《第17章电子信息类专业英语(第二版)课本-李白萍》由会员分享，可在线阅读，更多相关《第17章电子信息类专业英语(第二版)课本-李白萍（35页珍藏版）》请在金锄头文库上搜索。
1、Unit Seven Images and Televisions,Passage A Fundamental Concepts Passage B Compression/Decompression Techniques Passage C Television,Passage A Fundamental Concepts Computational image processing can be defined as the operation of mathematical functions on numeric representations of pictorial scenes.
2、 In general it is part of an overall process of visual perception, pattern recognition and image understanding. These form the essential components of computer vision. Fortunately image processing is conceptually rather simpler than many of the cognitive processes associated with computer vision and
3、 has been the subject of rather more practical and successful effort. Whilst it may be of scientific interest to exploit mathematics in the pursuit of perception, the objective basis of image processing is to apply an algorithm to a representation (usually digital) of a visual scene to produce follo
4、wing results.1 Understanding through quantitation. Perception through an improvement in a chosen index of quality. Efficiency through improved image coding.,Image Analysis The goals of image analysis can take a variety of forms: A complete symbolic description of an image at an adequate level of abs
5、traction. A list of interesting events, or objects, occurring within the image. A description of changes which have occurred between successive recordings of an image. The first of these definitions allows us some flexibility in the chosen level of abstraction. In image processing we will normally b
6、e operating at the lowest level, which names regions, pixels or lines and attributes characteristics. At higher levels we begin to bring in knowledge of the imaged domain. The general structure of the overall imaging system can be simplified as in Figure 7.1.,Figure7.1 A general model of image analy
7、sis,A system for pattern analysis and understanding may require up to seven basic components. A database which contains the outcomes of capture or processing. A module containing algorithms and protocols of processing. A module containing knowledge about the task domain and the candidate nature of t
8、he image. A module containing concepts of control and strategies of processing. A learning module which enhances knowledge in the task domain by analyzing the effectiveness of current processing strategies. A module which explains actions and results and allows interrogation. A user interface enclos
9、ing all of the above into a convenient operational entity.,Image Processing Image processing is largely concerned with: A pre-processing stage which, for example, might attempt to reduce noise or change/ enhance image contrast. The segmentation of the image into labeled areas of interest. The alloca
10、tion of features such as shape, texture, orientation and size.,Typical Application areas are: Inspection Verification of assembly, quality control in production, directional guidance systems, face and signature recognition and document processing. Medicine Tomographic reconstruction, cell recognitio
11、n, automated screening, image-guided surgery and multi-sensor imaging. Environment Remote sensing of land use, environment monitoring and international surveillance. Entertainment Virtual reality and image compression for broadcast purposes.,Image Formation Images are formed either by reflection or
12、transmission, in the first case a typical example is an object illuminated by visible light and recorded with a camera. The geometry of the resultant image is governed by the general translation of a 3D object (world coordinates) to a 2D scene (camera coordinates). Transmission images, on the other
13、hand, result from the passage of radiation through an object. The structure of a resultant image depends on the internal properties of the object. A typical example is the normal medical X-ray film, producing an image in two dimensions of electron density integrated along a ray path, thereby discrim
14、inating bone from soft tissue. A variant of this, tomographic imaging, produces a cross section (transverse) image and is the result of processing transmission profiles taken at a number of different angles of illumination, which are then built into a three dimensional data space.2,The images we dea
15、l with may be simple, as in a single printed character, or complex as in an aerial photograph or medical image. The two fundamental characteristics of images are their spatial resolution (number of pixels in orthogonal axes) and their intensity resolution (number of bits per pixel). How this digital
16、 structure is presented to the observer is a function of the visualization system, not that of the processing system. As an example, a magnetic resonance imaging system might be used to provide data on the movement of the knee joint. Each image plane is collected at a spatial resolution of 256256 pixels with up to 256 intensity levels. In each data volume there are 128 planes. Eight such data sets are collected, each at a different knee angle. This examination would take up over 67 Mbytes and it
