• Click:

Current position: Home  >  Research & Projects

Research & Projects

1. Meter image analysis and recognition system

(Mar. 2014 – present)  

The aim of this project is to detect and recognize the character and symbol in virtual LCD meter image for industrial automatic monitoring. To be specific, the character processing involves feature extraction by improved Harris, segment and optical character recognition. And then, graphical symbols are detected and classified by hierarchy template match. The dial parameter in meter area is obtained by using Hough and enhanced methods. At present, other optimal algorithms are under our pursuit.

 

2.Video communication and processing system for UAV

(Aug. 2011 – present)

This system is used to transmit HD video from UAV (Unmanned Aerial Vehicle) to receiver in ground stations. DVB-T, as a prefect video terrestrial broadcasting means, is introduced to solve the problem of wireless communication. H.264 codec with HD definition up to 1080P is used for video compression and decompression. A semi-RT image stitching algorithm is proposed, by which the tiles image with relative coordinates to form a panorama can be produced from UAV HD video.

 

3. High speed well-logging communication system

(Jun. 2006 – present)

This long-term project was divided into a series of sub-projects founded constantly by COSL (China Oilfield Services Limited) to research and verify the high speed wire line communication system. In this project, custom communication system including transmission architecture, OFDM (Orthogonal Frequency Division Multiplexing) modulator, light TCP/IP stack, time synchronization method and data processing algorithm are studied and verified for field application.

 

4. Real-time transcoding from MPEG-2 to AVS

(Dec. 2005 – Nov. 2006 )

The aim of this project is to transcode MPEG2 to AVS (Advance Video Standard of China) in multi-core processor. Based on the combination of GOP in video stream and hyper thread supported by multi-core processor, a parallel algorithm is presented to accelerate the transcoder. The intra and inter predication mode in H.264 are reused in AVS. A fast algorithm of I-frame transcoding in transform domain from MPEG-2 to AVS is proposed in this paper. The proposed algorithm includes AVS intra -prediction mode decision and fast coefficients transformation from MPEG-2 to AVS in transform domain.


伍瑞卿

Gender:Male Education Level:With Certificate of Graduation for Doctorate Study Degree:Doctor of Engineering Status:Professor Business Address:KB538 in 1st Scientific Research Building Contact Information: E-Mail: