Đang chuẩn bị liên kết để tải về tài liệu:
Mobile Robots Navigation 2008 Part 14
Đang chuẩn bị nút TẢI XUỐNG, xin hãy chờ
Tải xuống
Tham khảo tài liệu 'mobile robots navigation 2008 part 14', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | 508 Mobile Robots Navigation c -MPEG-4 d -VP7 Fig. 6. R D performance subjective quality at lowest bitrate. In Fig. 5 we show the R D Rate Distortion performance of the evaluated codecs with Foreman QCIF video sequence and no packet losses. This figure shows what we obtained in most of our loss-free tests at similar bitrates H.264 AVC offers the best quality nearly followed by VP7 and MPEG-4 offers the worst quality with difference. In Fig. 6 we can perceive the subjective quality difference between the codecs for the first frame of the video sequence compressed at the lowest bitrate. 4.1 Simulated packet losses evaluation In Fig. 7 we evaluate the behavior of the selected video codecs when the video delivery system begins to lose packets. We have used Foreman QCIF video sequence working in intra mode a medium compression level and a fixed packet size of 1272 bytes. We show the R D results with different PLR values. It can be seen that in absence of packet losses H.264 AVC and VP7 get similar results much better than the ones obtained by MPEG-4. It can also be seen that H.264 AVC does not fall as much as VP7 when PLR increases. Fig. 7. R D performance of codecs under packet losses. 509 Testing performance of current video codecs in teleoperated mobile robot applications a practical experience Packet Error Ratio Fig. 8. Resilience of intra and inter compression modes. Fig. 8 shows the video quality measures for Foreman QCIF video sequence using H.264 AVC codec with a medium compression level and a packet size of 512 bytes. Here we show the difference between intra and inter modes in terms of error resilience. The main reason that explains the lower error resilience behavior of inter mode coding is due to the propagation of errors to several frames. So if we find an error when decoding one frame this error will propagate to those frames that require the former one to be reconstructed. As a consequence the reconstructed video quality will be significantly reduced. .