- Low level image processing involves basic operations on images like noise reduction and contrast enhancement (Digital Image Processing)
- Mid-level processing segments and classifies objects within images (Digital Image Processing/Machine Vision)
- High level Machine Vision makes sense of groups of recognized objects and performs cognitive vision tasks
- Low level image processing involves basic operations on images like noise reduction and contrast enhancement (Digital Image Processing)
- Mid-level processing segments and classifies objects within images (Digital Image Processing/Machine Vision)
- High level Machine Vision makes sense of groups of recognized objects and performs cognitive vision tasks
- Low level image processing involves basic operations on images like noise reduction and contrast enhancement (Digital Image Processing)
- Mid-level processing segments and classifies objects within images (Digital Image Processing/Machine Vision)
- High level Machine Vision makes sense of groups of recognized objects and performs cognitive vision tasks
Muhammad Rzi Abbas Department of Mechatronics and Control Engineering
muhammadrziabbas@uet.edu.pk Lecturer, Mechatronics Dept. University of Engineering and Technology, Lahore Machine Vision • There can be numerous other aspects, but the major objective is: • Acquire it • Process it • Understand it Image Processing vs Machine Vision (Summary) • Low level process (Digital Image Processing) • Primitive operations where inputs and outputs are images • Major functions: image pre-processing like noise reduction, contrast enhancement, image sharpening, etc. • Mid-level process (Digital Image Processing and Machine Vision) • Inputs are images, outputs are attributes (e.g., edges) • Major functions: segmentation, description, classification/recognition of objects • High level process (Machine Vision) • Make sense of an ensemble of recognized objects; perform the cognitive functions normally associated with vision Machine Vision Steps Basics • Images are signals • Signals can be • 1-Dimensional (e.g. anything dependent on time), • 2-Dimensional (e.g. an image dependent on two coordinates on a plane) • 3-Dimensional (e.g. a volumetric object in a 3D space, or a video) • For most of the practical tasks and for the scope of this course an image can be modeled as f(x,y) Basics • The (gray-scale) image function values correspond to brightness at image points. • The function value can express other physical quantities as well • Temperature, • Pressure distribution, • Distance from the observer, • Distance from a particular point in the same image, etc. • Images bearing values regarding brightness at a point are called ‘Intensity Images’ Basics • 2D Intensity Images are merely a projection of the real 3D scene • Recognizing or reconstructing a 3D object from 2D projections is an ill-posed problem. • Another problem is understanding image brightness, because it depends on a number of factors • object surface reflectance properties (given by the surface material, microstructure, and marking) • illumination properties, • and object surface orientation with respect to a viewer and light source Basics • Some applications work with 2D images directly, for example: • an image of a flat specimen viewed by a microscope with transparent illumination, • a character drawn on a sheet of paper, • the image of a fingerprint, etc. • Many basic and useful methods used in digital image analysis do not therefore depend on whether the object was originally 2D or 3D Basics • Image processing often deals with static images, in which time, t, is constant. • A monochromatic static image is represented by a discrete image function f (x,y) whose arguments are two co-ordinates in the plane • Computerized image processing uses digital image functions which are usually represented by matrices, so co-ordinates are natural numbers. Basics • The domain of the image function is a region R in the plane
• where xm , yn represent the maximal image co-ordinates
Basics • The range of image function values is also limited; by convention, in monochromatic images the lowest value corresponds to black and the highest to white. • Brightness values bounded by these limits are gray-levels. Basics • The quality of a digital image grows in proportion to the • Spatial resolution: the proximity of image samples in the image plane • Spectral resolution: the bandwidth of the light frequencies captured by the sensor • Radiometric resolution: the number of distinguishable gray-levels • Time resolution: the interval between time samples at which images are captured References • Image Processing, Analysis and Machine Vision by Milan Sonka, Vaclav Hlavac and Roger Boyle, 3rd Edition, 2008. • Chapter 1 (Sections 1.2 and 1.3) • Chapter 2 (Section 2.1 and 2.2) • Machine Vision by David Vernon, Published in 1991 • Chapter 1 (Section 1.1) • Chapter 3 (Section 3.1 and 3.3) • Computer and Machine Vision – Theory, Algorithms, Practicalities by E.R.Davies, 4th Edition ELSEVIER, 2012 • Chapter 1 (Sections 1.1, 1.2.1 and 1.2.2)