Three-dimensional measurement of object surfaces with complex shape and color distribution based on projection of color fringe patterns

摘要

A challenging issue associated with three-dimensional (3D) fringe patterns profilometry (FPP) is the unwrapping of phase maps resulting from color object surfaces. This paper proposes a new colorprojection-based 3D FPP, making use of the three primary color channels [i.e., red, green, and blue (RGB)] associated with digital projectors. One channel (e.g., red) is used for projecting sinusoidal fringes required by phase shift profilometry (PSP); the other two channels are employed for generating binary stripe patterns. In order to achieve reliable phase unwrapping, each fringe of the sinusoidal patterns is identified by a unique binary sequence. These sequences are then encoded by a channel-encoding scheme used in the area of communication. The encoded sequences are embedded in the binary coding stripe images, which are projected together with the sinusoidal patterns. The three image patterns are reflected by the object surface and captured by an RGB 3-CCD camera. The reflected sinusoidal patterns are employed to yield a wrapped phase map, and the binary stripe patterns are used to retrieve the encoded sequences, which are then decoded to yield the original binary sequences for phase unwrapping. Compared with existing color-encoded algorithms, the proposed approach uses binary codes instead of fringe color to identify the fringes, which are less sensitive to the disturbances caused by object surface color and illumination noises. Furthermore, use of the channel-coding scheme provides extra resistance to the disturbances caused by object surface color and illumination noises. Experimental results are presented to confirm the effectiveness of the proposed technique.