ニューラルネットワークを用いた格子縞認識による高分子フィルムのひずみ分布計測

概要

論文の詳細を見る
A measuring system of strain distribution has been developed to analyze the mechanical behavior of polymer film. The system is composed of the material testing section and the image processing section. A lattice pattern is printed on the specimen surface. During material testing, the deformation of the pattern is recorded by a VCR camera. After the test, some parts of the recorded image data are digitized. Each datum is treated as follows : (I) smoothing to remove naise, (II) selection of the areas including a crosspoint using the image recognition technique based on the neural network, (III) determination of the coordinate of the crosspoint using the minimum square method, (IV) formation of the triangular mesh model, (V) determining the strain distributions of the specimen computed by applying a similer method used in the finite element method. To demonstrate the applicability of the present method, the strain distributions of a dumbbell-type specimen of polyethylene film under tensile testing are determined using this system.
一般社団法人日本機械学会の論文
1993-03-25