Image Processing and Human Perception

Image Processing

• Image processing is the manipulation of digital images by a computer
• Three key terms: Processing, Image, and Digital
• Need for image processing includes: improving pictorial information for human perception (e.g., image enhancement), use in autonomous machine applications (e.g., quality control, assembly), and efficient storage and transmission (e.g., quality control needs).

Human Perception

• Methods and techniques that enhance pictorial information for human interpretation and analysis
• Typical applications: noise filtering (e.g., image enhancement to improve image quality), content enhancement/contrast enhancement (e.g., increasing contrast for better image perception), and deblurring (e.g., defocused lens correction)
• Medical imaging (e.g., CT scans, ultrasounds)

Remote Sensing

• Aerial images or satellite images used for prediction or research. Examples include studying environmental changes (e.g., rivers), identifying fire regions (e.g., Borneo fires), and weather forecasting.

Machine Vision Applications

• Procedures for extracting image information for computer processing (visualization perception).
• Typical applications include automated industrial machine vision for product assembly and inspection, automated target detection and tracking, finger print recognition, processing aerial/satellite imagery for weather prediction and crop assessment.
• Automated inspection, e.g., bottling plant automation (checking if bottles are correctly filled)

Boundary Information

• Identifying boundaries(e.g., length/width).
• 2D projection of images.
• Dimension/Angle of a particular image.
• Surface characteristics of images (uniform or non-uniform). Components to consider: Corners, Angles, and Dimensions.
• For tolerance limitations, analyze boundary issues, identify objects based on defects, and inspect items (e.g., IC manufacturing) based on texture analysis.

Video Sequence Processing

• Emphasizes detecting moving parts in images.
• Example applications that use detection of moving parts in image sequences: security surveillance, finding trajectories of moving targets, monitoring organ boundaries (medical applications).
• Movement detection includes separating moving objects from backgrounds in images or video sequences, enhancing imagery with varied lighting conditions (e.g., sun, night, infrared, thermal), and correcting/improving image clarity (e.g., through image processing techniques).
• Tracking moving objects using image processing or video processing sequences, coordinate systems, and utilizing multiple cameras for 3D systems such as Azimuthal and Elevation.

Image Compression

• Images typically contain redundancy which can be exploited for compression.
• Reducing storage space required and transmission bandwidth.
• Intensity of an image is determined from neighboring intensity values (redundancy). Techniques include pixel redundancy, coding redundancy and psychovisual redundancy
• Three types of redundancy can be present in images: pixel, coding and psycho-visual. Removing these will leave only the necessary information, reducing storage and bandwidth requirements.
• Applications: reduction in storage space and reduction in required bandwidth to transmit the image.

Lossy and Lossless Compression

• Lossless compression: Removing redundancy from image, maintaining all original data
• Lossy compression: Removing redundancy and some information while achieving reasonable/acceptable image quality; naturally resulting in a certain loss/distortion when reconstructing the original image

History of Image Processing

• 1920s: Submarine cables used for transmitting digitized newspaper pictures. Specialized equipment used for coding and transmitting. This led to telegraphic printers.
• 1921: Printing procedure was modified to photographic reproduction from tapes, to improve quality and resolution
• 1929: Increased to 15 levels, improving picture quality.
• 1964: Computer processing techniques applied in space exploration (e.g., Ranger 7 images of the moon)

Digital Image

• Digital image = a multidimensional array of numbers (intensity image) or vectors (color image).
• Each component in an image (called a pixel) is associated with its own value.

Image Representation

• An image is a two-dimensional signal that varies over spatial coordinates (x, y), which can be represented mathematically (f(x, y)).
• In sunlight, light objects reflect in different ways; giving a continuous range of intensity values varying from zero to infinite (min. and max. values).

Image Acquisition

• A physical device (sensitive to specific bands of electromagnetic energy), such as an X-ray machine, an ultraviolet light sensor/scanners, MRI or PET machines, is used to acquire imagery.
• The resulting electrical signal from the device is converted to a digital form.

Storage

• Three types of image storage options exist - short term, online storage, and archival storage
• Short term involves using computer memory for instantaneous image use/processing (e.g., frame buffer within displays and storage on memory cards - 32 Mbytes).
• Online storage provides images for relatively faster recall, such as magnetic disks, winchester disks and magneto-optical (MO) storage.
• Archival allows for large storage amounts with infrequent access, such as high density magnetic tapes (e.g., 6400B/in). Other examples include write-once-read-many (WORM). Storage characteristics determine the appropriate system selection.

Processing

• Image processing algorithms are commonly run within software or a combination of software and hardware. Software algorithms were historically the dominant implementation method.
• Hardware implementations have improved speed, sometimes necessary for specialized applications like low light microscopy.
• More comprehensive systems incorporating both hardware and software have become increasingly common.

Image Processing Systems

• 1980s-1990s single-board systems designed for industry standards allowed for use in PCs and workstations.
• Used digitizers and frame buffers coupled with ALU chips (arithmetic logic unit)
• Commercially available systems typically include support for other software and graphics tools.

Communication

• Image transmission may be done using a 9.6 kbps modem
• Data compression is often used to improve transmission time.

Display

• Image display often uses cathode ray tubes (CRTs).
• Image may also be printed.

This quiz explores the fundamentals of image processing, including its techniques and applications in enhancing pictorial information for human perception. Key areas covered include noise filtering, contrast enhancement, and remote sensing applications. Gain insights into how these methods improve image quality and aid in various fields such as medical imaging and environmental studies.