The Automatic Recognition Of Online Product Identification

This thesis focuses on the design of automatic identification system for the online products. Inspired by the theory and practice of automatic identification of plate license and classification letters, the author tries to explore the theory about identifying online products and carries out some specific researches.First of all, the author gives a general introduction to the image processing technique which is applied in this system, and analyses the features of it. Therefore, the author establishes the research focus of this thesis, and analyses the locating and selecting methods of small target in complex background.Then, the author explores the processing flow, and gives the practical algorithm to solve the designing difficulties in this system. The analysis of the experiment result is also carried out in this part to approve its stability and feasibility. The selection of marked information adjustment of marked images and identification of information features are also concerned here.Finally, the author offers a prospect of this field on the basis of the latest development in the world.Generally speaking, in this project, the task the author has completed including:1st. Based upon the theory of image features analysis, the author puts forward two kinds of target locating algorithm, and realizes the location of automatic identification system in the marked area.2nd. According to the specific requirements, the author provides the processing flow and concrete framework of the system, and has proved its feasibility and stability by the experiment results.