Video Coding Technology And System Implementation In Digital Home-Networking Environments

This thesis focuses on the system implementation for the SOPCA digital TV system. For the MPEG-4 AVC/H.264 encoding, this thesis proposes rapid algorithms for motion estimation and mode decision, which improve the encoding speed greatly and only incur little video quality drop in accord with the requirement of SOPCA system. With the rapid progress of multimedia and network technologies, people need more and more powerful multimedia/intelligent functions from home facilities. The new MPEG-4 AVC/H.264 video coding standard has been accepted by the next generation DVD standard. It could be seen that MPEG-4 AVC/H.264 standard will take up an important status in the future home multimedia applications. Firstly, SOPCA digital TV system is introduced and the driver implementation is explained in detail. SOPCA is a key project supported by the Chinese IT industry ministry, which aims at achieving an information-controlling platform of home facilities. Based on the SOPCA technologies, the digital TV system is actually a digital set-top-box, which complies with the DVB standard. The OS used is embedded Linux. The system framework and driver implementation, which includes the drivers of I2C, demodulator and EEPROM, are described concretely. Their key functions' flow charts are this part's emphasis.Secondly, two fast integer-pel motion estimation algorithms for MPEG-4 AVC/H.264 encoding are proposed. The algorithms of UCMRGS and DCMHS are put forward to reduce the computation complexity and improve the speed for JVT software. The two algorithms utilized the different multi-resolution grid search pattern and take advantage of the dynamic cross search correlation among variable blocks, which could considerably reduce the computational cost of ME at the encoder, while at the same time give similar, and in some cases better, visual quality compared with the brute force full search algorithm. The two algorithms can be 10 and 20 times faster than the Fast Full Search in JVT software respectively.Lastly, the fast mode decision algorithms are proposed for MPEG-4 AVC/H.264 encoding. Based on some statistical features and local parameters, some unnecessary modes are skipped and the priority of the SKIP mode is improved and the "sub SKIP mode" is proposed for it. The algorithms improve the encoding speed greatly, while incur acceptable video quality drop.