Audiovisual Information Fusion Method Based On Perception Alternative

With the development of technology, the modernization of mechanical and electricalequipment tends to be: highly automated, high intelligence, high adaptability and highself decision-making, and this trend is also the electrical and mechanical equipmentoperating conditions and practical problems to be solved decision. Therefore, in order toallow efficient and modern mechanical and electrical equipment, stable job, you needreal-time monitoring on the job site, and according to monitoring results so that the deviceitself to make the appropriate decision-making control in order to better achieve itsfunction, reasonable and efficient completion of the actual task of mechanical andelectrical equipment can be similar to humans and other organisms have the ability to dealwith their own perception of information: the integration of sensory information in avariety of capabilities.Firstly, to understand and analyze the biological, physiological and brain science onthe perception of audiovisual information processing, transmission mechanism, andrelated research by revelation to build the system model for the integration of audiovisualinformation. Secondly, under the guidance of this model diagram, research and analysis ofthe existing fusion theory and methods, their advantages and disadvantages and adaptabilitywere analyzed to determine the fusion method suitable research topic. To solve practicalproblems as a starting point for the construction of the electromechanical system statusmonitoring audiovisual information collection, processing and integration system. Thislevel of integration in the system, according to the capacity of existing hardware forprocessing information and need to be solved practical problems, the determination ofaudiovisual data layer sensor information fusion. Perceived as an alternative to thephysiology revelation, using vOICe basic principles of sound and image consistency willbe encoded using principal component analysis algorithm for extracting characteristicvalues of visual sensor information, while the information fusion, sensor informationconstitutes describe audiovisual feature vector space, this group of feature vectors as theBP network to identify the input, processing by identifying itself algorithm outputs a set ofelectrical and mechanical equipment can accurately describe the state of the data, and dataencoding process, an intuitive way of electromechanical systems state will be described. Finally, under the guidance of theory knowledge to solve practical problems to build avisual sensor information fusion applications, system status for robot crawling movementof objects in the process of monitoring and judgment. The results show that the system canaccurately determine the operational status of equipment, compared with the traditionalmonitoring systems, reducing false positives, the probability of miscarriage of justice, butalso proved the feasibility of fusion theory audiovisual information system thesisconstructed.