Design And Research Of Heterogeneous Dual - Core Dual Video Data Processing Platform

Users rely on video image processing technology, more and more weight is being gradually replaced by word processing and audio processing technology, and has become a landmark technology division of the PC era with the mobile Internet era. Specific integrated chip based on the traditional ASIC design needs for special purposes, often limit the chip’s functionality, flexibility is poor, with the emergence of highly integrated chip technology, single-core to multicore development must be an important trend in embedded systems to meet the changing needs of users and complex market environment.In this paper, TI introduced for digital media development DM3730DaVinci system processor and the Android system platform, using video capture coding technology, communications technology between the dual-core and memory management technology, designed to support a set of analog and digital video signals double channel switching video processing platform, to achieve a two-way switch, an important function of video capture and video coding. In the hardware design to the pursuit of low power and high performance for the purpose of using heterogeneous dual-core processor, ultra-low power decoder chips and high-resolution camera as the main module. In software design, DaVinci Codec Engine software technology designed to encode, Codec Engine software is a multimedia software development framework abstracted into ARM-side application and DSP software algorithm ends two parts, can be independently developed dual-core, Codec Engine encapsulates the details of the underlying implementation, providing a large number of API interface, which greatly reduces the complexity of development, shorten the development cycle, but also has strong scalability.In this paper, the design of video processing platform based on the study of the shared memory communication mechanism between the dual-core and memory management algorithms and modules for TI’s CMEM insufficient memory allocation algorithms exist, the improved scheme, this scheme improves memory CMEM distribution rate, thereby increasing the efficiency of video processing. In this study design, in order to facilitate the video data processed for network transmission, to join the network transmission hardware module, but did not transplant the corresponding streaming protocol, this feature needs to be added in future studies.