Image analysis – Applications – Target tracking or detecting
Reexamination Certificate
2000-05-18
2004-10-12
Mariam, Daniel (Department: 2621)
Image analysis
Applications
Target tracking or detecting
C382S154000, C382S225000, C382S288000
Reexamination Certificate
active
06804380
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a system and method that acquire tie-point location information on a structure and, more particularly, to a system and method that acquire tie-point location information on a structure with minimal user assistance.
2. Description of the Related Art
It is often desirable to have computer models of large physical structures, such as a petroleum refinery. Computer models facilitate an understanding of the structure that is beneficial in a number of ways, including facilitating an understanding of how materials flow through the structure.
The computer models are constructed by acquiring geometric information about the to-be-modeled structure. One technique for acquiring the geometric information is to use a laser scanning system. In a laser scanning system, the to-be-modeled structure is scanned with a laser from a plurality of views.
Each scan includes a plurality of scan points that cover a view of the structure. Each scan point has a measured location in 3D space, usually in terms of an (x,y,z) point in the scanner's local coordinate system. In some cases, the 3D point data also includes data that indicates the intensity of the laser light reflected from the point.
When a view is scanned, the laser scanning system generates 3D point data for a large number of points. The resulting collection of points is often called a point cloud. The point cloud typically has points that lie on many different surfaces, depending on the view that was scanned. For example, a scan taken at a petroleum refinery may include a point cloud that has points on pipes, elbows, valves, pumps, and structural steel members.
Once each view of the structure has been scanned, the 3D point data are processed to generate a computer model of the structure. One of the first processing steps, known as scan registration, combines the views together so that the points from each view share a common coordinate system. For example; scan registration combines a view of the front-left side of a structure and a view of the front-right side of the structure so that both views share a common coordinate system. The registration process calculates the transformation needed to bring each scan into a desired coordinate system, and then applies the transformation to the 3D scan data to give a new collection where all scans share the same coordinate system.
Most conventional scan registration methods use a constraint-based approach, where the first step in the scan registration process is the identification of a number of constraints. Constraints are created by matching features that a number of views share in common. For example, matching points in different views can be used to create a constraint. In addition, co-linear or parallel lines in different views can be used to create a constraint, and co-planar or parallel planes in different views can be used to create a constraint. If points are matched to form a constraint they are often called tie points because they tie the data sets together.
In addition, objects can be modeled, and then a feature of the model used to create a constraint. For example, views of a petroleum refinery may each include a cylindrical storage tank or a spherical gas tank. In these cases, the user can model the cylindrical storage tank or the spherical gas tank in each view, find the centerlines of cylinders or the centers of the spheres, and match these features in the different views to create constraints.
Once the constraints have been identified, they can be used in a registration process that determines the optimal transformations for each data set so that the identified matching features join as closely as possible. Each view must share a number of constraints to solve for the rigid body transformation, and the minimum number of constraints required depends on the type of constraint (point or line). For example, if two views include three spherical gas tanks, the spherical gas tanks in each view can be modeled to find the points at the centers of the three spheres. Having the three matching points is enough to solve the rigid body transformation problem. It should be noted that constraints can also be formed by matching features from scan data with features acquired through some other mechanism, such as surveying. A sphere could be scanned and its center point determined. A constraint could then be formed between that center point location and the surveyed position of the sphere center point. The resulting registration transformation can be used to bring the scan data into the survey coordinate system.
One of the problems with conventional scan registration methods is that it is often a time consuming process to identify the features that can be used to create the constraints. Lines and planes are often difficult to find from the large number of points that make up a view. Geometric shapes, such as a cylinder or a sphere, are also difficult to find and, once found, frequently require a significant amount of time to model. Thus, there is a need for an apparatus and method for acquiring tie-point location information on a structure that reduces or eliminates the need for the user to manually identify features for use in creating constraints.
SUMMARY OF THE INVENTION
The present invention provides a system and method for acquiring tie-point location information on a structure that reduces or eliminates the need for the user to manually identify features for use in creating constraints. In the present invention, readily identifiable objects, known as tie-point targets, are applied to the structure. When the structure is scanned, the tie-point targets are identified and can be utilized to create registration constraints.
The system of the present invention includes a plurality of tie-point targets and a laser. The laser scans the structure with a laser beam to generate a plurality of points. The laser beam illuminates the tie-point targets during the scan. Each point, in turn, has point data that defines a spot on a surface of the structure.
The system of the present invention also includes a computer that controls the laser and processes the plurality of points. The computer includes means for forming a number of groups of target points from the plurality of points where each group of target points has a recognizable feature. The computer also includes means for selecting a number of the groups of target points for acquisition to form a number of selected groups of target, points.
In addition, the computer includes means for forming a designated point for the recognizable feature in each selected group of target points. The designated point represents the recognizable feature in each selected group of target points. Further, the computer includes means for labeling the designated point in each selected group of target points as a tie-point location.
The method of the present invention begins with the step of affixing a plurality of tie-point targets to the structure. Following this, the structure is scanned with a laser beam to generate a plurality of points. The laser beam illuminates the tie-point targets during the scan. Each point, in turn, has point data that defines a spot on a surface of the structure. The method also includes the step of forming a number of groups of target points from the plurality of points where each group of target points has a recognizable feature.
In addition, the method includes the step of selecting a number of groups of target points for acquisition to form a number of selected groups of target points. Further, a designated point is formed for the recognizable feature in each selected group of target points. The designated point represents the recognizable feature in each selected group of target points. The method also includes the step of labeling the designated point in each selected group of target points as a tie-point location.
A better understanding of the features and advantages of the present invention will be obtained by reference to the following detailed descript
Ioannou Dimitrios
Kung Jonathan Apollo
Thewalt Christopher Robin
Wheeler Mark Damon
Leica Geosystems HDS, Inc.
Mariam Daniel
Patel Shefali
Stallman & Pollock LLP
LandOfFree
System and method for acquiring tie-point location... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for acquiring tie-point location..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for acquiring tie-point location... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3291940