Cloud based Single Shot Multi Box Detection (SSD) Architecture Models for Jakarta Traffic Jam Monitoring

Abstract

Traffic jam is still one of the main problems in Jakarta Indonesia. To encounter this problem, we proposed system to monitor the road and perform calculation on vehicles speed and road density. To achieve this, we used SSD to detect vehicle traffic on Jakarta’s road which obtained from public IP Camera. The main contribution of this work are utilization of public IP Camera and remote traffic analytics via cloud based artificial intelligence system. As a result, the program can perform monitoring on roads condition in real time. The accuracy of the system in average is 80% with highest accuracy achieved 92% to detect vehicle speed and road density.