Estimating free space in an elevator's confined environment

This paper proposes a computer vision approach aimed at improving the operation of an elevator based on its spatial occupancy. Besides parameters such as weight and number of persons, we use floor space as an additional parameter to determine occupancy more accurately. Our approach combines the use of Horn-Schunck optical flow to monitor opening and closing of the elevator's door so that the floor segmentation algorithm, which uses a colour detection scheme based on HSV colour space to calculate the amount of free floor space by learning the specific variation in hue, saturation and brightness of the floor area. The projected outcome of this work is to reduce the number of unnecessary stops when the elevator is full in order to conserve power consumption and provide faster service to users.