.png){kind=icon}
.png){kind=icon}
Return
ID-SSLNet: Intensity Difference-Based Semi-Solid-State LiDAR Planar Markers Detection Network
.png){kind=icon}
PRE
AI
DOI：10.1109/JSEN.2026.3665301.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
With the rapid development of autonomous navigation technology, the demand for accurate environmental target perception has been steadily increasing. As a precise positioning sensor, semi-solid-state LiDAR (SSSL) is often treated with the same object detection approaches as mechanical LiDAR, while its unique characteristics have received limited attention. Owing to its higher imaging density, SSSL can capture finer patterns of planar markers composed of materials with varying reflectivity, thereby providing richer feature representations in the reflection intensity channel. Motivated by this observation, and building on an analysis of the intrinsic properties of SSSL along with insights from recent advances, we propose a reflection intensity-aware voxel feature encoder (RIA-VFE) and design an efficient height feature refinement backbone (HFRBackbone). These two components are then integrated into ID-SSLNet, a framework specialized for SSSL-based object detection. Extensive experiments on our custom dataset demonstrate the superior performance of the proposed methods and ID-SSLNet. Comparative studies on the PandaSet dataset (PS dataset) further show that ID-SSLNet achieves accuracy comparable to baseline methods in major categories while requiring only 68% of the parameters, and it significantly improves the detection of long-range targets with reflective differences under relatively small-sample conditions. Our discussion indicates that the intensity difference planar markers detection method of SSSL may have potential in tasks where cameras are ineffective or LiDAR is indispensable. The source code will be available at here
Keywords:
Autonomous driving
computer vision
object detection
semi-solid-state LiDAR (SSSL)
sign recognition