Suspicious Objects Classification Platform Based On Image Recognition And Service System Software Architecture

Computer image processing and recognition is an interdisciplinary cutting-edge high-tech, widely used in remote sensing, industrial inspection, robot vision, biomedical, public security and many other areas, especially the arrival of the information age of the21st century, makes it an important means of understanding the world and transform the world.Based on actual experience in the development of image processing and recognition software system, in accordance with the object-oriented software engineering methods, identify a common software system framework using in classification of dangerous goods. The system framework define various objects, makes them having flexible scalability and combination, the main contents include1. Propose dangerous goods classification framework based on image recognition algorithm, by analyzing characteristics of the dangerous goods classification algorithm, design this framework based on the image recognition system, makes the framework can be applied to a set of algorithms, rather than a single algorithm.2. Propose a dangerous goods classification domain model. For the domain of dangerous goods classification, makes domain analysis, determine the boundary of the current field, extract the current areas of common needs and changes in demand, and thereby establishes the model of the scene of the domain model, role model, information model, feature model, business model.3. Propose the software architecture of dangerous goods Classification based on image recognition, on the basis of dangerous goods classification domain model, do component design and subsystem design based on image recognition, finally complete the field design of software architecture.4. Proposed a method that select shape and structure characteristic of the image, used in the classification of metal objects in X-ray backscatter image, and get a good classification results in the experiments.