Application Research Of Face Detection In Elevator Group Control

The distribution of passenger flow is a critical parameter in Elevator Group Control System. The efficiency and service of elevators are furthest improved on condition that the Elevator Group Controller assigned elevators with the distribution. In this thesis the distribution is obtained by using the technology of face detection, namely the number of faces in real time images. This thesis improves the algorithm of face segment and presents a method to detect face.Image capture and pan-tilt control are firstly achieved based on HiKivision's image capture board. Then the Image Pretreatment, which is a former work for face segment and face detection, is fulfilled. The thesis adopts a median filter to smooth image and then compensates light for the image.Skin-color segment is one of the frequent methods to segment face. This thesis presents a face segment algorithm which uses skin color clustering and partial edge, the skin-color areas are firstly obtained by using skin color clustering, then according to size of the area to decide whether continue to segment the area by using partial edge. The algorithm, which aims at that one candidate area represents one face, solves the case of skin-color non-face area connected with face area.Face detection has been a basic yet difficult problem in computer vision field. Here this thesis presents a novel integrated evaluated function which has four factors, geometry character matching, ellipse matching, hair matching and face feature matching. The candidate area is confirmed by the value of function.