US 8,576,235 B1Granted

Visibility transition planning for dynamic camera control

AssigneeDisney Enterprises, Inc.

InventorsRobert Sumner, Markus Gross, Nils Thuerey, Thomas Oskam

Life of this patent

Estimating the term…

Loading patent…

Abstract

A camera control system is provided that uses global planning to compute large, occlusion free camera paths through complex environments, incorporating visibility of a focus point into the search strategy, so that a path is chosen along which the focus target is in view or, in general, to compute a visibility graph. A visibility roadmap data structure permits precomputation of coarse representations of collision-free paths through an environment, together with estimates of pair-wise visibility between scene portions. At runtime, path planning can be done using the precomputed roadmap values to find a coarse path, and then refined to the camera path using occlusion maps computed on-the-fly. Iterative smoothing, together with a physically-based camera model, can be used to have the path followed by the camera smooth in both space and time. The data structure can be adapted at runtime to deal with dynamic occluders that move in an environment.

Claims39

Dependent claims

In a computer-implemented animation or gaming system that maintains a geometric model space having coordinates in order to generate one or more images representing virtual objects in that geometric model space within a time allotted for a computing system to generate the one or more images from a camera view position in the geometric model space based on a computer-readable input of a set of virtual objects positioned in the geometric model space and parameters of the geometric model space, wherein the animation or gaming system will generate a display of the objects in the geometric model space as viewed from the camera view position and perform operations based on visibility, a method of computing a camera path, comprising:

dividing the geometric model space into clear subvolumes, wherein a subvolume defines a space within the geometric model space that has nonzero volume and a clear subvolume is a subvolume that does not contain any virtual objects that are not allowed to be traversed;
building a graph structure, representable in electronic storage, the graph structure having nodes representing transition locations between adjacent clear subvolumes and edges representing transitions from one transition location in a given clear subvolume to another transition location in the given clear subvolume;
computing visibility weights for pairs of clear subvolumes, independent of a camera focus point and/or independent of the camera view position, wherein a visibility weight corresponds to a visibility measure between the clear subvolumes of the pair;
generating a computer-readable pair data structure having values corresponding to the visibility weights;
determining a path time period over which the camera view position is to change from a start point to an end point along a camera path;
determining a camera focus point in the geometric model space, wherein the camera focus point is a point or location in the geometric model space that the animation or gaming system will attempt to keep in view, wherein view is determined, at least in part, based on at least an approximation of whether opaque virtual objects are positioned along a ray spanning from the camera focus point to the camera view position as the system moves the camera view position over the camera path;
determining a preferred coarse path along the graph structure from a starting clear subvolume to an ending clear subvolume taking into account the visibility weights for pairs of clear subvolumes wherein one of the pair is a subvolume containing the camera focus point;
and
determining the camera path from the preferred coarse path.

6 more elements

A computer-implemented animation or gaming system that includes storage for elements of a geometric model space, the space having coordinates in order to generate one or more images representing virtual objects in that geometric model space within a time allotted for a computing system to generate the one or more images from a camera view position in the geometric model space based on a computer-readable input of a set of virtual objects positioned in the geometric model space and parameters of the geometric model space, wherein the animation or gaming system further includes an output usable for generating a display of the objects in the geometric model space from output data, as viewed from the camera view position and usable to perform operations based on visibility, the system comprising:

one or more computer-readable tangible media configured to store subvolume details that represent a partition of the geometric model space into clear subvolumes, wherein a subvolume defines a space within the geometric model space that has nonzero volume and a clear subvolume is a subvolume that does not contain any virtual objects that are not allowed to be traversed;
a graph structure;
and visibility weights;
a processor coupled to the one or more computer-readable tangible media, the processor being configured to:
generate the graph structure, the graph structure having nodes representing transition locations between adjacent clear subvolumes and edges representing transitions from one transition location in a given clear subvolume to another transition location in the given clear subvolume;
generate values for visibility weights for pairs of clear subvolumes, independent of a camera focus point and/or independent of the camera view position, wherein a visibility weight corresponds to a visibility measure between the clear subvolumes of the pair;
generate a computer-readable pair data structure having values corresponding to the visibility weights;
the one or more computer-readable tangible media being further configured to store a path time period value representing a period over which the camera view position is to change from a start point to an end point along a camera path;
the one or more computer-readable tangible media being further configured to store a camera focus point in the geometric model space, wherein the camera focus point is a point or location in the geometric model space that the animation or gaming system will attempt to keep in view, wherein view is determined, at least in part, based on at least an approximation of whether opaque virtual objects are positioned along a ray spanning from the camera focus point to the camera view position as the system moves the camera view position over the camera path;
the one or more computer-readable tangible media being further configured to store a preferred coarse path, wherein the preferred coarse path is a path along the graph structure from a starting clear subvolume to an ending clear subvolume taking into account the visibility weights for pairs of clear subvolumes wherein one of the pair is a subvolume containing the camera focus point;
and
the processor being further configured to refine the camera path from the preferred coarse path.

9 more elements

Dependencies are inferred from the claim text.