An Algorithm of Extracting Signature and Seal of Electric Power Marketing Inspection Documents
Guanyao Liu, Jianyi Li · 2022 IEEE 6th Advanced Information Technology, Electronic and Automation Control Conference (IAEAC ) · 2022
At present, in the on-site inspection of electric power marketing data, manual inspection is needed to check whether documents conform to specifications, mainly to check whether there is a corresponding seal in the approval area of documents and whether there is a signature of the relevant person in charge. With the traditional technology, the inspectors check documents manually and pick out documents that do not meet the approval standards through visual inspection, which brings heavy workload, low efficiency, and a good chance to make mistakes. To realize automatic inspection, we need to scan printed documents into electronic document image, and extract the signature and seal from the image. However, some documents have both the seal and signature requirements, and the signature and the seal often have overlapping areas. The color of the area is different from the color of the original signature and the color of the seal. Therefore, when the signature and seal are extracted through color recognition, the complete seal outline and signature outline cannot be extracted. Based on this, this paper has proposed an algorithm for identifying signatures and seals, the Overlapping Signature and Seal Extraction Algorithm (OSSEA). When the signature and the seal are present on the document, and the signature and the seal have overlapping areas, the image of signature and the outline of the seal can be completely extracted.