arturo rodriguez - Academia.edu (original) (raw)

Papers by arturo rodriguez

Iie Transactions, 1993

An industrial vision system capable of extracting the position, orientation, and identification o... more An industrial vision system capable of extracting the position, orientation, and identification of non-overlapping parts in 2-D scenes subject to moderately unconstrained conditions is presented, Typical scenes exhibit non-homogeneous object surfaces, transparency, shadows and specular reflections. System requirements include recognition of partial and unconnected objects, independent of object scale and rotation, to counterbalance incomplete and imperfect image segmentations that result from these conditions. Part of the approach consists of removal of unwanted image data. Shadows that would otherwise prevent detection of high-discriminant features arc identified and removed. Redundancy is included in feature space to compensate for imprecise feature measurements. Results for a typical industrial application are presented.

Iie Transactions, 1993

An industrial vision system capable of extracting the position, orientation, and identification o... more An industrial vision system capable of extracting the position, orientation, and identification of non-overlapping parts in 2-D scenes subject to moderately unconstrained conditions is presented, Typical scenes exhibit non-homogeneous object surfaces, transparency, shadows and specular reflections. System requirements include recognition of partial and unconnected objects, independent of object scale and rotation, to counterbalance incomplete and imperfect image segmentations that result from these conditions. Part of the approach consists of removal of unwanted image data. Shadows that would otherwise prevent detection of high-discriminant features arc identified and removed. Redundancy is included in feature space to compensate for imprecise feature measurements. Results for a typical industrial application are presented.