Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
A system for engineering flow drawing understanding
Abstract
A system has been developed for automatic understanding of engineering flow drawings. Previous engineering drawing understanding systems considered only specific domains of drawings. Different systems have to be developed for drawings in different domains even if they are similar in nature. We have developed a system for understanding drawings of similar nature that can be classified as flow drawings. It covers a broad range of domains. Flow drawing is a class of drawings characterized by alternate instances of symbols and connection lines. It includes flowcharts and diagrams describing chemical plants, logic circuits, structured chemical formulas, etc. The flow drawing understanding system (FDUS) consists of subsystems that (1) preprocess the thinned and vectorized images of the scanned drawings, (2) segment symbols from the connection lines, and (3) classify symbols and traverse the drawings to correct errors. It contains a symbol library that stores information specific to the domain. FDUS segments symbols using a few simple yet effective generic drawing properties such as (a) symbols contain mostly closed simple shapes, slant lines and open lines, and (b) connection lines are mostly horizontal or vertical. A very high fraction of the symbols are completely segmented, with most others partially segmented using these simple rules. FDUS classifies symbols with a set of simple matchers and evaluates the goodness measure of each classification. Classified symbols are categorized as recognized and unrecognized. The system estimates from the recognized symbols a priority of processing for the unrecognized symbols and corrects errors automatically. A small number of residual errors are corrected interactively. The modular design allows more sophisticated matchers to be integrated easily. The performance is reported in terms of the number of errors interactively corrected. FDUS has been tested with logic circuit diagrams, electrical circuit diagrams, chemical plant flow diagrams, and flowcharts. The system correctly recognized 74.2% and 88.5% of the symbols in electrical circuit diagrams and chemical plant flow diagrams, respectively. Higher recognition rates of 94.5% and 100% were achieved with logic circuit diagrams and flowcharts, respectively.
Subject Area
Computer science|Mechanical engineering
Recommended Citation
Yu, Yuhong, "A system for engineering flow drawing understanding" (1994). ETD collection for University of Nebraska-Lincoln. AAI9510988.
https://digitalcommons.unl.edu/dissertations/AAI9510988