•  
  •  
 

Keywords

power work ticket, table segmentation, optical character recognition

Abstract

The key information of system operation contained in the power work ticket is an important basis for pre-operation risk assessment. In order to quickly and accurately extract the required information from the actual work ticket document, a method of power work ticket segmentation and information extraction is proposed. First, the binarization, expansion, and corrosion operations are utilized to extract the table frame lines from the power work ticket image. Then the segmentation operations are performed to obtain cell images of the work ticket based on the frame lines detection results. The optical character recognition (OCR) is further leveraged to detect the corresponding information in the cell. Finally, the recognition result is structured based on the regular matching method to realize the effective extraction and matching of power operation information. Experimental tests based on actual work tickets show that the proposed method can combine the extracted power operation information in a contextual and semantic combination, and the key information recognition effect is better than general commercial OCR software.

DOI

10.19781/j.issn.1673-9140.2022.06.023

First Page

198

Last Page

205

Download

Share

COinS