Scanner Image De-skewing via Linear Element Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document scanning technologies face challenges in aligning multiple documents of varying sizes and orientations simultaneously, leading to inefficient use of scanning space and potential OCR errors due to incorrect image orientation.
Innovation Solution
A system and method using a multifunction peripheral with an intelligent controller that detects linear elements in each document image and reorients them to align with the scanner's axis data, allowing for flexible placement of documents on the platen and automatic de-skewing during scanning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple documents of varying sizes and orientations are scanned simultaneously, then scanning productivity is improved, but image alignment precision deteriorates
Solution Approach 1:
The scan image is segmented into multiple document regions, and each region is independently processed to detect linear elements and calculate orientation angles. This allows each document to be aligned separately while maintaining overall scanning efficiency.
Solution Approach 2:
The system changes the orientation parameter of each document region by calculating the angle between linear elements and the horizontal axis, then applying a rotation transformation to align each region with the scanner's axis data, thereby correcting skew without compromising scanning productivity.
2Ease of operation
If documents are placed without precise alignment on the platen, then ease of operation is improved, but OCR accuracy deteriorates
Solution Approach 1:
The system performs self-alignment by automatically detecting linear elements in each document region and rotating the regions to match the scanner's axis orientation. This eliminates the need for manual precise placement while ensuring OCR accuracy through proper image orientation.
Solution Approach 2:
The system performs preliminary detection of linear elements and calculation of orientation angles before finalizing the scan image. This preliminary action allows the system to correct misalignment automatically, ensuring OCR accuracy without requiring precise manual placement.
3Manufacturing precision
If automatic de-skewing is implemented, then image orientation accuracy is improved, but device complexity increases
Solution Approach 1:
The patent replaces manual mechanical alignment with an automated image processing system that detects linear elements, calculates orientation angles, and applies rotation transformations. This substitution increases device complexity but achieves superior image orientation accuracy.
Solution Approach 2:
The system introduces an intermediary processing layer between scanning and OCR that automatically corrects orientation. This intermediary layer handles the complexity of alignment calculations and rotations, isolating the complexity from the user while ensuring accurate image orientation for OCR.
Data Source
AI summary
A system and method for aligning concurrently scanned documents, such as receipts, is accomplished on a suitable digital computing device such as a multifunction peripheral. An intelligent controller having a processor and memory operates a scan engine. One or more documents are placed, character side down, on a scanner platen. Axis data is stored and corresponds with a defined orientation of the platen, and thus a defined orientation of a scan image. The controller isolates two or more image portions from the scan image with each image portion defined by a boundary corresponding to a document. The controller detects a linearly oriented image component associated with each document image portion and reorients each image portion to generally align in accordance with each linearly oriented image component and the axis data.