TY  - EJOU
AU  - Liu, Xu 
AU  - Zhu, Fa 
AU  - Alfarraj, Osama 
AU  - Alblehai, Fahad 
AU  - Khan, Shakir 

TI  - Urban Roadside Crown Identification via Intensity-Calibrated Mobile Laser Scanning Point Cloud
T2  - Computers, Materials \& Continua

PY  - 
VL  - 
IS  - 
SN  - 1546-2226

AB  - Mobile laser scanning (MLS) plays a crucial role in urban vegetation monitoring by enabling the precise identification of roadside trees. However, in cluttered environments where tree crowns intermingle with various objects, methods relying solely on raw intensity data face substantial accuracy limitations. To address this issue, a dedicated calibration method is proposed, which models the relationship between acquisition geometry and intensity values through polynomial fitting. Specifically, calibration models were first established for a given LiDAR sensor using reference data from a standard diffuse reflection panel. A customized MLS platform equipped with this sensor was then deployed to acquire point clouds along urban roadways. After preprocessing, locally planar point clouds were retained for intensity calibration. The calibrated intensity values were subsequently used for tree crown identification. Experiments on an 80 m road segment show that the calibrated intensity improves the F1 score from 0.63 to 0.83, a 32% increase, demonstrating its effectiveness in real-world scenes.
KW  - Remote sensing; mobile laser scanning; crown recognition; intensity calibration; urban greenery monitoring

DO  - 10.32604/cmc.2026.079640