An algorithm based on Fourier series approximation is proposed for separating touching grain kernels. The algorithm starts with image acquisition of wheat grain samples in group of two, three, and four as a first step. Then pre-processing techniques are applied to threshold the required region and make the image smoother, and also to find the boundary points. Pre-processing is followed by chain coding where freeman coding is used to get the connected components and given to Fourier

more »

... for getting boundary contours. Curvature analysis is done to get the corner points. Then possible potential node pairs are found by using radical critical distance. Nearest neighbors among corner point is calculated by finding pair-wise distance among all possible nodal points. From potential pairs, true pairs are selected based on nearest neighbor distance criteria. Finally Bresenham's line drawing algorithm is used to draw segmentation line. The algorithm is successful in separating most of touching grain kernels.