Home > Technology
    Technology
        Computer vision technology
        time：06-12-2014souce：hangfa news
      • Computer vision is the study of how to make the machine to "see" science, further said that means using cameras and computers instead of human eyes on target identification, tracking and measurement, machine vision, and image processing, processed by computer to be more suited to the human eye or images that are transmitted to the instrument.
          NASA Mars Exploration Rover's dual-camera system
      • As a scientific discipline, computer vision research-related theory and technology, tried to build images or multidimensional data gets ' information ' artificial intelligence system. Where the information referred to in summer farm defined that can be used to help make a "decision" of information. Because perception can be considered is to extract information from sensory signals, and computer vision can also be considered is the study of how to make the artificial systems from images or multidimensional data in "perception" of science.
      • As an engineering discipline, computer vision seeks to establish a computer vision system based on theories and models. The components of this system include:
      • 1. process control (for example, industrial robots and unmanned vehicles) 2. event monitoring (for example, monitoring) 3. Organization of information (for example, indexing of image databases and image sequence) 4. modeling of objects and the environment (for example, industrial inspection, medical image analysis, and topology modeling) 5. sympathetic interaction (such as human-computer interaction, input devices)
      • Computer vision can also be seen as a biological vision of a supplement. In the field of biological vision, and vision have been studied in humans and animals, so as to establish the Visual system of sensory information in the course of using physical models. The other hand, in computer vision, relying on software and hardware implementation of artificial intelligence system has been studied and described. Biological vision and computer vision for interdisciplinary Exchange brings tremendous value to each other.
      • Computer vision includes the following branches: image reconstruction, event detection, object tracking, object recognition, learning, indexing, image restoration, and so on.
      • Current situation of computer vision
          Relations between computer vision and other areas
      • Computer vision is characterized by its diversity and imperfection
      • Pioneers in this field can be traced back to an earlier time, but it was not until late in the 1970 of the 20th century, when the computer's performance improved enough to handle large-scale data, such as images, computer vision to get the official's attention and development. However these developments often originate from different areas of need, and what is meant by "computer vision problem" has not been formally defined, quite naturally, "computer vision problems" should be how to solve no formula.
      • Nevertheless, people are starting to get part of the way to solve specific computer vision tasks, but these methods are usually applied only to a narrow group of targets (such as faces, fingerprints, characters, etc), which is widely used in different cases.
      • Application of these methods is often used as a component of certain weapons systems to solve complex problems (for example, medical image processing, quality control of industrial manufacturing and measurement). In most practical applications in computer vision, computer is preset to solve specific tasks based on machine learning is growing in popularity, however, once the research further development of the "universal type" computer vision applications might come true.
      • Artificial intelligence research is a major challenge: how to make the system "program" and "decision making"? To make it perform a specific action (such as moving a robot in a particular environment). This problem is closely related to computer vision problems. Here, as a computer vision system sensor, providing information for decision making. Other research interests include pattern recognition and machine learning (which is attached to the field of artificial intelligence, but there is a vital link in computer vision), thus, computer vision, often seen as a branch of artificial intelligence and computer science.
      • Concerns of computer vision aims to fully understand visible light and infrared part of the electromagnetic waves – mainly – the surface formed by the reflection of an image, and this process is based on optical physics and solid state physics, some sophisticated image sensing systems are applied to the theory of quantum mechanics, to parse the image represented in the real world. Meanwhile, many measurements in Physics problems can also be solved by computer vision, such as fluid movement. Also, computer vision can also be seen as a development of physics.
      • Another area of importance is the neurobiology, especially part of the biological vision system.
      • In whole 20th century in the, human on various animal of eyes, and neural Yuan, and and and Visual stimulus related of brain Department Organization are for has widely research, these research obtained has some related "natural of" Visual system how operation of description (despite still slightly too rough), this also formed has computer Visual in the of a child field--people tries to established artificial system, makes of in different of complex degree Shang simulation bio of Visual operation. Also in the field of computer vision, machine learning based methods also have a reference section on biological mechanisms.
      • Another related area of computer vision is signal processing. Signal processing method for many variables, in particular, is a time-varying signal processing, you can naturally be extended to computer vision, signal or binary variables in multivariate signal processing methods. However, due to image data of unique properties, many developed methods in computer vision, but can not find in the unit of signal processing methods to the corresponding version. This type of method is one of the main features is their non-linear, and multi-dimensional image information, two points above as part of computer vision, in signal processing has formed a special study.
      • In addition to the areas mentioned above, a lot of research can also be used as a purely mathematical problem. For example, a lot of problems in computer vision, its theoretical basis is statistically significant, optimization theory, and geometry.
      • How to make the existing method is implemented by various software and hardware, or to say how these practices could be modified, and to attain reasonable execution speed without loss of sufficient precision, is now the main topics of the field of computer vision.