In this paper, we describe a fully operational real-time object detection system using a MobileNet SSD model with the COCO dataset on a Raspberry Pi-3. The system is written in Python as it provides the OpenCV library that enables to capture and process video from the cam- era in real-time. This entails that every single frame in every video has to go through object detection where objects in the frame are labeled and boxed to enhance the understanding of the final output. In this regard, the MobileNet SSD model has been optimized by applying deep learning approaches to achieve enhanced efficiency and reliability of the results in any environment. The Raspberry Pi is a cheap but powerful hardware that can be employed to develop this advanced real-time object detection system for different applications. This is because the MobileNet SSD model is a lightweight network that can be implemented on Raspberry Pi because of its low computational power without affecting the detection rates or accuracy. The experimental results of the proposed system showed the possibility of the identification of the object and its class in real-time, which makes the system useful for surveillance, automation, and IoT systems. This work proves that complex object detection methods can be deployed on low-cost Mini ITX mainboards that are easily accessible, thus making many fields accessible to such models.