Research

Share    

Video and image enhancement using rain, haze and fog removal techniques

Objective

To provide a real time all-weather vision enhancement system for a moving platform.


Description

Weather conditions such as rain, haze and fog significantly deteriorate the performance of a modern intelligent system based on camera inputs. 

In this project, progress has been made in developing an algorithm for rain removal using the dynamic scene with large depth variations obtained from a moving camera. This algorithm is superior over other conventional methods as rain removal can be achieved optimally even in the presence of dynamic contents. 

Going forward, research effort will focus on the development of software algorithm in supporting the real-time removal of fog and haze in the frame sequence.


   
                                                      Vision enhancement for a moving platform



Result

System Performance

  • Fast static and moving cameras with speed ranges from 0 to 30 km/h
  • Enhancement in SURF feature matching after de-rain:

o   Distance range 5-20m: average enhancement 29.5%

o   Distance range 50-100m: average enhancement 6.7%

  •  3 Hz Processing rate @ VGA resolution  




Video




Principal Investigators


Associate Prof. Chau Lap Pui (NTU)

Telephone: 6790 4239
Office: S2-B2a-09
Email: elpchau@ntu.edu.sg ​
Mr. Paul Tan (STE)

Telephone: 6660 1052
Office: S1-B4a-03
Email: paultan@stengg.com