Abstract: A proposed algorithm is using elliptic Fourier series approximation for separating touching grain kernels. Algorithm is based on boundary curvature values. The image acquisition inefficiency can be removed by using image pre-processing techniques and Fourier approximation smoothes the boundary contour which avoids the local pseudo-corners of kernel caused by the presence of rough boundaries. Nodal points separating the touching instances are determined after calculating curvature values along the boundary of kernels. Final nodal points are determined by evaluating the curvature along the boundary of image and selecting those which falls below threshold. With multiple nodal points, the segmentations lines are drawn using nearest-neighbor and critical radian distance difference [rad] of chain coded boundary point. The algorithm uses Breshenham algorithm for drawing segmentation lines. The algorithm appears to be robust enough to separate most of the multiple touching scenarios.