1、Introduction toComputer Vision IntroductionPart IFeature Extraction (2)Edge DetectionCSc I6716Spring 2017Zhigang Zhu, City College of New York zhucs.ccny.cuny.eduIntroduction toComputer Vision IntroductionPart IFeature Extraction (2)Edge DetectionCSc I6716Spring 2017Zhigang Zhu, City College of New
2、York zhucs.ccny.cuny.eduIntroduction toComputer Vision Edge Detectionn Whats an edge?l “He was sitting on the Edge of his seat.”l “She paints with a hard Edge.”l “I almost ran off the Edge of the road.”l “She was standing by the Edge of the woods.”l “Film negatives should only be handled by their Ed
3、ges.”l “We are on the Edge of tomorrow.”l “He likes to live life on the Edge.”l “She is feeling rather Edgy.”n The definition of Edge is not always clear.n In Computer Vision, Edge is usually related to a discontinuity within a local set of pixels.Introduction toComputer Vision Discontinuitiesn A: D
4、epth discontinuity: abrupt depth change in the worldn B: Surface normal discontinuity: change in surface orientationn C: Illumination discontinuity: shadows, lighting changesn D: Reflectance discontinuity: surface properties, markingsACBDIntroduction toComputer Vision Illusory Edgesn Illusory edges
5、will not be detectable by the algorithms that we will discussn No change in image irradiance - no image processing algorithm can directly address these situationsn Computer vision can deal with these sorts of things by drawing on information external to the image (perceptual grouping techniques)Kani
6、zsa TrianglesIntroduction toComputer Vision Another OneIntroduction toComputer Vision Goaln Devise computational algorithms for the extraction of significant edges from the image.n What is meant by significant is unclear.l Partly defined by the context in which the edge detector is being appliedIntr
7、oduction toComputer Vision Edgelsn Define a local edge or edgel to be a rapid change in the image function over a small areal implies that edgels should be detectable over a local neighborhoodn Edgels are NOT contours, boundaries, or linesl edgels may lend support to the existence of those structure
8、sl these structures are typically constructed from edgelsn Edgels have propertiesl Orientationl Magnitudel PositionIntroduction toComputer Vision Outlinen First order edge detectors (lecture - required)l Mathematicsl 1x2, Roberts, Sobel, Prewittn Canny edge detector (after-class reading)n Second ord
9、er edge detector (after-class reading)l Laplacian, LOG / DOGn Hough Transform detect by votingl Linesl Circlesl Other shapes Introduction toComputer Vision Locating EdgelsRapid change in image = high local gradient = differentiationf(x) = step edge1st Derivative f (x)2nd Derivative -f (x)maximumzero crossing
