Image and image composition

The image information input system of the electronic color separation machine adopts a light source to convert the original image information into an optical signal regardless of the form, and then the optical signal is appropriately transformed to give an image as an electrical signal or a digital signal. The processing system, from the perspective of image processing, uses the raster image processing method and photoelectric conversion technology to decompose images and obtain image information. Therefore, before discussing the electronic color image input system, it is necessary to introduce Related concepts.

a representation of an image and an image

In everyday life, when we observe a scene from a certain point, the light emitted by the object (the light emitted by the luminant or the light reflected or transmitted by the object after being illuminated by the light source) enters the human eye and forms on the retina of the human eye. Elephant, this is the objective world seen by the human eye. We call it the scene. This “image” reflects the change of the brightness and color of the objective scene with the spatial position and direction, so “image” is a function of the spatial coordinates. Retinal imaging is a natural physiological phenomenon, and only when human civilization has developed to a certain period of time, it realizes its existence, and tries to record it by various means. This recorded variety of “images” is Call it an image. Therefore, images are the most important means that humans use to express and transmit information. Modern images include both images in the visible light range and images in the invisible light range converted by the adult eye with the help of suitable conversion means.

In a spatial image information, light intensity is its basic element, which varies with the coordinates of the image space (x, y, z), the wavelength λ of the light, and the time t, so the spatial image function can be Expressed as:

For flat images, it is expressed as:

If only the energy of light is considered without regard to its wavelength, the image is visually represented as a black-and-white (gray) image, called a black-and-white image or a monochrome image. Its image function is:

Where: Vs(λ) is the relative visual acuity function.

When considering the color effects of different wavelengths of light, the image is visually represented as a color image, and its image function is

among them:

     

     

      In turn, red, green, and blue primary visual acuity functions

An image whose image content changes with time is called a moving image. The image function is as shown in Equation 2-1. When the image content does not change with time, it is called a still image. Still images are Is the focus of this book's research. Its image function is

Since the printed images are all still images, the color processing is also performed after decomposing them into three primary colors, and each color can be regarded as a monochrome image (black and white image). Therefore, in the following research on the Rhodiola, still black and white The image is modeled.

The scope of the image space is infinite and the field of vision of the human eye is limited. Therefore, for the sake of research, we define the image in a bounded space where the vision can be detected, that is:

Where Lx,Ly is the brightness limit of the visual on x,y.

In summary, the value of an image function at a certain point is defined as the light intensity or gray level, which corresponds to the brightness of the image at this point, and can be represented by a positive real number, and the magnitude of the value is limited. Moreover, a large image gray value indicates a large luminance value, whereas a small image gray value indicates a small luminance value, that is:

Where: Bm - maximum brightness value

Therefore, the image function f(x,y) is a binary, bounded, non-negative continuous function. For reflection type images, the image function is:

Including: Ii - incident light brightness Io - reflected light brightness

In a color-reproduced image, the layout thereof is composed of images and texts, and the layout parts other than the characters and symbols that can be discharged with the character patterns can be referred to as images. Furthermore, the pixels in the image represent the nature of the image. Let P(i, j), 0 ≤ i ≤ M, and 0 ≤ j ≤ N be one pixel in the image, then (1) if P(i) ,j) {0,1},0≤i≤M,0≤j≤N, which means that the image has only two values, that is, the foreground and the background. This kind of image is the line graph in the printing industry in seconds. Image processing is called binary image; (2) If the image P(i,j) indicates that the image has a certain brightness change, this image is called continuous tone image in the printing industry, in image processing. It is called shading (grayscale) image, which indicates that the image not only has a brightness change but also a color change. Such an image is called a color image, and its representation method includes an RGB system, a YM-CBK system, a LUV system, and the like.

Two digital images

Images are classified into continuous images and discrete images. The so-called continuous image refers to an image having a continuous change of f(x,y) and a grayscale value I in a two-dimensional coordinate system . The typical representative of the continuous image is an image acquired by an optical lens system, such as a character, an aerial camera, and the like, which have no natural feeling when viewed with the eye, and are also called analog images, while discrete images are like Figure 2.3 shows that with T as a period, the x, y axes are divided into checkerboard grids, and only discrete gray values ​​at each intersection point are taken. Such an image is called a discrete image or a sample plot. Sampling Image. Therefore, when discrete grey values ​​are used to express the gray value of an image, such an image is called a quantization image. A so-called digital image is a decomposition of an image. As shown in Fig. 2.4, the small discrete points are referred to as pixels (Picture element), and the gradation value of each pixel is represented by an integer value which is a quantized discrete value.

The method of decomposing an image into pixels is shown in Fig. 2.5. According to the plane, a rectangular array , a Hexagonal array, and a Tringular array are arranged. The square array is most commonly used.

For an image, a digital array (Tringular Array) is obtained from the simulated image , with the square array being the most commonly used. For an image, to obtain a digital image from an analog image, spatial sampling and quantization must be performed as shown in Figure 2.6.

Space sampling

(1) The concept of spatial sampling

Sampling refers to an operation of transforming a spatially or temporally continuous image (analog image) into a set of discrete sampling points (pixels).

Prepressed images are basically two-dimensional plane information distribution. To input these image information into a computer for processing, the inverse two-dimensional image signal is first converted into a one-dimensional image signal, which must be realized by scanning . The most common method is to scan horizontally or vertically in a straight line from top to bottom at predetermined intervals on a two-dimensional plane to obtain an image gray value array, ie, a set of one-dimensional signals, and then determine For each specific interval, discrete signals can be obtained.

Assuming an image, if the number of pixels in the x direction is M and the number of pixels in the y direction is N when sampling , the image is represented by discrete M×N pixels, that is, the image When processing, only the image gradation of M*N points needs to be processed.

In the actual sampling process, the selection of sample point spacing is an extremely critical issue. Since the image contains the first varying degrees of subtle density variations, the sampling points need to be based on the degree to which they faithfully reflect the image. The sampling theorem (Sampling Theorem/Shannon Theorem) points out that if the spatial frequency of a dimension signal g(t) is limited below , the sampled value g ( iT), where i=..., -2, -1, 0, 1, 2..., can actually recover (reconstruct) g(t).

among them: