Machine vision method using search models to find features in three-dimensional images
US6539107B1 · kind B1 · utility
Assignee
Inventors
Key dates
| Filing date | Apr 26, 2000 |
| Grant date | Mar 25, 2003 |
| Priority date | — |
| Expiry date | Apr 26, 2020 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06T2207/30152
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A system and method for analyzing 3D captured image data of scanned surfaces to locate object of interest, such as solder balls. A 3D model is created of an object of interest, the 3D model including weighted “don't care” areas where a match is not required. The 3D models which are used include geometric shapes, including a frustum, and models created from actual 3D image data of real objects. The 3D captured image data is processed to locate objects matching the 3D models, including processing by normalized correlation. Once the objects are located, the system selects data points within the 3D captured image data that are a predetermined distance away from the located objects. These data points are analyzed to determine a surface plane which fits the selected data points, thereby locating the surface in the 3D capture image data.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.