No abstract

Before broadcasting a video signal, redundant data should be removed from the transmitted video signal. This redundancy operation can be performed using many video coding standards such as H.264/Advanced Video Coding (AVC) and H.265/High-Efficient Video Coding (HEVC) standards. Although both standards produce a great video resolution, too much data are considered to be still redundant. The most exhaustive process in video encoding process is the Motion Estimation (ME) process. The more the resolution of the transmitted video signal, the more the video data to be fetched from the main memory. This will increase the required memory access time for performing the Motion Estimation process. In This chapter, a smart ME coprocessor architecture, which greatly reduces the memory access time, is presented. Data reuse algorithm is used to minimize the memory access time. The discussed coprocessor effectively reuses the data of the search area to minimize the overall memory access time (I/O memory bandwidth) while fully using all resources and hardware. This would speed up the video broadcasting process. For a search range of 32 × 32 and block size of 16 × 16, the architecture can perform Motion Estimation for 30 fps of HDTV video and easily outperforms many fast full-search architectures.

No abstract

In the contemporary video surveillance system, there have been many efforts made to maintain high security and extend the coverage areas in most of the countries around the world. The deployment of many surveillance cameras and sensors capable of detecting abnormal and meaningful events on the territories' streets and airports is an aspect of internal security. There are two main problems that affect the homeland security system, and the security cameras and sensors are not enough to cover all areas in the country. This is because of the high cost of the video surveillance cameras and the sensor installations, and the non‐standardisation of security cameras and sensors manufacturing, which is due to the differentiated infrastructure of companies or organizations that provide home security. The authors introduce a design and hardware implementation of a motion estimation (ME) co‐processor that can be used for video surveillance cameras in homeland security. The proposed ME co‐processor, if adopted in video surveillance cameras, can be connected utilising an internet of video things infrastructure (IoVT). The proposed co‐processor is suited for high‐efficiency encoding video surveillance systems (H.265/HEVC). Furthermore, to reduce the memory I/O, data reuse Level A and Level B have been used in the proposed architecture while taking full advantage of the hardware resources. Moreover, an effective local memory has been used to reuse the data during the process of loading both the search area and the current block into the processing element array (PE array). The performance of the proposed architecture has been calculated using subjective and quantitative measures techniques and compared to the full search block‐based motion estimation (FSBB‐ME) algorithm. Moreover, the proposed architecture achieves a very high video resolution accuracy that is similar to the accuracy of the FSBB‐ME algorithm. Modelism‐version10.4a has been used for simulation and time verification testing proposes. The proposed ME co‐processor can be embedded in the compressing decompressing and high definition broadcast for video surveillance systems.