Shape reflector and surface contour mapping

摘要

The three dimensional surface shape of one or more layers of a reflective object is determined by examining one or more captured images reflected from the reflective object. Curved surfaces reflect a distorted image altered by the surface shape. By analyzing one or more captured images of the distorted reflected images, the shape of the surface that caused the distortion is estimated. A captured distorted image is compared to a reference undistorted image having known geometric properties. A system to capture and process such images is assembled from components including an image capture assembly such as a digital camera to capture reflected images and a positioning assembly on which to orient the components with respect to each other. Multiple surface layers of the reflective object are separately estimated using polarizations, luminance levels, chroma values or combinations thereof contained in one or more captured images.

主权项

1. A method for estimating a three-dimensional surface shape of two reflective surface layers of an object, the method comprising:
capturing, by an image capture device and when a polarizing filter is blocking a first reflected image pattern reflected from a first reflective surface of the object, a second reflected image pattern reflected from a second reflective surface of the object; capturing, by the image capture device and when the polarizing filter is blocking the second reflected image pattern, the first reflected image pattern, wherein the first reflected image pattern exhibits a different amount of luminance than the second reflected image pattern as a result of a difference between the first reflective surface and the second reflective surface; obtaining a first and second set of two-dimensional reflected image points from the first reflected image pattern and the second reflected image pattern respectively; calculating a set of three-dimensional direction vectors for the first set and second set of two-dimensional reflected image points; calculating a set of three-dimensional surface estimates for each two-dimensional image point along its corresponding three-dimensional direction vector, wherein each element of the set of three-dimensional surface estimates represents a different position of a surface of the object; and minimizing a curvature function of a curve connecting an estimate of the three-dimensional surface to a previously generated three-dimensional surface estimate.