- 學位論文
DVB-T數位電視接收系統實作研究
Studying and Implementation of DVB-T Client System
摘要
近年來數位電視的發展已蔚為世界潮流，相較於各國，我國雖於2004年6月1日開始無線數位電視地面廣播服務，然後，計劃於2010年能完全普及化。因為接收端設備之普及率與收視品質須待加強，此乃造成今日國內數位電視普及率偏低之主因。因此，本論文以我國數位電視為基礎，提出一種使用BDA（Broadcast Driver Architecture）驅動程式標準架構能符合歐規數位地面廣播（DVB-T）之接收器平台實作與研究。我們所提出的實作平台不但具有高度相容性與模組化置換性，並且以系統效能考慮為前提下，在解碼器中使用錯誤隱藏來改善傳輸錯誤時影像所造成的品質影響。本研究中使用了ZMEC (Zero Motion Error Concealment), SEC (Spatial Error Concealment), 適應性法(Adaptive Error Concealment), BMEC ( Block Matching Error Concealment), HBMEC (Hexagon-Based Search Block Matching Error Concealment)與WHBMEC (Weighted Hexagon-Based Search Block Matching Concealment)錯誤隱藏方法，從實驗證明，這些方法能有效提高畫質PSNR 21dB以上，更能提供更優異數位電視之播放品質。最後，在這些方法中，考慮影像品質與計算能力限制條件下，我們所提出的方法WHBMEC是較佳的。
並列摘要
The development of digital TV system has been already a trend in recent years. Comparison to that of other countries, In Taiwan, digital TV broadcast system has been ready for service at June 2004, and planed totally popularize in 2010. However, nowadays it is still far from the goal, since the poor popularity of receiving devices and image quality. Therefore, based on our digital TV system, we propose and implement a receiver platform with drivers of BDA (Broadcast Driver Architecture) to meet the DVB-T standard. Our proposed platform not only owns properties of high compatibility and modulize design but also uses the error concealment method of the decoder to improve the image quality cause of transmitting error under performance evaluation. In this research, having studied the methodologies of ZMEC (Zero Motion Error Concealment), SEC (Spatial Error Concealment), Adaptive Error Concealment, and BMEC (Block Matching Error Concealment), HBMEC (Hexagon-Based Search Block Matching Error Concealment), and WHBMEC (Weighted Hexagon-Based Search Block Matching Concealment), from our results, we know that these methodologies can improve over PSNR 21 dB of the picture quality, and offer more excellent broadcasting quality of digit TV. Finally, our proposed method WHBMEC can get the better performance among them over image quality and computation power limitation