Chunfeng Xu

Chunfeng Xu

Auto-generated profile, based on publicly available information

Chunfeng Xu

Beijing Institute of Technology (China)

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
1 January 2025 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
TZTian Zhan
Tian Zhan
CXChunfeng Xu
Chunfeng Xu
CZCheng Zhang
Cheng Zhang
KZKe Zhu
Ke Zhu
The Perspective-n-Point (PnP) problem has been widely studied in the literature and applied in various vision-based pose estimation scenarios. However, most existing methods ignore the anisotropy uncertainty of observations, as demonstrated in several real-world datasets in this letter. This oversight may lead to suboptimal and inaccurate estimation, particularly in the presence of noisy observations. To this end, we propose a generalized maximum likelihood PnP solver, named GMLPnP, that minimizes the determinant criterion by iterating the generalized least squares procedure to estimate the pose and uncertainty simultaneously. Further, the proposed method is decoupled from the camera model. Results of synthetic and real experiments show that our method achieves better accuracy in common pose estimation scenarios, GMLPnP improves rotation/translation accuracy by 4.7%/2.0% on TUM-RGBD and 18.6%/18.4% on KITTI-360 dataset compared to the best baseline. It is more accurate under very noisy observations in a vision-based UAV localization task, outperforming the best baseline by 34.4% in translation estimation accuracy.

10 November 2022
Verified
Conference Paper
Closed access
PDF file
JSJunqi Shi
Junqi Shi
LLLei Li
Lei Li
FLFuxiang Liu
Fuxiang Liu
CXChunfeng Xu
Chunfeng Xu
With the development of remote sensing technology, remote sensing images are developing towards higher resolution and larger data volume, and it is important to identify objects from remote sensing images quickly and accurately. Due to the significant differences between remote sensing images and natural images, the direct application of existing algorithms to remote sensing images is not ideal. Therefore, we take the YOLOv5 object detection algorithm as base, and have completed the following tasks: First, we present an adaptive image cutting data preprocessing method, which fills or cuts images into uniform size to cope with the resolution differences of remote sensing images. Second, we use the Mosaic data enhancement method to improve the algorithm's effect in complex backgrounds. Third, we use the Soft-NMS postprocessing algorithm to reduce the missed detection of dense objects. Furthermore, we transplant the algorithm to the hardware platform. After the above improvements, the mAP of our algorithm increases from 0.354 to 0.677 on the DOTA dataset, achieving a good object detection effect on remote sensing images; and with the help of the TensorRT, it has reached a detection speed of about 60 FPS on the NVIDIA Jetson AGX Xavier embedded hardware platform.

12 April 2022 in IEEE Photonics Journal
Verified
Journal Article
Open access
LZLinghua Zhang
Linghua Zhang
SHSen Han
Sen Han
XWXianyu Wu
Xianyu Wu
XQXianhao Qi
Xianhao Qi
JCJun Cheng
Jun Cheng
CXChunfeng Xu
Chunfeng Xu
Computer-generated hologram (CGH) is a diffractive optical element that has been commonly used as wavefront corrector in the interferometer and can be accurately manufactured to meet the requirements of high-precision detection. But the transmitted wavefront error (TWE) of the CGH substrate and the adjustment errors of cylindrical surface are influence factors of high accuracy measurement. In this paper, we propose the method of calibrating the CGH substrate TWE. After calibrating, the substrate TWE is 0.0052λ (RMS). Meanwhile, we obtain the general expression for the misalignment aberrations by describing the relationship between the wavefront of the measured cylindrical mirror and the corresponding adjustment errors. Then the adjustment errors of the test cylindrical surface can be completely separated by subtracting the fitting function from the measurement on the basis of least-square algorithm. After that, the PV value changes from 3.0565λ to 0.9825λ, and the RMS value changes from 0.5278λ to 0.1273λ, the measurement results have been greatly improved.

10 November 2022
Verified
Conference Paper
Closed access
PDF file
FLFuxiang Liu
Fuxiang Liu
CXChunfeng Xu
Chunfeng Xu
LLLei Li
Lei Li
JSJunqi Shi
Junqi Shi
The object tracking algorithms based on deep learning represented by SiamFC have demonstrated promising tracking capabilities. However, convolution networks take up a lot of memory, and it is difficult to run in real-time tracking on UAV platforms. Targeting this issue, we propose an object tracking algorithm called SiamUAV based on the siamese network in this paper. Firstly, based on the backbone network of the SiamFC algorithm, depthwise separable convolution is adopted to improve the tracking speed. Secondly, a spatial and channel squeeze & excitation block is introduced as an attention mechanism so that the backbone network can dynamically adjust to improve the tracking performance. Lastly, the algorithm is deployed on the NVIDIA Jetson AGX Xavier embedded platform with acceleration by TensorRT. The algorithm achieves essentially the same accuracy as the SiamFC algorithm. The tracking speed is improved by more than 70%, reaching 59 FPS on the embedded platform. This provides an excellent tracking speed while ensuring tracking accuracy.