Document Area Detection Using Adaptive Thresholds for Shadowed Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing systems face challenges in accurately distinguishing between single and multiple documents on a transparent platen, particularly when shadows are present, which affects the determination of document areas and feature quantities.
Innovation Solution
The method involves reading an image on a transparent platen, determining object areas, and calculating feature quantities by removing specific areas to differentiate between single and multiple documents based on threshold values, with a secondary threshold used when shadows are detected around thick documents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single threshold value is used to determine document areas, then the processing is simple, but the accuracy deteriorates when shadows are present from thick documents
Solution Approach 1:
The patent applies dynamics by making the threshold value adjustable based on detected conditions. The system dynamically switches between a first threshold value (for normal documents) and a second threshold value (for thick documents with shadows) based on shadow detection results, allowing the threshold to adapt to different document types and lighting conditions rather than using a fixed value
Solution Approach 2:
The patent changes the parameter of the threshold value based on detected document characteristics. When shadows are detected indicating thick documents, the system changes from using a first threshold value to a second threshold value, thereby adjusting the parameter to match the specific document type and improve measurement accuracy
2Productivity
If the first threshold value is used for all documents, then processing is fast, but accuracy deteriorates for thick documents with shadows
Solution Approach 1:
The patent applies preliminary action by performing shadow detection before the final document area determination. The system预先 detects whether shadows are present and identifies thick documents before committing to a threshold value selection, allowing it to prepare the appropriate threshold (first or second) in advance to ensure accurate feature quantity determination
Solution Approach 2:
The system dynamically adjusts the threshold value selection based on shadow detection results. Rather than always using the first threshold value for speed, the system quickly detects shadows and switches to the appropriate second threshold value when needed, maintaining high processing speed while improving accuracy for thick documents
3Device complexity
If shadow detection is not performed, then processing is simpler, but accuracy deteriorates when thick documents are present
Solution Approach 1:
The patent performs shadow detection as a preliminary step before document area determination. By detecting shadows in advance, the system identifies whether documents are thick and will cast shadows, allowing it to select the appropriate threshold value (second threshold) for accurate document type identification and subsequent processing
Data Source
AI summary
There is provided an image processing method which determines the area of a document from an image read on a transparent platen, the image processing method also determines a single object area or a plurality of object areas from the image on the transparent platen, and when the luminance of the area in which the object area is removed from the image on the transparent platen is detected to determine the document area, the image processing method detects whether or not an image of a shadow exists around the periphery of the object area on the image of the transparent platen, and the threshold value used for determining the document area is different depending on whether a shadow is detected.