Enabling Cooperative Driving Automation (CDA) requires high-fidelity and real-time perception information, which is available from onboard sensors or vehicle-to-everything (V2X) communications. Nevertheless, the accessibility of this information may suffer from the range and occlusion of perception or limited penetration rates in connectivity. In this paper, we introduce the prototype of Cyber Mobility Mirror (CMM), a next-generation real-time traffic surveillance system for 3D object detection, classification, tracking, and reconstruction, to provide CAVs with wide-range high-fidelity perception information in a mixed traffic environment. The CMM system consists of six main components: 1) the data pre-processor to retrieve and pre-process raw data from the roadside LiDAR; 2) the 3D object detector to generate 3D bounding boxes based on point cloud data; 3) the multi-objects tracker to endow unique IDs to detected objects and estimate their dynamic states; 4) the global locator to map positioning information from the LiDAR coordinate to geographic coordinate using coordinate transformation; 5) the cloud-based communicator to transmit perception information from roadside sensors to equipped vehicles; and 6) the onboard advisor to reconstruct and display the real-time traffic conditions via Graphical User Interface (GUI). In this study, a field-operational prototype system is deployed at a real-world intersection, University Avenue and Iowa Avenue in Riverside, California to assess the feasibility and performance of our CMM system. Results from field tests demonstrate that our CMM prototype system can provide satisfactory perception performance with 96.99% precision and 83.62% recall. High-fidelity real-time traffic conditions (at the object level) can be displayed on the GUI of the equipped vehicle with a frequency of 3-4 Hz.