计算机视觉课件_LC21_cvln_intro

1、3D Computer Vision,Dr. Shengyong Chen Email: ,2,Why vision?,Vision = geometry + measurements + interpretation,3,Content,Introduction Stereo Vision Structured light vision system Active vision for robots Laboratory,4,Introduction,The eyes Methods of 3D vision,5,The Eyes,6,The Eyes,7,Project

2、ion from the Retina,8,The Optic Chiasm,9,Retinotopic organization,The visual field maps onto visual cortex in a systematic way. More of V1 is devoted to processing the central visual field than to the periphery.,10,Mapping of Visual Areas,This map shows a flattened cortex with the known visual areas

3、 mapped onto it.,The Dorsal Stream: Includes areas MT, MST, VIP, 7a, etc. Processes motion, stereo, spatial relationships The where pathway.,11,12,Recommended books Shapiro, L. & Stockman, G. (2001). Computer Vision. Prentice Hall. Richard Hartley and Andrew Zisserman, Multiple View Geometry in Comp

4、uter Vision, Cambridge University Press, 2000. Forsyth and Ponce. Computer Vision: A Modern Approach, Prentice Hall, 2003.,Further,13,Other resources on-line Annotated Computer Vision Bibliography: /Vision- Notes/bibliography/contents.html The Computer Vision Homepage (Carnegie Mel

5、lon University): /cil/vision.html Intel Open Source Computer Vision Library: ,Further,14,Introduction,The eyes Methods of 3D Vision,15,Todays class: Camera and World Geometry,How tall is this woman?,Which ball is closer?,How high is the camera?,What is the camera rotation?,What

6、 is the focal length of the camera?,16,Figures Stephen E. Palmer, 2002,Dimensionality Reduction Machine (3D to 2D),3D world,2D image,17,Projection can be tricky,Slide source: Seitz,18,Projection can be tricky,Slide source: Seitz,19,Projective Geometry,What is lost? Length,Which is closer?,Who is tal

7、ler?,20,Length is not preserved,Figure by David Forsyth,B,C,A,21,Projective Geometry,What is lost? Length Angles,Perpendicular?,Parallel?,22,Projective Geometry,What is preserved? Straight lines are still straight,23,Vanishing points and lines,Parallel lines in the world intersect in the image at a

8、“vanishing point”,24,Vanishing points and lines,o,Vanishing Point,o,Vanishing Point,Vanishing Line,25,Vanishing points and lines,Slide from Efros, Photo from Criminisi,26,Vanishing points and lines,Photo from online Tate collection,27,Note on estimating vanishing points,28,Laser Ranging Systems,Lase

9、r ranging works on the principle that the surface of the object reflects laser light back towards a receiver which then measures the time (or phase difference) between transmission and reception in order to calculate the depth.,29,Laser Ranging Systems,Most laser rangefinders: Work at long distances

10、 (greater than 15m) Consequently their depth resolution is inadequate for detailed vision tasks. Shorter range systems exist but still have an inadequate depth resolution (1cm at best) for most practical industrial vision purposes.,30,Structured Light Methods,Project patterns of light (grids, stripe

11、s, elliptical patterns etc.) onto an object. Surface shapes are then deduced from the distortions of the patterns that are produced on Objects Surface. Knowing relevant camera and projector geometry, depth can be inferred by triangulation. Many methods have been developed using this approach. Major

12、advantage - simple to use. Low spatial resolution - patterns become sparser with distance. Some close range (4cm) sensors exist with good depth resolution (around 0.05mm) but have very narrow field of view and close range of operation.,31,Moire Fringe Methods,A moire projection system,32,Moire Fring

13、e Methods,Moire fringe patterns,33,Shape from Shading Methods,To employ photometric stereo techniques to produce depth measurements. Using a single camera, two or more images are taken of an object in a fixed position but under different lighting conditions. By studying the changes in brightness over a surface and employing constraints in the orientation of surfaces, certain depth information may be calculated.,34,Pa