Theroy And Application Of Mathematical Morphology Connectivity

Connectivity plays an important role in image processing and analysis, and particularly in problems related to image segmentation, image filtering, image coding, motion analysis, pattern recognition, and other application areas. In mathematics, connectivity is classically defined using either a topological or a graph-theoretic framework. In general, topological connectivity is useful for images defined over a continuous space, whereas graph-theoretic connectivity is useful for images defined over a discrete space. Although these classical concepts have been extensively applied in image processing and analysis, they are incompatible. In order to meet the present requirements in image processing, Compatibility is desired. This state of affairs motivated G. Matheron and J. Serra to propose an axiomatic approach to connectivity, which unified the traditional concepts of connectivity throughout complete lattice framework.The original theory of mathematical morphology is set-oriented, and is applied to binary images. Then it is extended to grayscale images based on umbra theory. But now, there exists a very successful theoretical framework based on complete lattice for mathematical morphology. Accordingly, the works begin with a deeply discussion about the theory and properties of mathematical morpholgy in the framework of complete lattice. And the works in the dissertation are focus on the theory of connectivity and its applications in industrial detection system.The major contributions in this dissertation are as follows:(1) A new kind of hyperconnectivity class named F-MASK hyperconnectivity class is proposed. In this new hyperconnectivity class, a hyperconnectivity opening operator based on F-MASK hyperconnectivity is defined and analyzed. Throughout the study of hyperconnectivity class and morphological reconstruction operator, F-MASK hyperconnected reconstruction operator is put forward. Finally the morphological filter properties of the new operator are proved, and filter construction method is illustrated.(2) A novel F-MASK hyperconnected opening operator based on multiscale morphological wavelets is presented. And it is applied in road detection for high resolution remote sensing image. The proposed hyperconnected morphological filter integrates the shape information of each connective component and character of major object. Accordingly it is applicable to extract object from complicated background.(3) A novel connected operator named path-reconstruction operator is proposed based on path opening and morphological reconstruction. In order to meet the requirement of real-time application, a quick algorithm for path-reconstruction operator is presented. Combined with the definition of standard connectivity, the novel operator uses the length of path and some properties of object (such as width, contrast et al.) to filter nosie. It is able to simplify images while preserving the contour information.(4) To take the advantage of second generation connectivity, second generation path-reconstruction operator is proposed. Because the new connected operator is furnished with a second generation connectivity class, it can efficiently use the clastering properties to segment images. Compared with traditional operators, the proposed algorithm could effectively extract narrow and long discontiguous objects from noises.(5) An intelligent distribution system is developed based on vision guided AGV. With the research on morphological theory, a realtime guide line recognizing algorithm and multi-branch road seclection method are put forward. To overcome the disturbance from environment and AGV's motion, an online character recognizing algorithm is proposed based on structure features, self calibrate and adaptive filling algorithm.