Extraction Algorithm of Geometric Information on Reinforced Fiber in Composite Materials with Image Processing Technique

摘要

In this paper, the internal heterogeneousness of the fiber reinforced composite materials was evaluated. The image processing system that measured the distributions of the fiber length and fiber direction was examined. The analytical object image gave binary processing. This distinguished the matrix and the fiber respectively as a black pixel and a white pixel. First of all, the algorithm that extracted the geometrical information on the fiber aiming to improve analytical abilities of the system was examined. They are the following two algorithms: One is an algorithm concerning a continuous judgment of the pixel that composes the straight line. The other is an algorithm concerning the identification of each fiber in the fiber intersection part. Images that are difficult to analyze became possible to evaluate by examining these algorithms. In addition, the accuracy of the analysis was examined by processing the thinning on the object image. By processing of making to the thinning, the improvement of an analytical ability was shown. As a result, the identification of the fiber, that the complexity intersects, became possible. In addition, the measurement accuracy concerning the state of distribution of the fiber could be improved.