One of application of automation technology is Autonomous Mobile Robot (AMR), which has a navigation so that it can move around the unknown environment. This AMR ability can be used for moving around the dangerous area for people such as a place where new gas is found. As an application of automation technology, AMR needs a sensor for recognizing the environement. One of the sensors used in the automation technology and application is camera, which has the same characteristic as human eyes. Drawback of camera sensor is a limitation in the field of view of the camera. Camera can only catch the image or data of an object facing to the camera, whereas AMR neesd to know the data in the other direction or field of view. This can be overcomed by using omnidirectional vision sensor system. This system can give information of environment around the camera (3600) in one image. This system can be done by facing the camera to a convex mirror. Fisrtly, AMR has to detect the existence of the objects around it so that the collision can be avoided. This research focused on regocnizing an object and determining the direction of that object related to AMR. The designed omnidirectional vision sensor system is only used as the tool for data acquisition and has not been integrated with AMR.