基于Spark的人脸检测并行处理系统

With the growth of data amount in the field of video surveillance, it becomes very necessary to build a distributed video surveillance system based on Spark. In this paper, we present a method for face detection parallel running based on Spark, and build a distributed video surveillance system for face detection based on Spark. The paper mainly studies the parallelization of face detection algorithm on Spark. We run the algorithm on Spark cluster composed of three nodes, and test the proposed algorithm. The experimental results show that the distributed video surveillance system based on Spark has greatly improved the efficiency of face detection. The system can be applied to security, finance, large-scale test management and other fields.