Study on the perception mechanism and method of virtual and real objects in augmented reality assembly environment

Augmented reality assembly, is a promising application of virtual reality in design and manufacturing and has drawn much more and more attentions from industries and research institutes. The focus of this study are as follows. (1) using the agent-oriented modeling for subassembly object and the behavior reasoning based on petri net model of environment, to build the system model of augmented reality assembly environment and construct the percepting mechanism of the virtual and real objects; (2) based on above percepting mechanism, to solve the epipolar geometry constraint by least robustness square estimation model, and to sense the depth of the occluded virtual objects or real objects in the environment through assembly scene recognizing, background eliminating and geometrical features restructuring of product contour; (3) combining awareness process of object features matching, constrained motion and features fitting with image based visual hull generating and volume querying, to detect the collision of virtual and real objects, including the dynamic deformable objects. The research results is of very important academic and applicative interest in the field of digital design, assembly and maintenace. It would be helpful to promote the application research of augmented reality and virtual reality technology.