Machine vision 机器视觉学习笔记（1）Digital Image Fundamentals

Yabooo0

已于 2022-11-23 18:29:03 修改

阅读量318

点赞数 1

文章标签：学习

于 2022-09-12 00:36:21 首次发布

本文链接： https://blog.csdn.net/Yabooo0/article/details/126742903

版权

用matlab做

Digital Image Fundamentals

Image Representation

A projection of 3D scene into 2D.

The image is modeled by a continuous function of two variables f( x, y) or three variables f( x, y, t) ps:( x, y) gives the coordinate of a plane. t is the time domain.

The image function value f correspond to brightness at an image point.

#f对应亮度(或者其他，比如温度)，图像由位置xy表示或者xyt加上时间域参数。

ps：An intensity image is a 2D image bearing information about brightness points.

Image can be considered as a 2D light intensity function, f(x,y), where x and y are spatial coordinates, and f at ( x, y) is related to the brightness or color of the image at that point.

The brightness or intensity of the image is determined by:

Illumination component: the amount of light incident on the viewing scene.#光照成分，入射

Reflection component: the amount of light reflected by the object in the scene.#反射，eg. Example values of typical surfaces: black silk (0.01), stainless steel (0.65), silver plate (0.90).

Image Digitalization
图像需要用离散数据化表达，eg. Matrix or 2D array for 2D image
Image digitalization is the process of converting an image into a numerical representation through sampling and quantization(sampling&quantization)

Sampling turns the continuous function f( x,y) into a matrix of N rows and M columns.

Quantization assigns an integer value to each continuous sample.(为样本赋值，整数)

——Each element of the matrix of samples is called pixel (for PICture ELement).

Pixel

Sampling

Picture Space and Image Space

Image Resolution

Raster Dimension

Raster or image dimension (= width x height) is the number of horizontal and vertical samples in the pixel grid.

examples:

Scaling (or resampling): the process to create an image with different dimensions from that of the source image#缩放（或重采样）

Scaling image down (or decimation): the process of reducing the raster dimension.#缩小图像规模（或称分片）——对贡献于每个输出像素的源像素的数值进行平均化

Scaling image up: the process of increasing the image size to create sample points between the original sample samples in the source raster.#放大——内插法，使用样本网格中的数值来猜测未知像素的值

Quantization

Quantizes the continuous range of brightness or intensity to K gray levels.#灰度量化

K = 2^n

where n is known as the color capability or color depth or bit depth; n is the number of bits used to indicate the color of a single pixel.

Quantization error is defined as the difference between actual analog value and quantized digital value.#量化误差是因为四舍五入或者截断

Brightness level = gray level in monochrome image

Color Depth

Indexed Color

色彩能力=每像素8比特，最大数值=256。每个值都是一个索引号，对应于文件查找表中的明确颜色值。

查找表（颜色查找表或颜色映射）存储在图像文件的头部。颜色查询表可以是一个内置在成像系统中的硬件设备。

色域图被定义为一个颜色列表，每个颜色都以一个整数像素值为索引。色域图中的每个条目被称为一个颜色单元。一个颜色单元代表一种颜色，通常由一组三个数值定义，分别代表红、绿、蓝的强度。

The finer the sampling (i.e., the larger M and N) and quantization (i.e., the larger K), the better the approximation of the continuous image function f(x,y) achieved.The number of quantization levels (brightness or gray levels) should be high enough for human perception of fine shading details in the image.

#采样越细（即M和N越大）和量化（即K越大），实现的连续图像函数f(x,y)的近似度越高。