Rotation Tolerant Three-Dimensional Object Pattern Recognition Using a Single Exposure off-axis Digital Holography
Shaotong Feng, Liang Wang and Shouping Nie
Twenty-one reconstructed perspective images were reconstructed from seven off-axis digital holograms using the convolution method, which is very different from the Fresnel approximation method. These images were used to construct a nonlinear filter. The experimental results showed that the true object can be differentiated from a false object by the correlation peak value and the rotation tolerant three-dimensional object pattern.