This paper proposes a markerless system whose purpose is to detect falls of elderly people at home. To track human movements we use Microsoft Kinect camera which allows to acquire at the same time a RGB image and a depth image. One contribution of our work is to propose a method for fusioning the information provided by several Kinects based on an occupancy grid. The observed space is tesselated into cells forming a 3D occupancy grid. We calculate a probability of occupation for each cell of the grid independently of its nearby cells. From this probability we distinguish whether the cells are occupied by a static object (wall) or by a mobile object (chair, human being) or whether the cells are empty. This categorization is realized in real-time using a simple three states HMM. The use of HMMs allows to deal with an aliasing problem since mobile objects result in the same observation as static objects. This paper also deals with the issue regarding the presence of several people in the field of view of camera by using the component labelling method. The approach is evaluated in simulation and in a real environment showing an efficient real-time discrimination between cells occupied by different mobile objects and cells occupied by static objects.