CS-323 Digital Image Processing Lecture Notes PDF

CS-323: Digital Image Processing L3&4: Image Formation: Acquisition, Sampling, and Quantization Dr. Ahmed Afifi  Light and Electromagnetic Spectrum Each wavelength corresponds to a spectral color: Ultraviolet...

CS-323: Digital Image Processing L3&4: Image Formation: Acquisition, Sampling, and Quantization Dr. Ahmed Afifi  Light and Electromagnetic Spectrum Each wavelength corresponds to a spectral color: Ultraviolet Infrared Each light source has a specific wavelength distribution, and each surface has its own distribution of wavelengths that it reflects. We perceive the color according to the power of the reflected wavelength.  A Simple Image Formation Model The value of f at spatial coordinates (𝑥, 𝑦) is a scalar quantity whose physical meaning is determined by the source of the image, and whose values are proportional to energy radiated by a physical source (e.g., electromagnetic waves). Accordingly, 0 ≤ 𝑓 𝑥, 𝑦 < ∞, 𝑓 (𝑥, 𝑦) = 𝑖(𝑥, 𝑦)𝑟(𝑥, 𝑦) Were, illumination 0 ≤ 𝑖(𝑥, 𝑦) < ∞ and reflectance 0 ≤ 𝑟(𝑥, 𝑦) ≤ 1 The intensity (gray level) of a monochrome image at any coordinates (𝑥, 𝑦), denoted by ℓ = 𝑓 (𝑥, 𝑦) lies in the range 𝐿𝑚𝑖𝑛, 𝐿𝑚𝑎𝑥.  Image Sensing and Acquisition There are three principal sensor arrangements used to transform incident energy into digital images. Single sensing element Array sensor Linear sensor  Image Sensing and Acquisition There are three principal sensor arrangements used to transform incident energy into digital images. 2D images can be formed by movement in two direction Single sensing element  Image Sensing and Acquisition There are three principal sensor arrangements used to transform incident energy into digital images. Linear sensor 2D images can be formed by movement in one direction perpendicular to the linear sensor strip. It can also be used in circular form as in CT scanning.  Image Sensing and Acquisition There are three principal sensor arrangements used to transform incident energy into digital images. Array sensor 2D images can be formed directly and do not require any movement. The most common sensor in digital cameras  Image Sampling and Quantization To create a digital image, we need to convert the continuous sensed data into a digital format. This requires two processes: sampling and quantization. Intensity variations To digitize a continuous image, we must Continuous image along line AB in the sample the function in both coordinates continuous image and in amplitude. Digitizing the coordinate values is called sampling. Digitizing the amplitude values is called quantization. Sampling and quantization Digital scan line.  Image Sampling and Quantization To create a digital image, we need to convert the continuous sensed data into a digital format. This requires two processes: sampling and quantization. The quality of a digital image is determined to a large degree by the number of samples and discrete intensity levels used in sampling and quantization. Continuous image projected onto a sensor array and the sampled and quantized output.  Digital Image Representation Suppose that we sample the continuous image into a digital image, 𝑓(𝑥, 𝑦), containing 𝑀 rows and 𝑁 columns, where (𝑥, 𝑦) are discrete coordinates. For notational clarity and convenience, we use integer values for these discrete coordinates: 𝑥 = 0, 1, 2, … , 𝑀 − 1 and 𝑦 = 0, 1, 2, … , 𝑁 − 1. The numerical representation of an image Image plotted as a surface. Or simply as Image displayed as a visual intensity array. Image shown as a 2-D numerical array. (normalized)  Digital Image Representation Coordinate convention used to represent digital images is as shown below. Because coordinate values are integers, there is a one-to-one correspondence between x and y and the rows (r) and columns (c) of a matrix.  Digital Image Representation Image digitization requires that decisions be made regarding the values for 𝑀, 𝑁, and for the number, 𝐿, of discrete intensity levels. There are no restrictions placed on 𝑀 and 𝑁, other than they must be positive integers. However, digital storage and quantizing hardware considerations usually lead to the number of intensity levels, 𝐿, being an integer power of two; that is 𝐿 = 2𝑘 where 𝑘 is an integer. We assume that the discrete levels are equally spaced and that they are integers in the range [0, 𝐿 − 1]. For example, when 𝑘 = 8, we can quantize the image into 256-levels in the range [0,255]. The number, 𝑏, of bits required to store a digital image is 𝑏 = 𝑀 ∗ 𝑁 ∗ 𝑘  Spatial and Intensity Resolution Spatial Resolution: It can be defined as the number of pixels utilized in construction of a digital image. Images having higher spatial resolution are composed with a greater number of pixels than those of lower spatial resolution. It can be measured by dots per pixel (dpi). 175 × 175 58 × 58 35 × 35 25 × 25 19 × 19 8×8 Same image with different spatial resolution  Spatial and Intensity Resolution Intensity Resolution: refer to the number of bits used to quantize intensity. For example, it is common to say that an image whose intensity is quantized into 256 levels has 8 bits of intensity resolution. 256-level 128-level 64-level 32-level 16-level 8-level 4-level 2-level Same image with different intensity resolution  Reading Material Main Book: Rafael C. Gonzalez Richard E. Woods, Digital Image Processing, 4th Edition. Chapter2: Sections 2.2, 2.3, 2.4 (Except interpolation)