Arrow Research search

Author name cluster

Hanbin Luo

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

2 papers
1 author row

Possible papers

2

EAAI Journal 2026 Journal Article

Vector attention-based point cloud network for semantic segmentation of sewer sonar data

  • Wenli Liu
  • Yueming Jiang
  • Hanlin Li
  • Lei Yang
  • Hanbin Luo

Sonar technology is unaffected by lighting or water conditions, making it ideal for inspecting water-filled sewers. Nonetheless, significant challenges remain in utilizing sonar point clouds effectively. This research introduces the Vector Attention-based Point Cloud Network (VAPCNet), a deep learning method for semantic segmentation of sewer sonar point clouds. It is based on a U-Net style encoder-decoder architecture and consists of the attention module, the contraction module, and the expansion module. Additionally, to mitigate data imbalance, a weighted focal loss was employed during training. VAPCNet demonstrates excellent performance on a sewer dataset collected by a sonar robot, achieving an overall accuracy of 95. 9 % and a mean Intersection over Union (mIoU) of 86. 4 %. It demonstrates robustness to point perturbations and supports a lightweight design by adjusting encoder stages without sacrificing accuracy. These advantages make VAPCNet an innovative solution for employing sonar technology in sewer detection, emphasizing its practical potential.

Details DOI

EAAI Journal 2024 Journal Article

Multisensor data fusion approach for sediment assessment of sewers in operation

  • Chen Li
  • Ke Chen
  • Hanlin Li
  • Hanbin Luo

Urban sewer systems are essential components of urban water infrastructure, but their operations are often affected by sediment. Existing sediment assessment methods generally adopt closed-circuit television (CCTV) or individual types of sensors, but they fail to accurately locate the sediment and quantify the sediment volume. More seriously, these methods become ineffective in an operating sewer. In this study, a sewer sediment assessment approach is developed based on multisensor data fusion (SA-MDF). The raw data are collected by a remotely operated vehicle (ROV) equipped with a rotating sonar device, a gyroscope, an accelerometer, and an odometer. Subsequently, a two-step process of multisensor data fusion is implemented. In the first step, the unscented Kalman filter (UKF) and Rauch-Tung-Striebel (RTS) are applied to fuse the gyroscope, accelerometer, and odometer data, thereby achieving precise localization of the ROV inside the sewer. In the second step, the point cloud data collected by sonar are fused with the sensor data to address the point cloud data offset caused by the sewage flow and robot motion jitter. The laboratory and field experiments demonstrated that the SA-MDF can effectively reduce the ROV location error from 32. 37% to 0. 7% and the sediment quantification error from 15. 25% to 6. 38%. As a result, the SA-MDF facilitates accurate sediment assessment of the sewer, which offers valuable support for sewer maintenance decisions.

Details DOI
v2026.09.13