IMAGE PROCESSING
In the 21 Century, there is the relentless wave of the `Information Revolution'. The so called `Information Highway' is being connected to our homes and it will eventually be connected to our high-performance PCs and TVs. It will become common for many people to have a high resolution TV with cinema-size screen and equiped with stereophonic hi-fi and to feel the emotional sensation that can now only be found at the movies. Home banking, advanced internet shopping and customized videos will be delivered to our home to allow the pleasure of watching desired programs in comfort. Visual image phones will provide the type of communication that until now has only been possible face to face. Also working in the comforts of home rather than traveling to the workplace will become more prevalent than ever.
Information consists of countless elements including letters, figures, sounds, voices, visual images and audio-visual images. In the `Information Revolution', all this information will be stored and transferred through technology that involves the combination of multimedia products using digitalization techniques. In fact, international competitiveness will inevitably depend on an advanced development that combines the forces of hardware(super-speed communication network) and software(ultra-efficient information network capacity).
Digital-image processing is indeed the most important technology that can assist in producing visual images in this multimedia era.
Human beings fundamentally rely on the sense of sight. We gather 99% of our information through our eyes. Image processing is a process that applies computers and various algorithms to a specific purpose after it acquires familiar visual images through a camera and/or scanner.
1. What is `Digital-image Processing'?
Digital-image processing is the application of the computer in image processing.
Since the 1960s, the digital-image processing has gradually become one of the most important research areas. However, as the image processing algorithm requires a vast processing capacity, its limited development has been in the hands of a few experts. But, with the rapid development of computers, many people has a taken keen interest in image processing. The development of image processing is being accelerated further with the fast advancing related technologies of the parallel-processing technique, cost-effective CCD magnetodiode, maximized capacity of memory chips and cost-effective high-resolution color display system.
Image processing, in broad terms, involves image recognition, analysis, manipulation and transmission. and other related area of adjustment.
(1) Image Manipulation
- Noise(static) reduction function
- Restoration of clear image
- Abstract correction (especially for the satellite sent images)
- Improvement of image contrast
- Artistic alteration
(2) Image Analysis
- Identification of printed and hand-written scripts
- Identification of accessory(parts) measurement through camera
- Precision check-up function for PCB board
- Cell-analysis function for medical fields
(3) Scene Analysis
This is one of the most interesting functions in the image processing as it identifies the number and variety of objects in the image. The optic system of robots and the optic sense of self-driven cars involve this function. However, there are some limitations and futher research and development is needed.
(4) Image Transmission
This is an area where images are sent through cables, satellites and other communication passages. One of its function is the `image compression algorithm' for compressing the huge data of digital images.
Image processing has a close relationship with computer graphics. The field of computer graphics involves creating images using computers where the computer vision researches the image processed, focusing on the areas of image recognition and understanding. Image processing, however, involves transforming the images that are provided by other facilities and can be seen as a reprocessing of the image and/or extracting information from the image.
2. Application of Image Processing
(1) Biological Field
In the fields of biology and biomedicine, image processing is used to analyse biological samples visually. There are some cases where the biological sample analysis is completely automated using image processing.
In order to improve the analysis, visually unidentifiable and/or unclear features from images are improved using the `contrast-balance' or `edge-sharpening' techniques of image processing. Automated identification, classification, categorization and DNA analysis can be performed by image processing.
(2) Military Field
Image processing allows the identification of enemy airports, battleships, bases and missile launches by an automated analysis of satellite pictures. It is also used for identifying targets for `Smart Missiles' and missile inducements.
(3) Document Processing
An automated gathering and processing of documents and pictures is useful for banks and insurance companies. Documents are digitally compressed and stored. Printed information on cheques and tax auditing files are automatically detected and identified.
(4) Factory Automation
Image processing is used to provide an automated inspection and supervision in production lines. This system reduces man-power while providing stability and precision in production.
(5) Medical Diagnostic Imaging
Medical X-ray and CT projections are digitalized to examine internal areas of the body. A number of projections from CT are automatically combined and digitalized to produce 3-dimensional images.
(6) Remote Sensing
A satellite photographs the outer-crust of the Earth at regular intervals and the images are used to analyse crop growth conditions, vegetation distribution and for resource explorations Also the crust of the Earth can be turned into a 3-dimensional model using the pictures taken by satellites.
(7) Video/Film effect
The movie industry uses a variety of image processing techniques to produce special visual effects. Unreal images and otherwise costly scenes are artificially produced using computer graphics and image processing. The main techniques are as follows;
-Mophing
Mophing is a special effect technique that visually transforms one material into another spontaneously producing a natural effect. This technique is widely used in motion pictures and advertisements.
-Image Composition
This technique composites a number of different pictures into one. Many kinds of images, such as computer animated pictures, can be composited together.
3. Structure of Image Processing System
The basic structure of the image processing system consists of sources of light, an image creation function, analog/digital transformers, an image-frame memory function and computers. The first step is converting the reflected lights of an object into electric signals by the image creation function. This converted signal is then, as it is in an analog state, digitalized by an analog/digital converter and is eventually stored in the image-frame memory function. The next step involves the image processing of the stored data. The following picture shows an example of image processing that is currently being used in medical practice.
(1) Lighting
The success of most of industrial image processing systems basically depends on providing appropriate lighting. There are rear-lighting, front-lighting, and dip angle-lighting methods.
(2) Image acquisition
The equipment used for the image acquisition are Image scanner(having light-source, A/D convertor and photographing functions in one), Digital camera(having photographing, A/D convertor and image memory functions in one) and Video camera(having only photographing function).
- The general structure of image acquisition function
The image acquisition function usually involves the process of converting the images that are formed on sensors into electric signals using the dot, line and plane light-sensor functions that are made possible by the PD(Photo Diode) and CCD(Charge Coupled Device) techniques. The plane light-sensor can produce 2 dimensional images at a time, but the dot sensor and the line sensor produce 1 dimensional images as they can only receive one dot or one line data at a time. Therefore, in order to gain 2 dimensional images. both dot and line sensors must be physically operated. Line sensors are used on scanners.
The acquired analog visual images should be space-sampled in x,y directions before being converted using an A/D convertor to reduce the quantum error between analog and digital signals. The digital image data acquired from the conversion by the A/D convertor are stored in the computer for image processing. The quality of the light and color intensities of the digital images relies heavily on the sizes and condensity of data, the breadth of sampling and the efficiency of the A/D convertor of the photographing function. To gain color images, R, G and B color filters need to be positioned in front of the conversion data elements in the photographing function. These filters separate the colors and enable the data to acquire the color intensity required.
Digital cameras, video cameras and scanners are mainly used for the image input facility. The most universal and standard equipment used are video cameras. The video cameras can be divided into two: a semiconductor camera and a vacuum-tube camera. CCD cameras(semiconductor cameras) use chips that contain Pixel sensors. The current technology can arrange around 300,000 sensor elements per square centimeters. The signals, that are read in each line by the detector, generate analog electric pressures. The diagram below is the universal CCD camera's block-diagram.
The older model cameras use vacuum tubes. The light passes through the glass and strikes the coated surface that is reactive to light. The electronic inject-rays also strike this coated surface and the resistance changes incurred are transformed to visual image signals. The structure of the video-conductor is shown in the picture below.
Top
Copyright (C) 2000~2003, iMusicSoft All Rights Reserved. webmaster@imusicsoft.com