Xiaojian Zhou

Xiaojian Zhou

Auto-generated profile, based on publicly available information

Xiaojian Zhou

College of Information Engineering, Hainan Vocational University of Science and Technology, Haikou, Hainan, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

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

Representative Publications

View All Publications
13 March 2025 in IEEE Access
Verified
Journal Article
Open access
SZShulong Zhuo
Shulong Zhuo
HBHao Bai
Hao Bai
LJLifeng Jiang
Lifeng Jiang
XZXiaojian Zhou
Xiaojian Zhou
XDXu Duan
Xu Duan
ZZZihan Zhou
Zihan Zhou
Object detection in low-light environments has been widely recognized as a critical research direction in the field of computer vision. In response to the challenges of reduced detection accuracy and high edge-deployment costs encountered by mainstream single-stage object detection models under low-light conditions, this paper proposes a lightweight object detection network based on YOLOv11, integrates StarNet, C3k2-Star, and a lightweight detail-enhanced convolution and shared convolutional detection head(LSDECD), so called SCL-YOLOv11 herein. First, the StarNet architecture is introduced into the Backbone to enhance the extraction of shallow image features and significantly reduce computational complexity. Next, Star Blocks from the StarNet framework are employed to optimize the C3k2 module in the Neck stage, thereby improving the localization accuracy of deep features without increasing network complexity. Meanwhile, the Minimum Point Distance based IoU(MPDIoU) loss function is adopted to mitigate gradient explosion risks while enhancing detection precision. Furthermore, a lightweight detail-enhancement convolution layer and a shared-convolution detection head are designed to improve the model’s capability in capturing fine-grained details. Finally, a knowledge distillation procedure is conducted using SCL-YOLOv11 as the student model and YOLOv8 as the teacher model, ensuring that the proposed SCL-YOLOv11 model achieves both low complexity and excellent detection performance. Experimental results on the ExDark dataset demonstrate that the proposed algorithm achieves 67.6% mAP@0.5 and 42.4% mAP@0.5:0.95, while reducing parameter count by 38.5% and computational cost by 25.4%. Consequently, the proposed model is highly suitable for edge deployment in industrial applications and embedded development systems.

2 January 2025 in IEEE Access
Verified
Journal Article
Open access
XZXiaojian Zhou
Xiaojian Zhou
JKJicheng Kan
Jicheng Kan
NUR FATIN LIYANA MOHD ROSELYNUR FATIN LIYANA MOHD ROSELY
NUR FATIN LIYANA MOHD ROSELY
XDXu Duan
Xu Duan
JCJiajing Cai
Jiajing Cai
ZZZihan Zhou
Zihan Zhou
With the rapid growth of China’s new energy vehicle industry, the quality of crimped wire connectors directly impacts the performance of wiring harnesses, which are critical to the overall vehicle quality. At present, reliable methods for inspecting crimped wire connectors are still primarily based on image recognition evaluations. To address this, we propose an Improved Lightweight Network based on YOLOv8 (ILN-YOLOv8) to achieve high-precision and high-efficiency detection of crimped wire connectors. Taking the original YOLOv8 model as a baseline, the new model enhances the ability to extract shallow features from small targets by increasing the P2 detection layer and improving the Feature Pyramid Network(FPN) and Path Aggregation Network(PAN) structures. Next, the improved Selective Boundary Aggregation(SBA) module replaces the Concat module in the Neck, enhancing the fusion of deep and shallow features. Additionally, the Efficient Local Attention(ELA) attention mechanism is introduced into the Cryptographic Service Provider(CSP) bottleneck with 2 convolutions(C2F) module in the Backbone, improving feature localization accuracy without increasing network complexity. The Minimum Point Distance based IoU(MPDIoU) loss function is used to calculate localization loss, improving detection accuracy while preventing gradient explosion. Finally, lightweighting of the ILN-YOLOv8 model is achieved using the slim-neck network, the backbone with Depthwise Separable Convolution (DWConv), and Lightweight Convolution (LightConv) modules. After pruning and knowledge distillation, the model’s complexity and computational load significantly decreased while accuracy improved, meeting the industry’s requirements for crimped wire connectors detection and achieves superior performance. Experimental results show that, compared to the original YOLOv8 model, the proposed method achieved 96.2% accuracy on a real-world crimped wire connectors dataset, with mAP@0.5 and mAP@.5:.95 improving by 6.1% and 7.4%, respectively, while Parameters and FLOPs decreased by 66.7% and 34.6%.

7 August 2024 in IEEE Access
Verified
Journal Article
Open access
SZShulong Zhuo
Shulong Zhuo
JSJinmei Shi
Jinmei Shi
XZXiaojian Zhou
Xiaojian Zhou
JKJicheng Kan
Jicheng Kan
The Printed Circuit Board (PCB), often regarded as the cornerstone of the electronic information industry, plays a pivotal role in determining the quality of electronic products. However, traditional detection methods struggle to identify minuscule defects on PCBs due to their small surface area and intricate layouts, leading to a decline in product quality. To address these challenges, we propose an innovative lightweight intelligent detection algorithm, named YOLOv7-tiny Improved Detection (YOLOv7-TID). For precise defect detection on PCBs. This new model, based on YOLOv7-tiny, incorporates several enhancements. First, a parallel network module is added to the neck network to bolster the backbone network’s ability to extract feature information from both deep and shallow layers of images. Second, the sampling scale for small targets is increased, and the FPN and PAN structures are improved to enhance the feature extraction network’s semantic feature extraction and localization capabilities. Additionally, the SimAM attention mechanism module should be introduced to improve the network’s focus on shallow features without increasing the number of parameters. The model is further optimized by using the slim-neck network and the DWConv convolution module to reduce its weight, and by employing the NWD loss function to calculate positioning loss and enhance the network’s detection capability for small targets. Finally, the lightweight model undergoes pruning and knowledge distillation. Experimental results show that, compared with the original YOLOv7-tiny, the new model achieves a detection accuracy of 96.4% on the PCB defect test dataset. The mAP@0.5 and mAP@0.5:.95 are increased by 6.2% and 6.0%, respectively. Additionally, the number of parameters is reduced to only 3.5M, and the computation load is decreased by 23.5%. This makes the model more suitable for industrial applications and embedded development systems.

5 January 2026 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
SZShulong Zhuo
Shulong Zhuo
HBHao Bai
Hao Bai
XLXiongwei Li
Xiongwei Li
XZXiaojian Zhou
Xiaojian Zhou
JCJiajing Cai
Jiajing Cai
XDXu Duan
Xu Duan
Unmanned aerial vehicles (UAVs) offer a safe and economical alternative to manual patrols of overhead transmission lines, yet reliable localisation of missing cotter pins remains difficult because the targets are pixel–scale and often masked by cluttered towers. This work presents SADL–YOLOv11, a lightweight single-stage detector embedded in an end-to-end UAV workflow. A visibility-aware waypoint optimiser first plans 3-D trajectories that maximise pin exposure while satisfying safety envelopes reconstructed from LiDAR point clouds. The captured images are analysed on board by the proposed network, which extends a YOLOv11 baseline with a slim-neck (GSConv, VoVGSCSP, DySample), a micro-scale P2 head, an Adaptive Fine-Grained Channel Attention module, a Normalized-Wasserstein Distance loss for sub-pixel box regression, and a detail-enhanced weight-shared head. Group-Taylor pruning and feature–logit distillation further compress the model for edge deployment. On the Pin_Defect_Dataset, SADL–YOLOv11 attains 99.5% mAP@0.5 and 87.8% mAP@0.5:0.95. Under adverse imaging on the Pin_Low_Quality_Dataset (low light, blur/over-exposure, and very small targets), it achieves 97.0% mAP@0.5 and 73.3% mAP@0.5:0.95. Cross-domain transfer to PKU-Market-PCB confirms generalisation: mAP@0.5 matches YOLOv11 (78.5%) and mAP@0.5:0.95 improves by +3.2 points. The detector sustains 66.8 FPS on an RTX 3090 and 18–28 FPS on an RK3568 edge SoC, delivering a practical, high-precision, IoT-ready solution for large-scale power-grid inspection. Source code of proposed SADL-YOLOv11 model and related data-sets are available at the following link: https://github.com/haobai816/ultralytics-yolo11-Pin-Pro_SADL-YOLOv11.git.

20 October 2023
Verified
Conference Paper
Open access
MZMei Zhong
Mei Zhong
JSJinmei Shi
Jinmei Shi
XZXiaojian Zhou
Xiaojian Zhou
In this paper, for the AGV navigation path planning algorithm design problems, explored and researched the idea and implementation process of the A* algorithm, constructed the experimental environment model of AGV navigation path planning by using grid graphs, and carried out experiments with the help of simulation tools to verify the impact of different neighborhood node expansion search methods on the path planning speed in the A* algorithm. From the experimental data comparison and analysis results, it is concluded that the node expansion search method plays a key role in the efficiency of path planning. The results of the exploration research will provide a certain reference basis for the correct selection of node neighborhood expansion search method for AGV navigation route path planning in the actual production environment.

18 October 2024
Verified
Conference Paper
Open access
XZXiaojian Zhou
Xiaojian Zhou
JZJingnan Zou
Jingnan Zou
ZZZhen Zhang
Zhen Zhang
Wildlife is a valuable natural resource and an important component of biodiversity and natural ecosystems. This design focuses on wildlife video data and proposes an automated deployment tool for the HBase database environment, quickly establishing a data foundation for training models. The design primarily includes using Shell scripts to automate the deployment of Hadoop and HBase clusters, simplifying the deployment process and improving efficiency. It also involves using Python to serialize wildlife videos and store them in the created or selected HBase tables. After performing data cleansing and serialization on the video data, it is stored in the corresponding table. This design introduces automated deployment and storage technology in the context of wildlife conservation, enhancing the efficiency of environment setup, reducing manual intervention errors, and improving reliability.