6.2.2 Basic Steps in Image Processing
Generally, the digital image processing flow could be divided into a series of steps
and subdivisions including image acquisition, image enhancement, image restoration, image compression, image segmentation, image representation, and object
recognition. The context diagram of these steps is illustrated in Fig. 6.1. This section
will give an overview of each step and component in the digital image processing
flow.
Image acquisition: Digital image acquisition or digital imaging is the first step in
any digital image processing system which performs the action of retrieving
image from hardware sources ranging from a personal camera to satellite. The
original image generated by the hardware source is an unprocessed raw image
which is waiting for preprocessing and further image processing operations. In
the stage of image acquisition, some preprocessing steps are involved, such as
scaling; then the preprocessed image will be generated as the input data of the
whole image processing system.
Image enhancement: Image enhancement in a digital image processing system
referring to the process of adjusting and manipulating image and making it more
suitable for displaying in a specific application. Generally, this stage covers a lot
of image processing methods; enhancement techniques adopted in an image
processing system may vary from one task to another. Typical image processing
methods applied in the stage of image enhancement include contrast stretching
such as density slicing and linear/nonlinear stretching, histogram processing such
as histogram equalization and histogram matching, spatial filtering such as
smoothing filtering and median filtering, edge enhancement and detection,
image transformation such as principal component analysis (PCA), and huesaturation-intensity (HSI) transformation.
Image restoration: Image restoration is the process of improving the appearance of
images by recovering a damaged image. The damaged image we talk about in the
process of image restoration usually refers to the images with corruptions like
motion blur, noise, and camera misfocus. Like image enhancement, the objective
of image restoration is to get a clean and suitable image. However, the way image
restoration improves images is different with image enhancement which is mainly
based on human subjective preferences. Major approaches used in image restoration include reducing noise, recovering resolution loss, and applying deblurring
functions on the damaged image.
Image compression: Image compression is the technique about reducing the storage
size of an image. There are many common methods in image compression such as
Huffman coding, arithmetic coding, Golomb coding, predictive coding, adaptive
dictionary algorithms, and wavelet transform. The compressed image, depending
on the compression method, could be lossy or lossless. Common lossy image
formats include JPEG (joint photographic experts group) and GIF (graphics
interchange format). Common lossless image formats include raw image file,
BMP (bitmap image file), and PNG (portable network graphics). Some format
84
C. Zhang and L. Lin
Generally, the digital image processing flow could be divided into a series of steps
and subdivisions including image acquisition, image enhancement, image restoration, image compression, image segmentation, image representation, and object
recognition. The context diagram of these steps is illustrated in Fig. 6.1. This section
will give an overview of each step and component in the digital image processing
flow.
Image acquisition: Digital image acquisition or digital imaging is the first step in
any digital image processing system which performs the action of retrieving
image from hardware sources ranging from a personal camera to satellite. The
original image generated by the hardware source is an unprocessed raw image
which is waiting for preprocessing and further image processing operations. In
the stage of image acquisition, some preprocessing steps are involved, such as
scaling; then the preprocessed image will be generated as the input data of the
whole image processing system.
Image enhancement: Image enhancement in a digital image processing system
referring to the process of adjusting and manipulating image and making it more
suitable for displaying in a specific application. Generally, this stage covers a lot
of image processing methods; enhancement techniques adopted in an image
processing system may vary from one task to another. Typical image processing
methods applied in the stage of image enhancement include contrast stretching
such as density slicing and linear/nonlinear stretching, histogram processing such
as histogram equalization and histogram matching, spatial filtering such as
smoothing filtering and median filtering, edge enhancement and detection,
image transformation such as principal component analysis (PCA), and huesaturation-intensity (HSI) transformation.
Image restoration: Image restoration is the process of improving the appearance of
images by recovering a damaged image. The damaged image we talk about in the
process of image restoration usually refers to the images with corruptions like
motion blur, noise, and camera misfocus. Like image enhancement, the objective
of image restoration is to get a clean and suitable image. However, the way image
restoration improves images is different with image enhancement which is mainly
based on human subjective preferences. Major approaches used in image restoration include reducing noise, recovering resolution loss, and applying deblurring
functions on the damaged image.
Image compression: Image compression is the technique about reducing the storage
size of an image. There are many common methods in image compression such as
Huffman coding, arithmetic coding, Golomb coding, predictive coding, adaptive
dictionary algorithms, and wavelet transform. The compressed image, depending
on the compression method, could be lossy or lossless. Common lossy image
formats include JPEG (joint photographic experts group) and GIF (graphics
interchange format). Common lossless image formats include raw image file,
BMP (bitmap image file), and PNG (portable network graphics). Some format
84
C. Zhang and L. Lin