, YOLOv5, YOLOv7, and YOLOv11 by 22.2%, 12.4%, 11.1%, and 36.2% in mAP@0.5, respectively. This work provides an efficient solution for industrial metal sorting and intelligent recycling in the circular economy.
- 2 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-14
Open AccessHybrid-YOLOv5 for object detection of non-ferrous metals in end-of-life vehicles
In recent years, the detection of non-ferrous metals in end-of-life vehicles (ELVs) has become essential for improving recycling in the circular economy. Traditional methods struggle with accurate detection due to the variety of metals and challenging industrial environments. This study proposes a Hybrid-YOLOv5-based algorithm for efficiently detecting copper, aluminum, and stainless steel in ELVs. The goal is to enhance detection accuracy and computational efficiency in metal sorting. By integrating the Coarse-to-Fine (C2F) module, Squeeze-and-Excitation (SE) module, and MobileNetV3 backbone, we significantly improve performance and speed. On a dataset of 2,500 infrared images, Hybrid-YOLOv5 achieves 84.2% mAP@0.5 and 60 FPS inference speed, outperforming YOLOv3, YOLOv5, YOLOv7, and YOLOv11 by 22.2%, 12.4%, 11.1%, and 36.2% in mAP@0.5, respectively. This work provides an efficient solution for industrial metal sorting and intelligent recycling in the circular economy.
- Open Access
- 26 February 2025
- Published by MDPI AG in Agriculture
- Vol. 15 (5) , 511
An Efficient Method for Counting Large-Scale Plantings of Transplanted Crops in UAV Remote Sensing Images
Counting the number of transplanted crops is a crucial link in agricultural production, serving as a key method to promptly obtain information on crop growth conditions and ensure the yield and quality. The existing counting methods primarily rely on manual counting or estimation, which are inefficient, costly, and difficult to evaluate statistically. Additionally, some deep-learning-based algorithms can only crop large-scale remote sensing images obtained by Unmanned Aerial Vehicles (UAVs) into smaller sub-images for counting. However, this fragmentation often leads to incomplete crop contours of some transplanted crops, issues such as over-segmentation, repeated counting, low statistical efficiency, and also requires a significant amount of data annotation and model training work. To address the aforementioned challenges, this paper first proposes an effective framework for farmland segmentation, named MED-Net, based on DeepLabV3+, integrating MobileNetV2 and Efficient Channel Attention Net (ECA-Net), enabling precise plot segmentation. Secondly, color masking for transplanted crops is established in the HSV color space to further remove background information. After filtering and denoising, the contours of transplanted crops are extracted. An efficient contour filtering strategy is then applied to enable accurate counting. This paper conducted experiments on tobacco counting, and the experimental results demonstrated that the proposed MED-Net framework could accurately segment farmland in UAV large-scale remote sensing images with high similarity and complex backgrounds. The contour extraction and filtering strategy can effectively and accurately identify the contours of transplanted crops, meeting the requirements for rapid and accurate survival counting in the early stage of transplantation.
- 26 February 2025
- Published by MDPI AG in Agriculture
- Vol. 15 (5) , 511
Open AccessAn Efficient Method for Counting Large-Scale Plantings of Transplanted Crops in UAV Remote Sensing Images
Counting the number of transplanted crops is a crucial link in agricultural production, serving as a key method to promptly obtain information on crop growth conditions and ensure the yield and quality. The existing counting methods primarily rely on manual counting or estimation, which are inefficient, costly, and difficult to evaluate statistically. Additionally, some deep-learning-based algorithms can only crop large-scale remote sensing images obtained by Unmanned Aerial Vehicles (UAVs) into smaller sub-images for counting. However, this fragmentation often leads to incomplete crop contours of some transplanted crops, issues such as over-segmentation, repeated counting, low statistical efficiency, and also requires a significant amount of data annotation and model training work. To address the aforementioned challenges, this paper first proposes an effective framework for farmland segmentation, named MED-Net, based on DeepLabV3+, integrating MobileNetV2 and Efficient Channel Attention Net (ECA-Net), enabling precise plot segmentation. Secondly, color masking for transplanted crops is established in the HSV color space to further remove background information. After filtering and denoising, the contours of transplanted crops are extracted. An efficient contour filtering strategy is then applied to enable accurate counting. This paper conducted experiments on tobacco counting, and the experimental results demonstrated that the proposed MED-Net framework could accurately segment farmland in UAV large-scale remote sensing images with high similarity and complex backgrounds. The contour extraction and filtering strategy can effectively and accurately identify the contours of transplanted crops, meeting the requirements for rapid and accurate survival counting in the early stage of transplantation.
- Open Access
Research on Magnetic Field of Permanent Magnet Rolls Arranged Periodically in Circumferential—Axial Direction
The separation of non-magnetic non-ferrous metals such as copper and aluminum from scrapped automobiles is a critical area of research due to the increasing number of end-of-life vehicles. Traditional eddy current separation methods have limitations, particularly in handling large-sized broken copper and aluminum parts. This paper proposes a novel magnetic roller model featuring a circumferential–axial periodic arrangement of permanent magnets. This study explores the external magnetic field distribution of this new roller design by constructing an equivalent current model, solving magnetic scalar potential equations, and employing simulation tools. The findings indicate that the new magnet array enhances both the magnetic field strength and the range of the external magnetic field, leading to improved separation efficiency of large-sized metal fragments. The results provide a theoretical basis for advancing the separation technology of large-sized broken copper and aluminum parts in scrapped automobiles, offering potential improvements in the recycling of non-ferrous metals from end-of-life vehicles.
Open AccessResearch on Magnetic Field of Permanent Magnet Rolls Arranged Periodically in Circumferential—Axial Direction
The separation of non-magnetic non-ferrous metals such as copper and aluminum from scrapped automobiles is a critical area of research due to the increasing number of end-of-life vehicles. Traditional eddy current separation methods have limitations, particularly in handling large-sized broken copper and aluminum parts. This paper proposes a novel magnetic roller model featuring a circumferential–axial periodic arrangement of permanent magnets. This study explores the external magnetic field distribution of this new roller design by constructing an equivalent current model, solving magnetic scalar potential equations, and employing simulation tools. The findings indicate that the new magnet array enhances both the magnetic field strength and the range of the external magnetic field, leading to improved separation efficiency of large-sized metal fragments. The results provide a theoretical basis for advancing the separation technology of large-sized broken copper and aluminum parts in scrapped automobiles, offering potential improvements in the recycling of non-ferrous metals from end-of-life vehicles.
- Open Access
Effect Analysis of Process Parameters on Geometric Dimensions during Belt-Heated Incremental Sheet Forming of AA2024 Aluminum Alloy
The effect of process parameters on geometric dimensions is complex during belt-heated incremental sheet forming, leading to control difficulties in the geometric dimension of parts. In this study, the geometric errors were analyzed under the action of different process parameters, and the corresponding influence law was obtained through macro- and microexperiments. On this basis, the micromorphology of the deformation region section was analyzed in detail, and the effect of process parameters on micromorphology characteristics was obtained during belt-heated incremental sheet forming. Meanwhile, the dislocation density of the deformation region section was analyzed through XRD, and the change law of the dislocation density was revealed under the action of different process parameters. Therefore, the macro and micro effect mechanism of process parameters on geometric dimensions was verified, and then the control method of the geometric dimension was obtained in belt-heated incremental sheet forming.
Open AccessEffect Analysis of Process Parameters on Geometric Dimensions during Belt-Heated Incremental Sheet Forming of AA2024 Aluminum Alloy
The effect of process parameters on geometric dimensions is complex during belt-heated incremental sheet forming, leading to control difficulties in the geometric dimension of parts. In this study, the geometric errors were analyzed under the action of different process parameters, and the corresponding influence law was obtained through macro- and microexperiments. On this basis, the micromorphology of the deformation region section was analyzed in detail, and the effect of process parameters on micromorphology characteristics was obtained during belt-heated incremental sheet forming. Meanwhile, the dislocation density of the deformation region section was analyzed through XRD, and the change law of the dislocation density was revealed under the action of different process parameters. Therefore, the macro and micro effect mechanism of process parameters on geometric dimensions was verified, and then the control method of the geometric dimension was obtained in belt-heated incremental sheet forming.