：In order to solve the problem of human face detection under weak light, a quick locating method based on the depth information by Kinect device is proposed. Firstly, 2D Chamfer Match method is used to rapidly preliminary locate the human face, and region-growing algorithm is applied on the preliminary locating face. Then, combined with extended Harr features, the AdaBoost algorithm is used to train the strong classifier under weak light, which realizes the precise face detection. The experiments show that rationally utilizing the depth information to detect the human face under the normal light can reduce the scope of the search window, which accelerates the detection speed, and the proposed method has strong robustness and timeliness.