Image Processing Lecture 01 PDF

Image Processing BY DR. MUKESH KUMAR ECE, VIAET, SHUATS  Introduction of Image Processing Introduction: Processing of images which are digital in nature by a digital computer. There are three terms to find: Processing Image Digital Need of i...

Image Processing BY DR. MUKESH KUMAR ECE, VIAET, SHUATS  Introduction of Image Processing Introduction: Processing of images which are digital in nature by a digital computer. There are three terms to find: Processing Image Digital Need of image processing: Improvement of pictorial information for human perception. (image enhancement for better look) Image processing for autonomous machine application. (Industrial application: Quality control, assembly autonomous) Efficient storage and transmission. (need certain amount of disk storage/for low bandwidth)  Human perception It is mainly employ methods or technique which is capable of enhancing pictorial information for human interpretation and analysis. Typical applications: Noise filtering: Some of images are very noisy, so that it is necessary to filter them for better image quality by image enhance technique. Content enhancement/Contrast enhancement: Some of images are very low contrast so we get better image perception or visualization by enhancement technique Through enhancement technique: Low contrast – High contrast Black & white image – Color image Deblurring: Defocused lens /Motion blurred Medical imaging: CT scan(for find various type dieses) , Ultra sonogram(To check growth of baby) Cont…………… Remote sensing: Aerial image(aerial image also taking for prediction) which is taking from satellite.(Different color for different region & study of various kind/ Changement of environment, river etc.) which is not possible to take image directly. So we get by satellite & by image enhancement technique to find them and resolved this problem: Terrain image: For hilly region Borneo fire: Identify the region of fire region & save the property. Weather forecasting Atmospheric study Astronomy, etc. Image processing technique help to rectify for clear image  Machine Vision Application The interest is on procedures for extraction of image information suitable for computer processing(visualization perception). Here we do not take a enhancement technique but we get extraction of image information by computer processing. Typical applications: Industrial Machine vision for product assembly and inspection Automated target detection and tracking Finger print recognition Machine processing of aerial and satellite imaginary for weather prediction and crop assessment etc. Automated Inspection: Ex.- Bottling plant automation(to get information about bottles- filled or not, properly filled or partially filled). If any bottle which is empty or partially filled any goods and delivered to the costumer, then company reputation will be lost and then company get loss. Cont…………… Boundary information: Ex.- Woods, Identify the boundary (Length/width of boundary) 2D projection of image(automated inspection) Dimension/ Angle of particular image Surface characteristics(uniform or non-uniform) Corner Angle Dimension For tolerance limitation Boundary of image or not Identify by defects on object Surface characteristics by texture processing Inspection of IC manufacturing  Video sequence Processing The Major emphasis of image sequence processing is detection of moving parts. It has various/many application: Detection and tracking of moving target for security surveillance purpose. To find out the trajectory of a moving target. Monitoring the movements of organ boundaries in medical application, etc…… Movement detection: We have able to separate moving background object to detect purpose of security and many more. Image is taken during sun light as well as night. In image we can use infrared image or thermal image or signal and through image processing technique, we can rectify the image. Cont…………… Track the image by window: If object is moving then we can track by window trajectory process by image processing or video processing sequence. Window Co-ordinate system Azimuthal & Elevation (With the help of two different camera to identify X-Y- Z co-ordinate system(3-D system))  Image compression An image usually contains lot of redundancy that can be exploited to achieve compression. To reduce space required to store that image & transmit the image over low bandwidth channel. Intensity of image find to by its neighboring by prediction (known as redundancy) Pixel redundancy Coding redundancy Psycho visual redundancy There are three kinds of redundancy is present in images. In any image two types of information are stores. One is information about image & second is redundancy(intensity of head/head/face etc.). If we remove these type of redundancy by the image, then we have remain only information about image, which is very less to require capacity of storage or very less bandwidth required for transmit. Application: Reduce storage Reduce in bandwidth Lossless Compression: If we remove redundancy from the image, then we have remain only information about image, this is known lossless compression. Lossy Compression: In this compression if we remove redundancy as well as some information of image, which is still acceptable. The quality of image which is reconstructed which will be not a original image, naturally we will some loss/some distortion.  History of Image Processing In 1920’s , submarine cables were used to transmit digitized newspaper pictures between London & New York- Bartlane System. Specialized printing equipments were used to code the picture for cable transmission and its reproduction on the receiving end. This picture was produced by telegraphic printer. In 1921’s , printing procedure was changed to photographic reproduction from tapes perforated at telegraph receiving terminals. This improved both tonal quality and resolution. Bart lane system was capable of coding % district brightness levels. This was increased to 15 levels by 1929. (That means 15 different intensity levels, So quality level of pictures were increased.) Improvement of processing techniques continued for next 35 years In 1964 computer processing techniques were used to improve the pictures of moon transmitted by Ranger 7 at Jet propulsion laboratory. This is time from where the image processing technique of the digital image processing has got a boost. This was the basis of modern image processing techniques. Digital Image Digital image = a multidimensional array of numbers (such as intensity image) or vectors (such as color image) Each component in the image 10 10 16 28  9 656 7026 56  43 37  78 called pixel associates with  32 99 54 70  67  96 56  15 256013902296  67 the pixel value (a single number in   21 54 47  42  the case of intensity images or a 32 158587853943  92 54  65 65 39  vector in the case of color images). 32 65 87 99  Image representation Origin An image can be defined as a 2D y signal that varies over the spatial coordinates x and y, and can be written mathematically as f (x, y). F (x, y) x Refractivity Intensity In sun light, Light from the source falls on the object surfaces, it gets reflected reaches are eye then only we can see that particular object. There are infinite numbers of point in particular image, which can give information, which has direction on X-Y co-ordinates. Every point the intensity value is also a continuous between some minimum and some maximum. Theoretically minimum value will be taken “0”(zero). & Maximum value will be taken “ ”(infinite). So we can not store in digital computer infinite possible intensity value(infinite no.’s) of point.  Nature of f(x,y):  The amount of source light incident on the scene being viewed  The amount of light reflected by the objects in the scene Illumination & reflectance components:  Illumination: i(x,y)  Reflectance: r(x,y) f(x,y) = i(x,y)  r(x,y) 0 < i(x,y) <  and 0 < r(x,y) < 1 (from total absorption to total reflectance)  So. That we try to take some samples value by grid: Spatial discretization by grids. Intensity discretization by quantization. Grids superimpose on picture. (But here every point the value of particular grid is in continuous so that it is not possible by grid.) So after sampling the discretization of intensity values of different samples is known as quantization. Quantization:- The discretization of intensity values of different samples, the process techniques known as quantization. Assume that an image is sampled so that the resulting digital image has M rows & N columns. The values of the co-ordinates now become discrete quantities. For notational clarity and convenience, We shall use integer values for these discrete co-ordinates. Thus, the values of the co-ordinates at the origin are. The next coordinate values along the first row of the image are represented as. Note:- notation is used to signify the second sampled along the first row. So image is represented in matrix, like this: Each element of this matrix array is called an image element, pixel, or pel. An image is represented by a rectangular array of integers. An integer represents the brightness or darkness of the image at that point. M : # of rows, N : # of columns, Q: # of gray levels – M = 2 m, N = , Q =(q is the # of bits/pixel) – Storage requirements: M x N x Q (e.g., M=N=1024, q=8, 1MB)  Elements of Digital Image Processing Acquisition:  Physical Device: o Sensitive to a band in the electromagnetic energy spectrum(such as x- ray, ultraviolet, infrared violet) o Electrical signal output, Scanner, MRI, PET  Digitizer: o Electrical output Digital form  Storage: Three categories of digital storage: 1. Short term storage for use during processing:  Computer memory  Frame buffer(video rate, 30frames/sec.) Instantaneous image (Zoom, scroll, pan)  Physical site of card  Usually 32 Mbytes 2. On-line storage for relatively fast recall:  Magnetic disks:  Winchester disk : Hundreds of Mbytes  Magneto-optical(MO) storage:  Laser and specialized material technologies  Gbytes of storage  Characteristics : frequent access to data  Jukebox : hold 30 to 100 optical disks 3. Archival storage:  Massive storage but infrequent need for access  High density magnetic tape : 6400 B/in  Short shelf life (about 7 years)  Need of controlled storage environment  Write-once-read-many (WORM):  1 GB on 5.25 in. disk  6GB on 12 in. disk  10 GB on 14 in. disk  Not erasable  Shelf life : up 30 years Processing :  Processing of image:  Algorithmic form  Mainly implemented in S/W  H/W implementation: need for speed : low-light microscopy (image averaging)  H/W + S/W : today’s image processing system  Image Processing System :  late in 1980s - early in the 1990s  the form of single board designed to be compatible with industry standard buses. PC and workstations. The principal imaging H/W : digitizer + frame buffer + ALU  image Processing S/W  Obtained commercially  Combined with other S/W (spread sheet, graphics) Communication:  9.6 kbps modem  512 x 512 8 bit image : 5 minutes  Data compression, Decompression Display:  CRT  Printing device  References 1. Rafael, C. Gonzlez., and Paul, Wintz, “Digital Image Processing”, Addison-Wesley Publishing Company, 1987. 2. Rafael C. Gonzalez, Richard E. Woods, “Digital Image Processing”, Pearson, 2011. 3. Jain Anil K., “Fundamentals of Digital Image Processing”, Prentice Hall, 1989. 4. William K. Pratt., “ Introduction to Digital Image Processing”, CRC Press, 2013. 5. S. Sridhar “Digital Image Processing” Oxford University Press, 2011. 6. http://nptel.ac.in/courses/117105079/ (Video lecture on DIP)