Sun Jin

Dr. Sun Jin


Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
8 September 2022 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ZLZhiheng Li
Zhiheng Li
TGTongcheng Gu
Tongcheng Gu
SJSun Jin
Sun Jin
WXWubin Xu
Wubin Xu
HXHe Xin
He Xin
XHXiangyu Hui
Xiangyu Hui
Thus far, few studies have been conducted on fine-grained classification tasks for the latest convolutional neural network ConvNeXt, and no effective optimization method has been made available. To achieve more accurate fine-grained classification, this paper proposes two attention embedding methods based on ConvNeXt network and designs a new bilinear CBAM; simultaneously, a multiscale, multi-perspective and all-around attention framework is proposed, which is then applied in ConvNeXt. Experimental verification shows that the accuracy rate of the improved ConvNeXt for fine-grained image classification reaches 87.8%, 91.2%, and 93.2% on fine-grained classification datasets CUB-200-2011, Stanford Cars, and FGVC Aircraft, respectively, showing increases of 2.7%, 0.3% and 0.4%, respectively, compared to those of the original network without optimization, and increases of 3.7%, 8.0% and 2.0%, respectively, compared to those of the traditional BCNN. In addition, ablation experiments are set up to verify the effectiveness of the proposed attention framework.

25 September 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
YBYe Bowen
Ye Bowen
SJSun Jin
Sun Jin
SJSun Jin
Sun Jin
SYShuaiyu Yan
Shuaiyu Yan
DZDeng Zhang
Deng Zhang
For the visual measurement of moving arm holes in complex working conditions, a histogram equalization algorithm can be used to improve image contrast. To lessen the problems of image brightness shift, image over-enhancement, and gray-level merging that occur with the traditional histogram equalization algorithm, a dual histogram equalization algorithm based on adaptive image correction (AICHE) is proposed. To prevent luminance shifts from occurring during image equalization, the AICHE algorithm protects the average luminance of the input image by improving upon the Otsu algorithm, enabling it to split the histogram. Then, the AICHE algorithm uses the local grayscale correction algorithm to correct the grayscale to prevent the image over-enhancement and gray-level merging problems that arise with the traditional algorithm. It is experimentally verified that the AICHE algorithm can significantly improve the histogram segmentation effect and enhance the contrast and detail information while protecting the average brightness of the input image, and thus the image quality is significantly increased.

7 August 2023 in Processes
Verified
Journal Article
Open access
PDF file
CWChuanmei Wen
Chuanmei Wen
ZZZhimin Zhu
Zhimin Zhu
XFXuezhong Fu
Xuezhong Fu
TLTianliang Long
Tianliang Long
SJSun Jin
Sun Jin
Bolted joints are widely used in aeroengine rotor systems to connect multiple components into an integrated structure and provide sufficient stiffness. The mechanical properties of a bolted joint have a significant effect on rotor dynamics. For modern aeroengine designs, the blade-tip clearance is gradually reduced to improve efficiency, which may lead to rubbing damage and affect safe operation. The mechanical properties of a bolted joint change significantly during the blade–casing rubbing process and influence the dynamic properties of the rotor system. Based on the finite element (FE) modeling method, a 15-node bolted joint rotor system model is established in this paper, in which the bolted joint is represented by a 2-node joint element, and the blade–casing rubbing force is considered. The Newmark method is used to solve the motion equations. The dynamic model is validated by comparing the frequency response characteristics for different numbers of blades with the results provided in other published studies. Based on the established model, the effects of the rotational speed, number of blades, and rubbing stiffness on the dynamic responses, normal rubbing forces, and bending stiffness of the bolted joint are evaluated by numerical simulation. The results show that the response amplitude and bending stiffness of the bolted joint change significantly under blade–casing rubbing faults, and the mean value of the vibration response deviates significantly from 0 as the number of blades increases. Meanwhile, the amplitude of the frequency component fVC and the maximum value of the normal rubbing force also increase as the number of blades increases. The main contribution of this paper is the establishment of a new model for a bolted joint rotor system, considering the time-varying bending stiffness of the bolted joint and the blade–casing rub fault, comparing the simulation results to obtain some general results bridging the current research gap. Meanwhile, the numerical results in this paper can provide a cognitive basis for the blade–casing rubbing fault mechanism of a bolted joint rotor system under the influence of speed, number of blades, and rubbing stiffness. The nonlinear dynamic characteristics observed in the present paper can be applied to the blade–casing rubbing fault diagnosis of turbomachines.

16 December 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
BXBo Xu
Bo Xu
WXWubin Xu
Wubin Xu
SJSun Jin
Sun Jin
HZHanwen Zhang
Hanwen Zhang
YXYuanbin Xiao
Yuanbin Xiao
WZWeixin Zhou
Weixin Zhou
The road environment of open-pit mines is complex and unstructured. Unmanned construction machinery driving faces huge challenges. Improving the accuracy and speed of obstacle detection during driving is of great significance to ensuring the safety of mine automation construction and improving the efficiency of overall unmanned operations. In view of the fact that the current obstacle detection algorithm struggles to strike a balance between high precision and real-time performance, and there are problems such as difficulty in model deployment or unsuitability for practical applications, a lightweight open-pit mine driving obstacle detection algorithm based on improved YOLOv8s is proposed, which is committed to improving the driving safety of unmanned engineering machinery in open-pit mines. In order to enhance the ability of the backbone to capture features, the idea of the guidance module (CGBlock) of contextual information is introduced to construct a new CGC2f module; the efficient squeeze excitation (ESE) attention mechanism is embedded in the feature fusion layer to make the model pay more attention to the channels containing important feature information; in order to enhance the model’s learning ability for obstacles of different sizes in the open-pit mine, a more suitable dynamic head network (DyHead) is used at the output end; in order to further improve real-time performance, the layer-based adaptive amplitude pruning (LAMP) score algorithm is used to prune redundant weight parameters. To verify the effectiveness of the algorithm in this paper, an experimental verification is carried out on the constructed open-pit mine driving obstacle dataset. The results show that compared with YOLOv8s, the mAP50 of this algorithm reaches 95.3%, the detection speed is increased by 40.2%, the model parameters are reduced by 71.2%, and the calculation amount is reduced by 73.7%. It meets the requirements of real-time and high-precision obstacle detection in open-pit mine driving and provides technical support for smart mine driving.

19 October 2021 in IEEE Photonics Journal
Verified
Journal Article
Open access
JLJuliu Li
Juliu Li
FXFan Xiangsuo
Fan Xiangsuo
HCHuajin Chen
Huajin Chen
SJSun Jin
Sun Jin
MLMin Lei
Min Lei
XZXu Zhiyong
Xu Zhiyong
Small target detection in high strength clutter background is in great in remote imaging system, a new improved spatio-temporal filtering was proposed in this paper. Firstly, traditional anisotropy filtering has poor suppression effect in strength edge contour region, so a new diffusion filtering function proposed in paper. According to the analysis with difference of each component of the image, a new anisotropy diffusion function is constructed in this paper. When the difference of background and target is small, this algorithm will give in large diffusion coefficient to filter most background clutter and retain target signal well which achieves background prediction better. Secondly, because the traditional spatiotemporal filter algorithm cant follow the motion object in the fixed search pipe diameter what will make lose the target detection, a new weight constraint function of adaptive change of the search diameter in this paper is built which can change the search diameter with the moving of target, and improve the detection accuracy. Finally, experiments show that compared with traditional algorithms and detected in different scenes, this method can enhance small target detection effectively.

25 March 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
YXYuanbin Xiao
Yuanbin Xiao
SJSun Jin
Sun Jin
WXWubin Xu
Wubin Xu
WZWeixin Zhou
Weixin Zhou
BXBo Xu
Bo Xu
HZHanwen Zhang
Hanwen Zhang
This study introduces an enhanced ORB-SLAM3 algorithm to address the limitations of traditional visual SLAM systems in feature extraction and localization accuracy within the challenging terrains of open-pit mining environments. It also tackles the issue of sparse point cloud maps for mobile robot navigation. By combining point-line features with a Micro-Electro-Mechanical System (MEMS) Inertial Measurement Unit (IMU), the algorithm improves the feature matching’s reliability, particularly in low-texture areas. The method integrates dense point cloud mapping and an octree structure, optimizing both navigation and path planning while reducing storage demands and improving query efficiency. The experimental results using the TUM dataset and conducting tests in a simulated open-pit mining environment show that the proposed algorithm reduces the absolute trajectory error by 44.33% and the relative trajectory error by 14.34% compared to the ORB-SLAM3. The algorithm generates high-precision dense point cloud maps and uses an octree structure for efficient 3D spatial representation. In simulated open-pit mining scenarios, the dense mapping outperforms at reconstructing complex terrains, especially in low-texture gravel and uneven surfaces. These results highlight the robustness and practical applicability of the algorithm in dynamic and challenging environments, such as open-pit mining.