부산대학교 도서관

We present a methodology to detect and describe patterns in flow-like images. These patterns are called, in dynamical system theory, critical points or single points. This method involves four steps: estimation of flow directions, pattern matching, pattern confirmation, and description according to dynamical system theory. At the second step, to achieve the critical point class, node, saddle or center point, we use templates to match the patterns. The last step of the method is necessary only if we wish a numerical representation of the patterns. We ran the method over several real images of some fluid flow classes and present the results. The methodology has detected adequately the desired patterns and presented an efficient computational time.