Fig. 1. Flowchart of the proposed method

Fig. 2. Key process of generating fringe pattern

Fig. 3. Simulation results for different Gaussian filter sizes. (a) 5×5; (b) 9×9; (c) 13×13

Fig. 4. Patterns generated by two methods. (a) Binary pattern obtained by Floyd-Steinberg error-diffusion dithering method; (b) binary pattern obtained by Floyd-Steinberg error-diffusion dithering method and Gaussian filter; (c) binary pattern obtained by proposed method; (d) binary pattern obtained by proposed method and Gaussian filter

Fig. 5. Comparison between Floyd-Steinberg error-diffusion dithering method and proposed method. (a)(d)(g)Cross-sections of ideal sinusoidal pattern and two binary patterns obtained by error-diffusion dithering method and the proposed method after applying Gaussian filter with sizes of 5×5, 9×9, and 13×13,respectively; (b)(e)(h) gray value differences of error-diffusion dithering method and the proposed method after applying Gaussian filter with sizes of 5×5, 9×9, and 13×13, respectively; (c)(f)(i) phas

Fig. 6. 3D shape measurement system

Fig. 7. Measurement results under different defocusing amounts. (a) Mild defocusing; (b) moderate defocusing; (c) severe defocusing

Fig. 8. Defocusing patterns by two methods. (a) Error-diffusion dithering method; (b) proposed method

Fig. 9. 3D measurement results of complex object. (a)(c)(e) Results of error-diffusion dithering method using mild defocusing, moderate defocusing, and severe defocusing, respectively; (b)(d)(f) results of proposed method using mild defocusing, moderate defocusing, and severe defocusing, respectively