The validation model of information measuring channel in technical vision systems
Due to the development of technical vision systems (TVS) and information technology digital images are increasingly used as information models of real and simulated objects allowing to predict, detect and correct the spatial, brightness and color characteristics for different stages of the life cycle. Technical (machine) vision implements the complex process of conversion of video information, which contains six major phases: 1) receiving (perception) of information; 2) preprocessing; 3) segmentation; 4) description; 5) pattern recognition; 6) interpretation. The technical vision system is a sensor device generating images of working scenes and objects, their transformation, the computer processing and interpretation and transfer of results of the control device of the robot. TVS can be classified by principle of operation, functional purpose, autonomy, range, method information, number of cameras, method of placement, method of signal processing, etc. In accordance with the principle of action of TVS it is possible to allocate on the basis of a bistable (logical) systems, coordinators, survey-comparative systems and Biosystems. However, the general basis of these systems is an obtaining of the object digital image as a source of information about it. The process of digital images creating is a technology of the flow of series-parallel operations including data conversion, information and its losses. Therefore, it is necessary to identify and minimize the major sources of losses to obtain reliable and accurate information about the state of the object. This goal can be achieved by properly selecting the elements of information-measuring channel and the organization of the experiment depending on the research objectives. The validation model of a measuring channel based on the presentation of the technical vision system elements as parts of an information measuring channel is described in this article. This model is based on a modular principle and allows to perform measurements in vision systems, depending on the set measuring task.