Generated Based On The Shadow Of The Synthetic Objects In The Images Of Complex Scenes

Shadow generation of composited object is always a research hotspot in Computer Graphics and Computer Vision, which has very important application value. The 3D-based shadow generation method appeared in recent years, reconstructs 3D scene and object model by the way of taking shot from multi-views of scene and object, then composites the 3D object model into the 3D scene to implement the shadow generation of 3D object in the scene. It has got very wonderful effect of reality.But nowadays, followed the development and popularization of network year after year, more and more network fans like to make some funny or interesting video clips which shot with their own video device by themselves, and put them into their own homepages or Blogs to share them with other friends on line. These guys don't demand very high precision of shadow generation of composited object, the thing they paid attention is more convenient and faster in making special effects, such as shadow generation of composited object.Based on this requirement, although the 3D modeling-based shadow generation could get very impressive effect, it is too strict for device, spot, photographer to popularize for general people. For solving this problem, the image-based shadow generation method of composited object emerges as the times require. It is very fit for the demand because it is simple and easy to manipulate, and still has good effect.In this paper, I have done a deeper and more meticulous research on this method, and present the image-based shadow generation method of composited object in complex scene. Approved by experiments, this method could generate realistic shadow of composited object handily in complex scene (including walls, steps…etc).Finally, I have applied this method to panorama-based video composition, and satisfy the requirement of network fans successfully.