Research On The Principles And Algorithms For Recognizing Curve Object

As an interesting research in computer vision, the object recognition has been found wide applications in various technical sections, such as robot automatic picking, automatic navigating, and detecting object automatically, assembling automatically, analyzing medical and remote sensing image,and so on.Three-dimensional information of an object will be lost while it is projected into a two-dimensional image, so the purpose of recognition is to identify three-dimensional object based on the hidden information in two-dimensional images.This dissertation aims to research on the principles and algorithms for recognizing two-dimensional curve objects under similar transformation and affine transformation.First, object contour information is obtained by edge detecting, denoising and contour tracing. Because the feature points, such as corner, inflexion point and tangent point, are invariant under project transformation, and can describe the contour well, therefore they are extracted for rough matching between model contour and object contour.In order to recognize the contour whose feature points are the same but the curvatures between the feature points are different, A new method for representing and recognizing the contour curve is presented in this paper. First, feature points of the contour are extracted for the rough matching; Then the sampling points of the sub-curve are obtained based on the precision requirement using the given minimal area threshold. A new recognition vector of sample points is defined, and a novel recognition vector matrix is constructed based on the recognition vector of sample points; Last the similarity of the corresponding sub-curves is calculated by compared the recognition vector matrix. The curve is recognized by recognizing their each sub-curve. The matching method is a process from simple to complex, thus many redundancies calculation are avoided. The experiment results show the algorithm is suitable for similar transformation and can recognize the target curve well from some similar contours.In this paper a novel method for recognizing curve is proposed under the affine transformation. First feature points are matched both on model contour and object contour, then the affine region is divided for obtaining a series of sub-region's centroids by using feature points and the whole contour's centroid. Because area ratio of triangles, which constituted by centroid of the whole contour area and centroids of sub-area, is invariant under affine transformation, we introduce it to describe the contour curve. And then a new recognition vector matrix is defined, and a different measure function is defined by us too. Finally the different measure value between model curve and object curve is calculated, which can be employed to judge whether the model and object are matched or not. Theory analysis and experiment results show: the method is feasible, which can not only recognize the different curve object but also distinguish the object with the similar shape according to the different measure value.The recognition algorithms presented in this paper are suitable for similar transformation and affine transformation respectively, and can set up one-to-one correspondence between model features and object features.