gu jiaze

Dr. gu jiaze

Guangzhou Institute of Technology, Xidian University, Guangzhou, China, Guangzhou, Guangdong, China
+1
orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

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

Representative Publications

View All Publications
28 June 2022 in Remote Sensing
Verified
Journal Article
Open access
PDF file
Wanrun LIWanrun LI
Wanrun LI
WZWenhai Zhao
Wenhai Zhao
gu jiazegu jiaze
gu jiaze
BFBoyuan Fan
Boyuan Fan
YDYongfeng Du
Yongfeng Du
The structural condition of blades is mainly evaluated using manual inspection methods. However, these methods are time-consuming, labor-intensive, and costly, and the detection results significantly depend on the experience of inspectors, often resulting in lower precision. Focusing on the dynamic characteristics (i.e., natural frequencies) of large wind turbine blades, this study proposes a monitoring method based on the target-free DSST (Discriminative Scale Space Tracker) vision algorithm and UAV. First, the displacement drift of UAV during hovering is studied. Accordingly, a displacement compensation method based on high-pass filtering is proposed herein, and the scale factor is adaptive. Then, the machine learning is employed to map the position and scale filters of the DSST algorithm to highlight the features of the target image. Subsequently, a target-free DSST vision algorithm is proposed, in which illumination changes and complex backgrounds are considered. Additionally, the algorithm is verified using traditional computer vision algorithms. Finally, the UAV and the target-free DSST vision algorithm are used to extract the dynamic characteristic of the wind turbine blades under shutdown. Results show that the proposed method can accurately identify the dynamic characteristics of the wind turbine blade. This study can serve as a reference for assessment of the condition of wind turbine blades.

16 October 2025 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
gu jiazegu jiaze
gu jiaze
LFLulin Fan
Lulin Fan
JZJing Zhao
Jing Zhao
XCXIANGHAI CAO
XIANGHAI CAO
Conventional remote sensing object detection models are predominantly constrained by their dependence on closed-set annotations, limiting their capability to accurately detect objects that were not present in the training dataset. To enable the model to generalize to novel object categories, we propose CoseDet, a novel open-vocabulary object detection framework that integrates the semantic richness of visual-language pretraining model with the precise positioning capability of modern detection architectures. Specifically, CoseDet augments a Faster R-CNN detector with a ResNet50-FPN backbone by incorporating RemoteCLIP-based embeddings through a pseudo-word mechanism, which aligns high-dimensional visual features with robust textual semantics. Furthermore, a convolutional block attention module is employed to refine feature representations, and explicit modeling of surrounding regions is utilized to capture crucial contextual dependencies. Comprehensive experiments based on four datasets demonstrate that CoseDet not only outperforms state-of-the-art methods but also provides a robust and generalizable solution for open-vocabulary object detection in complex remote sensing scenarios.