How to determine camera's next best view based on current information is a challenging problem in visual field.A next best view approach was proposed based on the occlusion information of a single depth image.Firstly

to establish the model for occlusion region external surface

the quadrilateral meshes for occlusion region external surface were obtained according to the occlusion information of a depth image in initial view.Secondly

the model for next best view was constructed by considering both the visible quadrangle and the loss information in next view.Finally

the next best view was achieved by solving the model with gradient descent method.Compared with the existing methods

the proposed approach does not limit the camera position on a fixed surface or need the priori knowledge of visual object.Experimental results demonstrate its feasibility and effectiveness.