Research On Decision-making Method Based On Image Information

Visual sense is one of the main methods of obtaining information from the environment for human beings, and the image is a main way to obtain visual information. Especially in the context of complex industrial process where it is difficult to model accurately, the experienced technicists often use the image as the basis for the control of productive process and fault diagnosis. Image information is often needed to make important decisions. The study focuses on how to make full use of images for automatic decision-making, namely the decision-making method based on image information, which is also derived from the practical needs of engineering project.The study discusses the general situation of the development of the image information processing and recognition technology and decision-making methods, especially the development in the field of using the image for decision-making. The study applies the case-based reasoning to the decision-making, provides case-based decision-making methods, and analyzes the representation, retrieval and correction methods of decision-making case and the maintenance methods of case base.The study introduces the image information into the case-based reasoning technology, takes the image case representation and retrieval as the focus of the study, uses the image information to retrieve the required information efficiently and accurately, and makes decisions by the screening and identification of computer automatically. The study presents the representing method based on the image information, focuses on how to extract image feature information to describe the image, studies the method of extracting the color, texture and shape features of the image, and uses these three characteristics as the property description of the case to structure the model of the image case. Combining content-based image retrieval technique, the study discusses the metrics method of similarity, analyzes the applicable conditions of different metrics methods, focuses on case retrieval methods based on image information, and gives retrieval methods based on the characteristics of image information about color, texture, and shape. Using Matlab2013 software to simulate, the study verifies the feasibility of the case and the validity of the retrieval methods.The study discusses the image information decision-making methods in view of case-based reasoning, and develops an effective decision-making idea for using image information to make automated decisions.