MEMTV: a Research on Multi-level Edge Computing Model for Traffic Video Processing

Abstract

Real-time traffic video data analysis and processing technology is an important part of intelligent urban transportation, real-time urban traffic planning, reducing urban traffic congestion, has certain application value. Aiming at the problem that network transmission link resources are limited and it is difficult to meet the real-time analysis and processing of massive video data uploading to the cloud. In this paper, we propose MEMTV, a multi-level edge computing model based on big data machine. The model obtains scene semantic information by processing traffic video data at the edge intelligent terminal, and uploads the scene semantic information as textual metadata to the edge server for analysis, while the edge server only does the inference of the model, and the update training of the model is performed in the cloud, which does not affect the edge data processing while ensuring the high reliability of the prediction model. The experimental results show that the MEMTV can effectively process traffic video stream data.