報告書番号

U03075

タイトル（和文）

画像計測のためのメッシュレス法に基づくデータ平滑化法の開発

タイトル（英文）

Development of Data Smoothing Technique Based on the Meshless Method for Image-based Measurement

概要 （図表や脚注は「報告書全文」に掲載しております）

近年，画像計測技術が，精度ならびにデータ処理能力において著しく進展した．同技術によれば，対象表面における面的変位情報を得ることが可能であるが，ひずみ分布などを得るためには数値解析的処理が必要とされる．通常，変位計測データには，画像デジタル化等に伴う誤差が含まれており，有限要素などを用いた補間方法を適用して空間微分量を求めると，誤差が増大する．従って，画像計測を高度化するためにも，誤差を低減可能な後処理法の開発が望まれる． 本報告は，新たに開発したメッシュレス法に基づくデータ平滑化法について述べたものである．2次元弾性変形場におけるひずみ解析を対象に，本法と既存の補間方法によるものとを比較した．本法によれば，既存の方法を適用した場合に，ひずみの厳密解に対する相対誤差が50％程度ある場合でも，誤差を15％程度まで低減させ，デジタル画像計測の後処理法として適用性を有していることがわかった．

概要 （英文）

Recently, digital image-based measurements, such as close range photogrammetric technique and full-field optical methods, have been improved remarkably both in accuracy and data acquisition from digital images. These measurement methods make it possible to acquire the full-field displacement information of 2D or 3D measuring objects. However, the information obtained by these techniques, is limited to that of the object surface, and it is not possible to obtain information of the strain distribution without aid of numerical approach. Usually, the measured displacement data often contain the errors associated with the in-site measurement conditions and subsequent digital image processing. Then, strain analysis methods by using conventional finite element or finite difference-based interpolation scheme and so on, often cause oscillation in the approximate function for the spatial derivatives. Therefore, in order to establish a sophisticated image-based measurement system, it is important to develop more useful post-processing method with capability of minimizing the measurement errors and smoothing of strain distribution. In this study, a newly developed data smoothing technique based on the element free Galerkin method is presented. The proposed method is applied to a 2D elastic deformation field, which contains measurement errors. The results of strain analyses by the proposed methods are compared with those by the conventional methods. Numerical examples show that the proposed method could decrease the relative errors to the exact solutions of shear strain, from approximately 50% to 15% , and smooth the spatial distribution of shear strain with more reasonable accuracy, as compared to the conventional methods. As a result, it is found that the proposed method is applicable as post-processing technique for digital image-based measurement.

報告者

キーワード