Background: The proposed system describes a surveillance system developed using Raspberry Pi and a camera, which keeps monitoring a certain highly secured region continuously. When the system recognizes a change in motion (human motion) compared to its previous frame, the system starts recording video and stores it primarily in its memory and also in the cloud (for the reason that even if the burglar tries to destroy the system his image/video will be saved in the cloud storage), and the user receives alert mail from the system stating “human motion detected” along with the captured image attached with the alert mail. The system contains database of face patterns of local suspects which is compared with the face pattern of the person triggering the system, and image processing is done in real time to correctly identify the detected face; the system also keeps tracking the face throughout the region even if the person moves out of the frame by a camera mounted over a servo motor. The system turns on a buzzer alarm when the burglar attempts to cause damage to the system. The system allows the user to remotely access the camera to monitor live streaming video output and control the rotation of the camera. Methods/Statistical analysis: In this project, different types of surveillance systems which already exist are analysed, and the methods of having a portable surveillance system were developed using Raspberry Pi. Image processing methods for facial identification and face recognition is used. Findings: A study based on various image processing techniques is done; it is found that Haar-cascade and linear binary pattern are the suitable algorithm for performing image processing in real time. Application/Improvements: For better surveillance, face tracking in introduced, which can track the detected face throughout the region even if the person goes out of the camera frame, and remote accessing with control of the camera through IoT is introduced. © Springer Nature Singapore Pte Ltd. 2018.