An image based approach which detects Indian coins of different denomination has been proposed in this paper. This consists of matching an input coin image with a database of coin images (templates) in two phases. The first phase involves, identifying matching radius database coins. In the second phase, template matching is performed by correlating the edges of input and matching radius database coin images. Template matching phase involves two parts, coarse matching and fine matching. This provides rotation invariance and does away with the requirement of placing the front face of the coin up. If the correlation coefficient obtained by template matching satisfies the threshold, then coin stands recognized. The algorithm has been developed in MATLAB 7.9.0 and the obtained results are recorded. The proposed method has been compared with existing methods (Difference, LBP, FFT) and comparison results have been recorded.
展开▼
