Smart station has its concern to passengers safety and security as it is the fundamental needs of the passengers satisfaction. CCTV (circuit closed television) has been commonly used as surveillance system in public places including railway stations. CCTVs could be optimized using video analytics to automized surveillance using machine learning. Trespassing is one of threatening activities that is happening in railway stations. The objective of this paper is to add trespassing recognition feature to VIANA existing platform using YOLOv5 as the detection model and DeepSort as the tracking model.