Neuronal Edge Detection with Median Filtering and Gradient Sharpening

Edge detection is the process of finding out edge of the objects from images. It is the fundamental of pattern recognition in digital image processing field, and it plays an important role in the application of computer vision. A method for edge detection based on median filtering and gradient sharpening was proposed in this paper. At first median filtering was used to remove the noise because it can protect the edge of the image at the same time. Secondly, gradient sharpening was employed to enhance the strength of the edge pixels, and then the initial edge of the image was obtained by simple binarization. The system used a connected component labeling method to remove the residues black block within the enclosed boundaries of the image. By comparing the results of the experiment with the manual extraction results and error analysis, the shortcoming of this algorithm was found for future work. Practice has proved that the edge detection method of this paper has a good effect to speckle images.