In optical see-through MR,virtual objects and virtual objects are mixed optically and shown to observers. So the observsers can see images in which virtual images and real images are mixed. But optical see-through MR which has been used could not visualizing occlusion between virtual images and real images[1-2],because objects' 3-D shapes and locations in real world were not measured so that real objects' and virtual objects' 3-D locations could not be defined.
Fig1 Optical See-Through MR which has been used
Fig2 A Real Image Occluding a Virtual Image
Fig3 A Virtual Image Occluding a Real Image
For example,we use a CG of a ble sphere as a virtual image,and use a petbottle as a real image. In the method which has been used, the real image is seen through the CG regardless of their locations(fig1). But it is ideal to visualize occlusion between virtual images and real images;the real image hide a part of the CG which must be not able to be seen when the real image is located in front of the CG(fig2). Of course,the CG should hide a part of the real image which must be not able to be seen when the CG is located in front of the real image(fig2). So in this research,we suggest a method of visualizing occlusion between real objects and virtual objects in optical see-through MR by measureing 3-D shape of real images in real-time,and made a trial system named ZARD(Zkeying AR Display).
