Ageing contributes to a lack of independence on the part of the elderly, whether in their own homes or care settings. The world is looking into alternative ways to increase the independence of the elderly. However, there are many risks associated with the independent living of the elderly. This paper presents an activity monitoring system which uses computer vision techniques to monitor the activities of vulnerable older people. The algorithms used in the system can detect their movements and track their activities which can build their activity profile. Once a deviation from their daily pattern is detected, the system can alert caregivers and family members. The proposed system has the potential to be a building block for remote elderly care services, in which they can live independently yet will have the attention of caregivers, if needed. Further trials are underway in the United Kingdom and South Africa, which will help to generalise the results of this research.