Concurrent acquisition and processing of multi-spectral shadow information for 3D computer vision

This paper presents a novel method of recovering the three-dimensional information of objects using multispectral shadow information. In the method, multiple light sources of different power spectral distribution and illumination direction are concurrently projected onto the object. This increases both the image acquisition and subsequent processing speed. A novel method of optimizing the allocation of the power spectral distribution of light sources is formulated. This not only ensures robust three-dimensional (3D) visual information recovery, but also overcomes the problem of the drifting of reflectance characteristic of the object and background. A fast 3D vision inspection and measurement technique is suggested which further enhances the processing speed