Abstract: In order to improve the noise immunity and accuracy of edge detection algorithm, a one pixel width morphological edge detection algorithm based on adaptive filter is presented. First, the image is filtered with median filter and weighted average filter respectively. Pulse noise and Gaussian noise are suppressed by adjusting the weighted value of median filter results and weighted average filter results adaptively. Then, a morphological gradient with orientation was defined on the basis of that it is able to detect edges with different direction when using structure elements with different tropism, and the gradient is used to extract image edge which is thinned with one pixel width by implementing non-maxima suppression along the orientation of the gradient finally. Simulation results indicate that the proposed algorithm is not only able to detect edge accurately, but also has better noise immunity and faster processing speed.