Abstract: In order to enhance the immersion and authenticity of the augmented reality applications,real-time collision detection between virtual and real objects is essential.This paper proposes a collision detection(CD) and response(CR) algorithm of virtual objects to arbitrary-shaped real objects based on augmented reality(AR) and monocular vision.The proposed algorithm is developed from the existing 2D monocular CD and CR algorithm.To reduce the computational cost,our algorithm only needs to calculate four feature points.Through object segmentation,feature extraction,CD,CR and other steps,3D real collision effect can be obtained.Experiments in AR marker and markerless applications show that the computational cost of our approach is similar to the traditional 2D collision detection ones.The algorithm implements AR 3D CR estimation and processing based on monocular vision.