Gas-Rd.png

Gascoigne Road & Flyover

#Semantic Segmentation

01 /

 
ro_gas_rd-1.png

Gascoigne Road & Flyover

#Semantic Segmentation

Large scale of point clouds is acquired by a mobile laser scanner along Gascoigne Road with near 325m length. Half of the data is being used as training set for segmentation of different classes such as light poles, traffic lights poles, pedestrian fences, road curbs, vegetation, etc. The semantic segmentation result by the trained neural network is promising. The light poles, traffic lights poles, pedestrian fences are distinguished by color as orange, violet, yellow respectively.

 


Bruce-Lee.png

Sculpture of Bruce Lee

#3D Model Reconstruction

02 /

 
BL_view1.png

Sculpture of Bruce Lee at Hong Kong Heritage Museum

#3D Model Reconstruction

Reconstruction and refinement of 3D model of "Bruce Lee" from 60 thousands of point clouds captured by terrestrial laser scanner and mobile laser scanner. The scan process takes 15 minutes with 4 static setups and 1 mobile scanning.

 


CC_Suc.png

The Flying Frenchman Monument

#3D Model Reconstruction

03 /

 
scu_volume.png

The Flying Frenchman Monument at Hong Kong Cultural Centre

#3D Model Reconstruction

By 3D model reconstruction from 660 thousands of point clouds, the volume - 4.9514m³ & weight - 41.59ton of "Flying Frenchman" is estimated with the 3D model.

 


Sun-Chui-Estate.png

Sun Chui Estate Precast Sculpture

#3D Model Reconstruction

04 /

 
scEst_model_view1.png

Sun Chui Estate Precast Sculpture

#3D Model Reconstruction

The volume & weight of the precast sculpture are being estimated from point clouds. The volume is approximate 38m³ and the weight is around 87.4ton. These information is crucial for future removal or relocation. 

 Learn More