H.264视频编码的三步全搜索运动估计改进算法及性能仿真研究

随着H.264标准的确定,网络传送的阻塞、错码等问题已成为了网络视频传输的研究热点,与此同时FPGA强大的并行能力实现了巨大复杂的运算,使得全搜索运动估计成为了急待解决的问题。为了进一步优化H.264视频编码的实时性及改进视频压缩性能的好坏。设计了一种全搜索运动估计模块的算法,确定了峰值信噪比PSNR值来实现图像质量评定;完成了新的三部全搜索算法设计;通过Modelsim仿真验证完成存储单元功能模块、数据选择功能模块的设计。系统测试表明:波形仿真结构验证了整体结构的正确性;FPGA的结果进一步完成了运动矢量采集成功的验证。对固定块大小范围的运动估计、结构以及数据流的设计研究对于FPGA的进一步大批量应用具有一定的意义。

Research of H.264 video coding three-step full search motion estimation algorithm and performance simulation

With the determination of the H.264 standard, network issues such as blocking, coding becomes a research hotspot network video transmission, at the same time strong parallel FPGA capabilities to achieve a huge operation, which makes the full search motion estimation became a pressing issue. To further optimize the H.264 video coding and improvement of real-time video compression performance is good or bad, has designed a full search motion estimation algorithm for module, I mainly do the following work: PSNR values for image quality evaluation were identified; completed three new search algorithm design; storage modules, through the Modelsim simulation design of data selection module. System test showed that the waveform simulation structures show the overall structure of the correctness FPGA completing further motion vector collection success as the result of validation. To fixed block size motion estimation of the range in the text, structure, as well as further research on the data flow design for FPGA, high volume applications have a certain significance.