Canny Edge Detection Algorithm

The Canny edge detector is a multi-stage algorithm for finding the edges, the boundaries of sudden brightness change, within a digital image. It first smooths the image with a Gaussian filter to suppress noise, then computes the intensity gradient at every pixel to find both the strength and direction of the brightest rate of change; it thins the resulting ridges of gradient magnitude down to single-pixel-wide lines through a step called non-maximum suppression, which keeps only the pixels that are local maxima along the gradient direction, and finally applies a double threshold followed by edge tracking by hysteresis, keeping strong edges outright, discarding weak isolated responses, and keeping weak edges only where they connect to a strong one. John Canny developed the algorithm in 1986 alongside a computational theory explaining why the method finds edges accurately, avoids marking spurious edges from noise, and marks each true edge only once, and it remains one of the most widely used edge detection algorithms in computer vision.