This paper proposes an efficient image interpolation algorithm using optimal edge detector. The proposed interpolation algorithm is done in two steps. In the first step, the missing pixels with diagonal neighbors are interpolated and in the second step, the missing pixels with axial neighbors are interpolated. In both the steps classification of edge and smooth pixels is performed using canny edge detector. For the pixels classified as edge pixels, the direction of the edge is found using various structuring elements. The edge pixels are then interpolated along the directional orientation of the edges. The smooth pixels are interpolated using proportionate variation based interpolation technique in which more weights are assigned for the direction with minimum variation. The proposed interpolation algorithm is applied in the NTSC color space for color images. Conventional image interpolation algorithms like nearest neighbor and bi-cubic interpolation algorithms produces artifacts like edge blurring, zig- zag effects etc. The proposed interpolation algorithm produces super resolution images with improved image quality and less distortion. Experimental results show that in addition to the significant increase in visual effects, this algorithm also manifests improvements in quantitative analysis. Quantitative analysis is done using metrics like PSNR and correlation coefficient. From this analysis it is evident that the proposed algorithm performs better than conventional interpolation algorithms.