公开(公告)号：KR1020090114162A

公开(公告)日：2009-11-03

申请号：KR1020080039963

申请日：2008-04-29

Applicant: 한국과학기술원

Inventor： 김진근 ,  곽효경 ,  이방연 ,  김재홍

IPC: B28C7/00 ,  B28C7/04

CPC classification number: B28C7/04 ,  B28C7/024 ,  G06F19/00

Abstract: PURPOSE: A method for proportioning concrete is provided to improve rapidity and accuracy about proportioning of the concrete by establishing an available area of database, and to improve of the performance of the concrete. CONSTITUTION: A method for proportioning concrete includes the following steps of: making database including concrete property(S10); predicting the concrete property from the database(S21); establishing an available area of the database(S22); and selecting optimized proportion of the concrete with a direct search method(S30). Artificial Neural Network is used for predicting the concrete property.

Abstract translation: 目的：提供比例混凝土的方法，通过建立数据库可用区域，提高混凝土的性能，提高混凝土配比的快速性和准确性。 构成：比例混凝土的方法包括以下步骤：制作数据库包括具体的属性（S10）; 从数据库中预测具体的属性（S21）; 建立数据库的可用区域（S22）; 并用直接搜索方法选择混凝土的优化比例（S30）。 人工神经网络用于预测混凝土性质。

公开(公告)号：KR100869469B1

公开(公告)日：2008-11-19

申请号：KR1020080030185

申请日：2008-04-01

Applicant: 한국과학기술원

Inventor： 김진근 ,  이방연 ,  김윤용 ,  김정수

IPC: G01N33/36

CPC classification number: Y02A30/336

Abstract: A processing method for improving detection performance of an image of fiber is provided to classify a type of fiber reasonably, to detect the fiber more accurately by using a watershed algorithm and a morphological reconstruction and to obtain high resolution to materials having a low contrast ratio. A processing method for improving detection performance of an image of fiber comprises: a step of obtaining an image of fiber from fiber composite(S10); a step of detecting the image of fiber by using a binarization technique(S20); a step of classifying the detected image of fiber by using a watershed algorithm(S30). A processing method for improving detection performance of an image of fiber improves reliability and an evaluation about a measurement about the fiber composite used in a real spot, and reliability and an evaluation about a measurement about a fiber distribution property.

Abstract translation: 提供了一种用于提高光纤图像检测性能的处理方法，以合理分类纤维，通过使用分水岭算法和形态重构来更精确地检测光纤，并获得具有低对比度的材料的高分辨率。 一种提高光纤图像检测性能的处理方法，包括：从纤维复合材料获得光纤图像的步骤（S10）; 通过二值化技术检测光纤的图像的步骤（S20）; 通过使用分水岭算法对检测出的光纤图像进行分类的步骤（S30）。 用于提高纤维图像的检测性能的处理方法提高了关于实际使用的纤维复合材料的测量的可靠性和评估以及关于纤维分布性能的测量的可靠性和评估。