31-07-2012, 01:30 PM
Smart Cameras as Embedded Systems
Smart Cameras as Embedded Systems.docx (Size: 341.59 KB / Downloads: 63)
Abstract.
A "smart camera" is basically a video camera coupled to a computer vision system in a tiny package. This communication begins stating the main differences between smart cameras and standard smart vision systems. A smart camera architecture is described whereby as combination of an on-board microprocessor and PLD’s allow for the embedding of image processing algorithms in the camera. A static thresholding algorithm is presented which demonstrates the ability to track non-uniformity in the inspection target. A multi camera inspection
System application is presented where a maximum of twenty smart cameras may be networked together to a single host computer. Finally, a prediction is made of technological and application future evolution on smart cameras.
Introduction
The smart camera – a whole vision system contained in one neat housing – can be used anywhere, in any industry where image processing can be applied. Companies no longer need a cabinet in which to keep all their computing equipment: the computer is housed within the smart camera. In the pharmaceutical industry and in clean rooms – when not even dust is allowed – this can be a big advantage. A single square metre of space can be comparatively very expensive – if there is no need for a component rack or cabinet, simply a smart camera, then this could save a lot of money. In particular, there would not be the usual cabling involved for other vision systems, and set-up is simple . Later in this communication are stated some advantages of using smart cameras or PC-based systems in vision applications. In usual vision systems scenarios, only a small fraction of a picture frame will be the region
of interest (ROI). In fact, often no visual image of the ROI is even required. The object of a vision system, after all, is to make a decision: "Is there a blob"? "Where is the blob"? "Is this a defect"?
What if all that pixel pre-processing and decision-making could be done within the camera? If all the processing were done inside the camera, the blob analysis of a gigabit image might result in only a few hundred bytes of data which need to be sent somewhere. Such compact packets of data could be easily transmitted directly to a machine control without even passing through a PC. Information should be processed where the information occurs – i.e. the brain should be behind the eyes! The answer to the problem stated earlier is a stand-alone, smart camera.
To illustrate this, a smart camera’s embedded system architecture is shown along with an example of a hardware embedded image processing algorithm: “Static Gray Scale Thresholding”. Many global companies including Glaxo and Ciba-Geigy, and motor companies such as Ford and Volkswagen are users of smart cameras. Others include Sony Music and 3M; the latter has in place up to 150 smart camera systems . A small example of a vision system for web inspection is shown where twenty smart cameras are connected together. Many changes are yet to come concerning smart cameras as in technology as well as in there future applications. An overview, concerning these issues, is made closing this communication. We have to recognise however that smart cameras will not be the answer for all vision applications.
What Is a Smart Camera?
Smart camera is a label which refers to cameras that have the ability to not only take pictures but also more importantly make sense of what is happening in the image and in some cases take some action on behalf of the camera user. For example, a camera that can monitor a door entry and trigger an alarm or send an e-mail to a user when an entry is attempted outside of opening hours would qualify to be a “smart camera” because it can figure out what is happening (detecting a prohibited entry) and take action (triggering an alarm or sending an e-mail). While this may be a good description of what a smart camera is, it is not a technical definition for it. There are many “definitions” of smart cameras in the public space, offered by media, camera manufacturers, developers, etc., but there does not seem to be a well-established or agreed-upon definition for smart cameras.
Many definitions emphasize the fact that smart cameras have built-in image processing ability. We believe these are not sound technical definitions because virtually all digital cameras, consumer or industrial, have built-in image processing capability. For us, what separates a smart camera and a non-smart camera is the nature of the tasks performed by the built-in image processor and the primary outcome or output generated by a smart camera. For the purpose of this book, we define a “smart camera,” or an “intelligent camera,” as an embedded vision system that is capable of extracting application-specific information from the captured images, along with generating event descriptions or making decisions that are used in an intelligent and automated system.
There are several important aspects in this definition, which are analyzed below:
• “Vision system” means that the camera has, obviously, the ability to “see,” or take pictures. “Vision” is not limited to visible light; it can also include other light spectrums such as infrared and thermal imaging. “System” means that all components of the camera do not have to be physically built into single camera housing, even though strictly speaking, they are.
• “Embedded” means that the camera, as an embedded system, employs all necessary components, such as microprocessor(s), memory, power, communication interface, so that it can function in an autonomous and automatic way.
• “Generating event descriptions or decisions” means that the primary function of the camera is not to produce better quality images or videos for people to see, but to detect whether a pre-defined event has occurred, and to do something about it.
Classification of Smart Cameras
The ASIP, or application-specific information processing, is the most essential and defining component of a smart camera. It consists of one or more embedded microprocessor(s) and supporting memory, data buses and other components; its purpose is to provide an efficient computing platform on which powerful and intelligent image processing and pattern recognition algorithms run. The ASIP is at the heart of a smart camera. Strictly speaking, a smart camera is a particular type of embedded vision system, in which all necessary system components – image sensor, capture front end, ASIP, communication interface, I/O – are or can be integrated into a single practical camera casing. However, there exist other embedded vision systems, especially some compact vision systems as described , which could also be classified as smart cameras even though they do not appear to be stand-alone cameras. In fact, many of these systems are referred to as smart cameras within academia and in the research literature. Distributed smart cameras have recently attracted much interest from the research community among multiple cameras, each of them being either a smart or a standard camera. We believe that in some cases these networked cameras can be viewed as a single “virtual” smart camera, where the system ASIP or part of it is supported by the coordination among cameras and network topology. Based on the above discussions, we propose to classify smart cameras into three categories: integrated smart cameras, compact-system smart cameras, and distributed smart cameras. The integrated smart cameras category can be further divided into three types. This integration-level-based classification is presented below and shown in Fig. 2.3. The inclusion of compact-system smart cameras and distributed smart cameras into the classification can be controversial, but this classification does seem to cover most reported work on smart cameras and actual cameras
classification of smart cameras based on levels of integration.
From top to bottom, the camera systems have decreasing level of integration.