Extraction and recognition of partial shapes of machine parts on an assembly line
Tatsuya Hirata, Hidetoshi Ito, Naohiro Ishii · Systems and Computers in Japan · 1989
Abstract Various studies have been made on the storage and recognition problem for the two‐dimensional shape. An example is the Fourier descriptor which has a versatility in the description and recognition of the closed curve, being robust against expansion‐contraction or rotation. The usefulness of the method has been demonstrated from various standpoints. However, it is not a powerful tool for the partial shape, and little study has been done on that aspect since the contour representing a partial shape is an open curve. A method to represent the feature of the partial shape is the feature vector. In this method, the feature of the partial shape is described by a series of feature points. However, it is not for the description of a closed curve. In the method proposed in this paper: the contour is extracted from the partial image; the condition is stated for the projection shape to be detected from the contour line; and based on the derived condition, the projection shape is extracted. Then new methods for sampling and segmentation are proposed for recognizing the extracted projection shape. In the recognition of the contour line to be extracted in this paper, a new and simple method is proposed based on the concept of the feature vector. By the proposed method, the open curve of the partial shape is represented, and the recognition is performed by examining the matching to the model. It is shown that the proposed method can well recognize and process the partial shape of the machine part on an assembly line.